Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A comparison of Pan and zoom and rubber sheet navigation Nekrasovski, Dmitry 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2006-0088.pdf [ 7.06MB ]
Metadata
JSON: 831-1.0051176.json
JSON-LD: 831-1.0051176-ld.json
RDF/XML (Pretty): 831-1.0051176-rdf.xml
RDF/JSON: 831-1.0051176-rdf.json
Turtle: 831-1.0051176-turtle.txt
N-Triples: 831-1.0051176-rdf-ntriples.txt
Original Record: 831-1.0051176-source.json
Full Text
831-1.0051176-fulltext.txt
Citation
831-1.0051176.ris

Full Text

A Comparison of Pan and Zoom and Rubber Sheet Navigation b y D m i t r y N e k r a s o v s k i B . C . S . , C a r l e t o n Univers i ty , 2000 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M a s t e r of Science i n T h e F a c u l t y of G r a d u a t e Studies ( C o m p u t e r Science) T h e U n i v e r s i t y O f B r i t i s h C o l u m b i a J a n u a r y 2006 © D m i t r y N e k r a s o v s k i 2006 ii Abstract A s i n f o r m a t i o n v i s u a l i z a t i o n tools arc used l o v i sua l i ze datasets of i n c r e a s i n g size, there is a g r o w i n g need for techniques that faci l i tate efficient n a v i g a t i o n . P a n a n d z o o m n a v i g a t i o n enables users to d i s p l a y areas of interest at different resolut ions . F b c u s + c o n t e x l techniques a i m to o v e r c o m e the d r a w b a c k s o f p a n a n d z o o m b y d y n a m i c a l l y i n t e g r a t i n g areas o f interest a n d context regions . T o date , e m p i r i c a l c o m p a r i s o n s of these two navigat ion p a r a d i g m s have been l i m i t e d in scope a n d inconc lus ive . In two c o n t r o l l e d studies , we evaluated n a v i g a t i o n techniques representa-t ive of the p a n a n d z o o m a n d focus (-context approaches , ' the p a r t i c u l a r fo-cus-(-context t echnique e x a m i n e d was r u b b e r sheet n a v i g a t i o n , i m p l e m e n t e d in a way that af forded a set of nav iga t ion act ions s i m i l a r to p a n a n d z o o m n a v i -ga t ion . T h e two techniques were used by 10 subjec t s in each s t u d y to p e r f o r m a nav igat ion- in tens ive task in a large tree datase t . S t u d y 1 inves t igated the effect of the a m o u n t of screen real estate d e v o t e d to context regions for each n a v i g a t i o n technique . P e r f o r m a n c e wi th b o t h techniques was not s igni f icant ly affected by this factor, bu t was influenced b y technique-spec i f ic s trategies de-ve loped by subjects . S t u d y 2 c o m p a r e d the p e r f o r m a n c e of the two techniques . P a n a n d z o o m n a v i g a t i o n was found to be faster t h a n r u b b e r sheet n a v i g a t i o n a n d was rated by subjects as easier and less m e n t a l l y d e m a n d i n g . W e discuss the i m p l i c a t i o n s of these results , inc lud ing the r e l a t i o n s h i p between n a v i g a t i o n technique , task, a n d user strategy, and propose d i rec t i ons for f u l u r e work . Contents A b s t r a c t ii C o n t e n t s i i i L i s t of T a b l e s v i L i s t o f F i g u r e s v i i A c k n o w l e d g e m e n t s ix 1 I n t r o d u c t i o n 1 1.1 M o t i v a t i o n 1 1.2 O v e r v i e w 2 2 R e l a t e d W o r k 7 2.1 P a n a n d Z o o m N a v i g a t i o n ( P Z N ) 7 2.2 R u b b e r Sheet N a v i g a t i o n ( R S N ) 9 2.3 G u a r a n t e e d V i s i b i l i t y 13 2.4 E v a l u a t i o n of N a v i g a t i o n T e c h n i q u e s -. 14 2.5 E v a l u a t i o n of T r e e V i s u a l i z a t i o n s 17 3 T a s k a n d Dataset : \ . 19 3.1 Task 19 3.1.1 T a s k 1: Determining the lowest c o m m o n ances tor 20 3.1.2 T a s k 2: D e t e r m i n i n g the topo log ica l d i s tance between nodes 21 3.1.3 T a s k 3: D e t e r m i n i n g whether two subtrees are adjacent . 21 Contents iv 3.1.4 T a s k 4: D e t e r m i n i n g w h e t h e r a subtree c o n t a i n s u n m a r k e d nodes 2 2 3.2 D a t a s e t 2 3 4 S t u d y 1 2 5 4.1 H y p o t h e s e s 2 ^ 4.2 Interfaces 2 6 4.2.1 N a v i g a t i o n 2 9 4.2.2 O v e r v i e w s a n d F o c i • 30 4.2.3 G u a r a n t e e d V i s i b i l i t y a n d Leve l s of C o n t e x t 30 4.2.4 R S N - N o O V Interface 34 4.2.5 P Z N - N o O V Interface 34 4.2.6 R S N + O V Interface 35 4.2.7 P Z N + O V Interface 35 4.3 T a s k a n d D a t a s e t 36 4.4 A p p a r a t u s 36 4.5 P a r t i c i p a n t s : 36 4.6 D e s i g n 37 4.7 P r o c e d u r e . . : 38 4.8 M e a s u r e s 39 4.9 R e s u l t s • 4 0 4.10 S u m m a r y o f R e s u l t s 4 2 4.11 Strategies 4 5 4.12 D i s c u s s i o n 4 6 5 S t u d y 2 4 9 5.1 H y p o t h e s e s 4 9 5.2 Interfaces : 5 0 5.3 T a s k a n d Dataset ">n 5.4 A p p a r a t u s 54 5.5 P a r t i c i p a n t s 5 4 C o n t e n t s v 5.6 Design 54 5.7 Procedure 55 5.8 Measures 57 5.9 Results 57 5.10 Summary of Results 61 5.11 Discussion 61 6 C o n c l u s i o n s a n d F u t u r e W o r k 64 B i b l i o g r a p h y 67 A S t u d y 1 T r a i n i n g P r o t o c o l .' 74 B S t u d y 2 T r a i n i n g P r o t o c o l 81 C S t u d y 1 Q u e s t i o n n a i r e s 90 D S t u d y 2 Q u e s t i o n n a i r e s 110 v i List of Tables 4.1 Interface c o m b i n a t i o n s e x a m i n e d in Studies 1 a n d 2 25 4.2 L a t i n square used to c o u n t e r b a l a n c e the o r d e r of presentat ion for level of contex t in S t u d y 1 37 4.3 Levels of contex t used in S t u d y 1 38 4.4 Resul t s o f A N O V A s of level of context by interface in S t u d y 1 . . 41 4.5 N a v i g a t i o n strategies deve loped by subjects in S t u d y 1 46 5.1 M e a n s a n d effects of n a v i g a t i o n technique in S t u d y 2 59 v i i List of Figures 1.1 P a n a n d z o o m n a v i g a t i o n ( P Z N ) 3 1.2 R u b b e r sheet, n a v i g a t i o n ( R S N ) 3 2.1 P a n n i n g m e t a p h o r s e x a m i n e d by J o h n s o n 8 2.2 R u b b e r sheet n a v i g a t i o n with o r t h o g o n a l s t r e t c h i n g 11 2.3 R u b b e r sheet, n a v i g a t i o n w i t h p o l y g o n a l s t r e t c h i n g 11 2.4 R u b b e r sheet n a v i g a t i o n in T r e e J u x t a p o s e r 12 2.5 G u a r a n t e e d v i s ib i l i ty in C i t y L i g h t s [51] 15 2.6 G u a r a n t e e d v i s ib i l i ty in p a n a n d z o o m interfaces 15 3.1 T a s k 1: D e t e r m i n i n g the lowest c o m m o n ances tor ( L C A ) 20 3.2 T a s k 2: D e t e r m i n i n g the topo log ica l d i s tance between nodes . . . 21 3.3 T a s k 3: D e t e r m i n i n g whether two subtrees are adjacent . . . . . . 22 3.4 T a s k 4: D e t e r m i n i n g w h e t h e r a subtree c o n t a i n s u n m a r k e d nodes . 22 4.1 S t u d y 1 R S N - N o O V interface 27 4.2 S t u d y 1 P Z N - N o O V interface 27 4.3 S t u d y 1 R S N + O V interface 28 4.4 S t u d y 1 P Z N + O V interface 28 4.5 M o t i v a t i n g scenario for use of mult ip le foci in S t u d y 1 31 4.6 C a l c u l a t i o n of levels .of context, in S t u d y 1 R S N interfaces . . . . 33 4.7 C a l c u l a t i o n of levels of contex t in S t u d y 1 P Z N interfaces . . . . 33 4.8 M e a n p e r - t r i a l c o m p l e t i o n t imes by interface in S t u d y 1 41 L i s t of Figures v i i i 4.9 M e a n p e r - t r i a l c o m p l e t i o n t imes a n d trend l ine for the R S N -N o O V interface 43 4.10 M e a n p e r - t r i a l c o m p l e t i o n t imes a n d trend l ine for the P Z N -N o O V interface 43 4.11 M e a n p e r - t r i a l c o m p l e t i o n t imes a n d trend line for the R S N + O V interface 44 4.12 M e a n p e r - t r i a l c o m p l e t i o n t imes a n d trend line for the P Z N + O V interface 44 5.1 S t u d y 2 R S N - N o O V interface . . . ! 51 5.2 S t u d y 2 P Z N - N o O V interface 51 5.3 S t u d y 2 R S N + O V interface 52 5.4 S t u d y 2 P Z N + O V interface 52 5.5 C a l c u l a t i o n o f levels of context in S t u d y 2 R S N interfaces . . . . 53 5.6 C a l c u l a t i o n o f levels of context in S t u d y 2 P Z N interfaces . . . . 53 5.7 M e a n p e r - t r i a l c o m p l e t i o n t imes b y interface in S t u d y 2 58 5.8 M e a n p e r - t r i a l c o m p l e t i o n t imes by n a v i g a t i o n technique a n d b y b lock 60 ix Acknowledgements R e s e a r c h is a lways a co l laborat ive process , a n d this thesis is no e x c e p t i o n . I w o u l d like to t h a n k al l the people w h o have he lped m e in m a k i n g it h a p p e n . M y superv i sors , J o a n n a M c G r e n e r e a n d T a m a r a M u n z n e r , p r o v i d e d cop ious a m o u n t s o f adv ice , g u i d a n c e , a n d s u p p o r t t h r o u g h o u t this pro jec t . It has been a fantast ic l e a r n i n g exper ience a n d a real pleasure to work w i t h b o t h o f t h e m . Adam B o d n a r c o l l a b o r a t e d wi th m e on most aspects of the work d e s c r i b e d in litis thesis. H i s hard work , creat ive ideas, a n d ready wit have been essential to the success o f this pro jec t (and the preservat ion of m y sanity!) Frango i s G u i m b r e t i e r e c o m m i t t e d a great deal of t ime to he lp shape a n d d i -rect this work . H i s exper t adv ice a n d a t t en t ion to deta i l have p r o v e n inva luab le . K a r o n M a c l e a n somehow found t i m e to act as second reader for this thesis a m o n g all her o ther c o m m i t m e n t s . H e r he lpfu l input is grea t ly a p p r e c i a t e d . M a n y o ther peop le c o n t r i b u t e d to this work in a var ie ty of ways . M o t i v a t i o n for the studies c a m e from discussions w i t h phylogenet ic biologists , espec ia l ly D a v i d H i l l i s , W a y n e M a d d i s o n , a n d members o f their research g r o u p s . Early-feedback was p r o v i d e d by P a t r i c k B a u d i s c h , B e n B e d e r s o n , J e a n - D a n i e l Fekete . C a t h e r i n e P l a i s a n t . a n d K a t h e r i n e St . J o h n . L a t e r on . s t u d y des ign a n d ana lys i s was inf luenced by conversat ions w i t h L e a h F i n d l a t e r , H e i d i L a m . B a r r y P o . a n d M e l a n i e T o r y , whi le J a m e s Slack assisted w i t h i m p l e m e n t a t i o n issues. M e m b e r s of the Imager , M U X , a n d H C T labs h e l p e d out as p i lo t s t u d y subjects a n d created a s t i m u l a t i n g a n d fun e n v i r o n m e n t for c o n d u c t i n g this research. L a s t b u t most c e r t a i n l y not least. J u l i e Lavo ie a n d m y parents , A l e x a n d Sofia N e k r a s o v s k i , have g iven m e b o u n d l e s s love a n d cons tant e n c o u r a g e m e n t . T h i s is for y o u . Chapter 1 i Introduction 1.1 M o t i v a t i o n In format ion v i sua l i za t ion uses c o m p u t e r - s u p p o r t e d interact ive v i sua l represen-tat ions of abstract d a t a to a id cogni t ion [13]. I n f o r m a t i o n v i sua l i za t ion tech-niques are now b e i n g a p p l i e d to fields such as phylogenet ic biology, w h i c h is concerned w i t h d iscover ing evo lut ionary relat ionships between species, a n d re-quires increas ingly sophis t icated v i sua l i za t ion tools for this purpose . A s infor-m a t i o n v i sua l i za t ion tools are used to v isual ize datasets of increased size a n d complexi ty , there is a growing need for techniques that faci l i tate r a p i d a n d effi-cient navigat ion in such datasets. T w o p r i m a r y approaches have been proposed i n the in format ion v i sua l i za t ion l i terature to enable such nav iga t ion . P a n a n d z o o m nav igat ion ( P Z N ) [24] relies on a c o m b i n a t i o n of p a n n i n g a n d z o o m i n g operat ions to enable users to v iew discrete por t ions of the dataset at differ-ent resolutions. P a n a n d z o o m nav iga t ion techniques are often p a i r e d w i t h overv iew w i n d o w s to prov ide users w i t h contextua l i n f o r m a t i o n a b o u t ar-eas outs ide the region of current interest. However , overview windows take u p screen real estate a n d m a y not prov ide enough reso lut ion to c learly identi fy fea-tures of interest. F o c u s - f context ( F + C ) nav iga t ion techniques [11] combine h igh-reso lut ion displays of areas of interest t o the user (focus regions) a n d con-tex tua l i n f o r m a t i o n a b o u t the rest of the dataset (context regions) into a single unif ied view. In order to a c c o m p l i s h this w i thout sacrif ic ing screen real estate, m a n y focus+context interfaces re ly o n d i s tor t ion to d y n a m i c a l l y integrate focus a n d context regions as users navigate t h r o u g h the dataset . S o m e researchers Chapter 1. Introduction 2 have suggested that focus+context techniques m a y also benefit f rom the use of overview windows [5]. T h e r e has been no consensus in the l i terature as to w h i c h a p p r o a c h is superior , a n d some researchers have suggested that their relative effectiveness is h igh ly dependent on the p a r t i c u l a r nav iga t ion task for w h i c h they are used [22, 25]. T h e relat ive per formance of p a n a n d z o o m a n d focus+context nav iga t ion techniques c a n be inf luenced by a variety of factors. T h e s e inc lude the t y p e of dataset b e i n g nav igated a n d its v i sua l representat ion, the n a t u r e a n d level of difficulty of the navigat ion task, the interact ions afforded b y the nav iga t ion tech-niques, a n d the fract ion of screen real estate al lotted to context regions, whether integrated w i t h focus regions or presented in a separate overview window. T h i s last factor is referred to i n this thesis as level of context . T h e dataset types that most c o m m o n l y mot ivate research in the area of nav iga t ion i n the i n f o r m a -t i o n v i sua l i za t ion l i terature due to their use in a variety o f a p p l i c a t i o n d o m a i n s inc lude t ex tua l d o c u m e n t s [5, 25], m a p s [22, 24], a n d graphs , especial ly tree s tructures [30, 43]. A variety of nav iga t ion tasks c a n be p e r f o r m e d w i t h each of these dataset types . E x a m p l e s inc lude v i sua l search, brows ing , c o m p a r i s o n , a n d m o r e c o m p l e x c o m p o u n d tasks c o m p r i s e d o f mul t ip l e instances o f these tasks. O n e category of task that m a y be p a r t i c u l a r l y sui table for assessing per formance of nav igat ion techniques due to their potent ia l for requ ir ing significant a m o u n t s of nav igat ion is tasks that involve u n d e r s t a n d i n g the topology of a g r a p h or tree s tructure . T h e interact ions afforded by variants o n the p a n a n d z o o m a n d focus+context nav igat ion m e t a p h o r s often inc lude some or a l l of p a n n i n g , z o o m -i n g i n , a n d z o o m i n g out , a l l of w h i c h can be i m p l e m e n t e d i n a variety of ways d e p e n d i n g on the design goals a n d intended dataset of a par t i cu lar n a v i g a t i o n technique. L e v e l o f context affects the a m o u n t of contex tua l i n f o r m a t i o n avai l -able to users d u r i n g nav igat ion , a n d m a y r e m a i n stat ic or change d y n a m i c a l l y d e p e n d i n g on user interact ions . In recent years , e m p i r i c a l eva luat ion has ga ined increased p r o m i n e n c e in in format ion v i sua l izat ion l i terature [16]. R e c e n t l y pub l i shed user studies have evaluated c o m m e r c i a l l y avai lable v i sua l i za t ion tools [30], c o m p a r e d the usabi l i ty Chapter 1. Introduction 3 of different v i sua l i za t ion techniques for specific tasks [25], a n d e x a m i n e d pat-terns associated w i t h the use of v i sua l i za t ion tools i n the field [34]. A n u m b e r of these studies have either c o m p a r e d the per formance of different nav iga t ion tech-niques or e x a m i n e d interfaces w i t h a n d wi thout a n overview based o n the same nav iga t ion technique. However , these investigations have typ ica l l y invo lved a re lat ive ly narrow subset of var iat ions o n the p a n a n d z o o m a n d focus+context p a r a d i g m s , as well as differences i n v i sua l representat ion a n d interact ions af-forded b y interfaces us ing each of the two types of nav iga t ion , m a k i n g s tudy results difficult to interpret . F u r t h e r m o r e , the datasets a n d tasks used i n most of these studies have been devised specif ical ly for s t u d y purposes , w i thout re-g a r d to the needs of users i n a p a r t i c u l a r d o m a i n , a n d no s t u d y has eva luated nav igat ion technique a n d presence of overview as o r t h o g o n a l factors. F i n a l l y , to date level of context has not been e x a m i n e d as a factor i n e m p i r i c a l studies of nav iga t ion techniques. T h e work descr ibed i n this thesis represents a first a t t e m p t to fill these gaps i n the l i terature . 1.2 O v e r v i e w T h i s thesis describes exper iments p e r f o r m e d to quant i ta t ive ly evaluate the ef-fect of n a v i g a t i o n technique (pan a n d z o o m vs. focus+context ) , presence or absence of a n overview, a n d level of context o n user per formance a n d satisfac-t ion . W e chose to specif ical ly examine presence or absence of overview a n d level of context due to the lack of e m p i r i c a l results concern ing the relat ive influence of these factors o n per formance w i t h each of the two n a v i g a t i o n metaphors . T h e specific focus+context technique that we chose to evaluate is r u b b e r sheet nav iga t ion ( R S N ) [45], w h i c h allows users to stretch or squish focus areas as t h o u g h the dataset was l a i d out o n a r u b b e r sheet w i t h its borders na i led down . T h e differences between the two nav igat ion techniques used i n our exper iments are i l lus trated i n F i g u r e s 1.1 a n d 1.2. W h i l e a z o o m act ion i n P Z N causes areas outs ide the selected region of interest to move off-screen, the equivalent o p e r a -t ion i n R S N causes these areas to be compressed a r o u n d the edges of the view Chapter 1. Introduction 4 F igu re 1.1: Se lec t ing (left) a n d result of zooming in to (r ight) a rec t i l inear region w i t h p a n a n d z o o m nav iga t ion . A reas outs ide the zoomed region are pushed off-screen. 1 F i g u r e 1.2: Se lec t ing (left) and resul t of zooming in to (r ight) a rec t i l inear reg ion w i t h rubbe r sheet nav iga t ion . A reas outs ide the zoomed region are compressed a round the edges of the v iew. bu t r ema in v is ib le . W e chose R S N as the most appropr ia te representat ive technique because of the fact t ha t it is the on ly d is tor t ion-based focus+contex t technique to date to be comb ined w i t h g u a r a n t e e d v i s i b i l i t y [37], a p rope r t y tha t ensures tha t regions of interest rema in v is ib le independent of user nav iga t ion ac t ions . W h i l e guaranteed v i s ib i l i t y is a re la t ive ly new concept i n the i n fo rma t ion v i sua l i za t i on l i te ra ture , a recent s tudy [5] suggests that it m a y p rov ide benef i ts in te rms of b o t h per fo rmance and user preference. In an effort to use the best ava i lab le i m -p lementa t ions of each nav iga t ion technique for our exper iment , i n our s tud ies we p rov ided guaranteed v is ib i l i t y in interfaces us ing b o t h P Z N a n d R S N nav iga -Chapter 1. Introduction 5 t ion . O u r use of R S N was also m o t i v a t e d b y this technique's s imi lar i ty to P Z N in terms of user interact ion . O u r i m p l e m e n t a t i o n of r u b b e r sheet nav igat ion was based on the framework of Slack et a l . [47], wh ich , l ike P Z N interfaces, provides mul t ip l e levels of magni f i ca t ion a n d a n interact ion m o d e l based o n rec tangular selection areas. W e were also interested i n c o m p a r i n g R S N to P Z N because of the lack of e m p i r i c a l eva luat ion of the former i n the l i terature . A deta i led discus-sion of re lated work i n i m p l e m e n t a t i o n a n d evaluat ion issues for b o t h nav iga t ion techniques as well as guaranteed v i s ib i l i ty can be found in C h a p t e r 2. T h e task used i n o u r s t u d y is a topo log ica l nav iga t ion task m o t i v a t e d b y the requirements of phylogenet ic biologists, w h o require sophis t icated v i sual izat ion tools to s u p p o r t their work. O u r discussions w i t h phylogenet ic biologists lead us to develop a set of c o m p o u n d topo log ica l tasks re lated to their needs, of w h i c h we t h e n selected a navigat ion- intens ive task sui table for o u r compar i son . T h e dataset used in o u r s tudy is a large tree dataset also der ived f r o m phylogenet ic biology. W e deve loped abstract versions of the task a n d the dataset in order to al low us to p e r f o r m a quant i ta t ive s t u d y w i t h non-expert users. F u r t h e r details a b o u t the task a n d dataset used i n o u r studies can be found i n C h a p t e r 3. W e c o n d u c t e d two exper iments , each invo lv ing 40 subjects a n d the same task a n d dataset . E a c h of the exper iments also used the same four interfaces, represent ing a l l combina t ions of P Z N a n d R S N w i t h a n d wi thout overviews, a l though the design of the interfaces was refined between the studies. T h e exper iments were des igned to measure per formance b y record ing the c o m p l e t i o n t imes a n d n u m b e r of nav iga t ion act ions requ ired to p e r f o r m the task w i t h each exper imenta l interface. W e also gathered d a t a o n self -reported measures such as perceived m e n t a l a n d phys ica l effort, ease of nav iga t ion , a n d ease of use. O u r first s tudy, discussed i n C h a p t e r 4, represents the first eva luat ion of the effect of level of context on per formance w i t h different nav igat ion techniques. T h e results of this s t u d y show that level of context d i d not have a significant i m -pact on per formance for any of the interfaces. However , the s tudy's results were s trongly affected b y differences in nav igat ion strategies deve loped b y subjects to deal w i t h the perce ived c o m p l e x i t y of the interfaces. T h i s effect m o t i v a t e d us C h a p t e r 1. Introduction 6 to s impl i fy the interfaces a n d develop deta i led t r a i n i n g strategies for our second study. O u r second study, descr ibed i n C h a p t e r 5, is the first to evaluate the ef-fects of nav iga t ion technique a n d the presence of a n overview as or thogona l factors. W e found that subjects per formed signif icantly faster us ing P Z N t h a n R S N regardless of whether a n overview was present. A d d i t i o n a l l y , subjects re-qu ired fewer nav iga t ion interact ions a n d repor ted a lower m e n t a l effort w i t h P Z N whi le c o m p l e t i n g the task. O u r results also indicate t h a t overviews d i d not appear to improve per formance , but were st i l l perceived as beneficial . W e discuss the impl ica t ions of these results, i n c l u d i n g the re lat ionship between nav-igat ion technique a n d task, a n d m a k e recommendat ions for future evaluat ions of R S N . C h a p t e r 6 outl ines the l imi ta t ions of o u r exper iments , lists some possi-bil it ies for future work s t e m m i n g f rom our studies, a n d concludes this thesis. T h e research project that c o m p r i s e d the two studies discussed i n this thesis was c o n d u c t e d b y the a u t h o r j o i n t l y w i t h A d a m B o d n a r . W i t h i n this project , the a u t h o r was responsible for invest igat ing the effects o f nav iga t ion technique, while B o d n a r invest igated the effects of presence or absence of overview. T h i s thesis therefore emphasizes the aspects of the studies re lated to nav igat ion tech-niques, whi le those aspects related to presence of overview are presented i n greater deta i l in B o d n a r ' s master's thesis [12]. A s a result , Sect ions 4.3, 4.4, 4.5, a n d 4.8 of C h a p t e r 4 a n d Sections 5.3, 5.4, 5.5, a n d 5.8 of C h a p t e r 5 are jo int ly authored wi th B o d n a r , while C h a p t e r 3 is based on a version jo in t ly a u t h o r e d w i t h h i m . S u b s t a n t i a l port ions of this thesis also appear i n a paper pub l i shed i n the proceedings of the 2006 S I G C H I conference on H u m a n fac-tors in c o m p u t i n g systems, w h i c h was jo in t ly a u t h o r e d w i t h B o d n a r , J o a n n a M c G r e n e r e , F r a n c o i s G u i m b r e t i e r e , a n d T a m a r a M u n z n e r [38]. T o s u m m a r i z e , the re mainder of this thesis is organized as follows. C h a p t e r 2 discusses related work a n d provides a b a c k g r o u n d for o u r studies. C h a p t e r 3 discusses the task a n d dataset used in b o t h our studies. T h e design a n d results of each exper iment are deta i led a n d discussed in C h a p t e r s 4 a n d 5. F i n a l l y , C h a p t e r 6 suggests direct ions for future work a n d concludes this thesis. Chapter 2 7 Related Work This chapter e x a m i n e s design issues in p a n a n d z o o m a n d r u b b e r sheet n a v i g a -t ion interfaces, discusses the concept of g u a r a n t e e d v i s ib i l i ty a n d its implemen-tations in greater de ta i l , a n d presents a n overv iew of relevant work in e m p i r i c a l eva luat ion of n a v i g a t i o n techniques a n d tree v i sua l i za t ions . 2.1 Pan and Zoom Navigation (PZN) In the context of i n f o r m a t i o n v i s u a l i z a t i o n , n a v i g a t i o n can be def ined as the traversa l of a n i n f o r m a t i o n s t r u c t u r e b y se lect ing p a r t s of the current v i ew of the s t ruc ture [19]. T h e d o m i n a n t m e t a p h o r for n a v i g a t i o n in i n f o r m a t i o n v i s u a l i z a -t ion t o d a y is p a n a n d z o o m n a v i g a t i o n , w h i c h has been used in a var ie ty o f e x p e r i m e n t a l sys tems (see survey in H o r n b a e k . B e d e r s o n . a n d P l a i s a n t [24]). as well as a n u m b e r of c o m m e r c i a l a p p l i c a t i o n s [1 ; 20]. P a n a n d z o o m n a v i g a t i o n combines two classes o f n a v i g a t i o n techniques: p a n n i n g , wh ich a l lows users to change the v i s ib le reg ion of the datase t t h r o u g h h o r i z o n t a l a n d v e r t i c a l t rans-lat ions, a n d z o o m i n g , w h i c h changes the scale at w h i c h the 'dataset is v iewed to al low users to v iew regions of interest at greater or lesser re so lu t ion . In this thesis, p a n a n d z o o m n a v i g a t i o n is used to m e a n the c o m b i n a t i o n of these two classes of r i g i d t w o - d i m e n s i o n a l t r a n s f o r m a t i o n s , as opposed to n a v i g a t i o n m e t h o d s that a d a p t these techniques for use in d i s tor t ion -based interfaces [3. 4]. A l t h o u g h p a n n i n g is one of the most bas ic techniques for n a v i g a t i n g d a t a , there have been re lat ive ly few a t t e m p t s to descr ibe a n d c o m p a r e different p a n -n i n g var iat ions . J o h n s o n [27] prov ides a survey of this l i terature a n d descr ibes the fol lowing p a n n i n g m e t a p h o r s , i l lus trated in F i g u r e 2 .1: Chapter 2. Related Work 8 Push Background Push Camera Touch Edge Camera F i g u r e 2.1: P a n n i n g m e t a p h o r s e x a m i n e d b y J o h n s o n [27]. 1. P u s h b a c k g r o u n d : T o view an off-screen reg ion , users p a n in the d irec-t ion oppos i t e to where this region lies, as t h o u g h m o v i n g the b a c k g r o u n d of the v i s u a l i z a t i o n . 2. P u s h camera: T o v iew an off-screen reg ion , users p a n in the d i r e c t i o n this region lies, as t h o u g h m a n i p u l a t i n g a c a m e r a t h a t is v i e w i n g the dataset . 3. T o u c h edge camera: T o view an off-screen reg ion , users t o u c h the edge of the v iew in the d i r e c t i o n this region lies. J o h n s o n c o m p a r e d these three metaphors in, a contro l l ed s t u d y p e r f o r m e d o n a touch display . Resu l t s showed that P u s h B a c k g r o u n d p a n n i n g was super ior to the other k i nds of p a n n i n g in terms of b o t h p e r f o r m a n c e a n d user preference. C o n s i s t e n t w i t h this f inding , we use the P u s h B a c k g r o u n d m e t a p h o r in o u r i m p l e m e n t a t i o n of p a n a n d z o o m nav iga t ion . A variety o f z o o m i n g n a v i g a t i o n approaches have been descr ibed in the in-format ion v i s u a l i z a t i o n l i t era ture . T h e s e p r i m a r i l y differ in terms o f how the scale of objects in the dataset is m a n i p u l a t e d as users p e r f o r m z o o m in a n d z o o m out act ions . H o r n b a e k et a l . [24] descr ibe the fo l lowing approaches to i m p l e m e n t i n g z o o m i n g in t e r m s of scale changes: 1. G e o m e t r i c z o o m i n g : T h e most, c o m m o n a p p r o a c h , where the apparent size of objects increases l inearly when users z o o m into an area o f interest, a n d decreases at the s a m e rate when they z o o m out . Chapter 2. Related Work 9 2. S e m a n t i c z o o m i n g : I n t r oduced in P e r l i n a n d F o x ' s P a d sys tem [40], th is a p p r o a c h reveals new features i n add i t i on to inc reas ing the size of ex is t ing ones as users zoom in . T h i s techn ique is p a r t i c u l a r l y su i t ab le for m a p da tase ts , where users are of ten interested in different k i n d s of features d e p e n d i n g on the scale a t w h i c h they are v i e w i n g the d a t a . 3. C o n s t a n t dens i ty z o o m i n g : T h i s a p p r o a c h , first used by Woodru f f , L a n d a y . and Stonebreaker [50], uses a more comp lex re la t i onsh ip between scale a n d appearance, where a cons tan t numbe r of ob jec ts is v is ib le re-gard less of zoom ing ac t ions . T o da te , no s t u d y has c o m p a r e d these three approaches to de te rm ine whether semant ic o r cons tan t dens i ty z o o m i n g offer benef i ts compa red to s imp le geomet-r ic z o o m i n g . T h e imp lemen ta t i on o f pan and z o o m nav iga t i on d iscussed in th is thesis therefore relies on geomet r i c z o o m i n g for cons is tency w i t h the ma jo r i t y of the sys tems documen ted in the l i te ra ture . T w o m a i n approaches for i m p l e m e n t i n g scale changes d u r i n g z o o m nav iga-t ion ac t ions have been desc r ibed . In j u m p z o o m i n g [40], changes of scale occur ins tan taneous ly , w i t h o u t in te rmed ia te steps, wh i le in a n i m a t e d z o o m -i n g [8], the t rans i t i on f rom the o ld to the new scale is s m o o t h l y an ima ted . A s tudy by B e d e r s o n and B o l t m a n [7] compa red these two me thods i n a topo logy reca l l task. A l t h o u g h no di f ference in comp le t i on t ime was found , resul ts i nd i -ca ted that, the users p roduced topology recons t ruc t ions of h igher qua l i t y w i t h an ima ted z o o m i n g . Based on th is resu l t , we use s m o o t h a n i m a t e d t rans i t ions d u r i n g z o o m nav iga t ion ac t ions in ou r imp lemen ta t i on of pan a n d z o o m . 2.2 R u b b e r S h e e t N a v i g a t i o n ( R S N ) A s p rev ious ly ment ioned , the m a j o r a l te rna t ive to pan and z o o m nav iga t ion is the focus - f -context app roach , f i rst i n t roduced by Spence a n d A p p e r l e y [48]. U n l i k e pan and zoom inter faces, w h i c h ei ther on ly present, users w i t h regions of cur ren t interest o r p rov ide con tex tua l i n f o rma t i on in a separa te overv iew Chapter 2. Related Work 10 w i n d o w , focus+context techniques integrate focus regions a n d context w i t h i n a s ingle v iew [11]. T h r o u g h this in tegrat ion , the focus+contex t a p p r o a c h a i m s to reduce the cogni t ive l oad required for users to m a i n t a i n a g lobal represen-t a t i o n o f the dataset a n d the ir n a v i g a t i o n a l h i s tory [33]. M o s t f ocus+contex t interfaces in the l i t erature integrate focus a n d context regions us ing d y n a m i c a l l y chosen dis tort ions . E x a m p l e s of such d i s t o r t i o n - b a s e d techniques i n c l u d e fish-eye v iews [18 ; 44]. h y p e r b o l i c geometry [31], n o n l i n e a r magn i f i ca t ion [29] ; a n d a n u m b e r of other a p p r o a c h e s [35]. O t h e r focus+context approaches that d o not re ly o n d i s tor t ion i n c l u d e aggregat ing context regions into g lyphs [14. 43] a n d s h o w i n g contextua l i n f o r m a t i o n t h r o u g h layers of lenses [10]. T h e e v a l u a t i o n discussed in this thesis is intended to be p r i m a r i l y re levant to the l i t erature on d i s tor t ion-based focus+contex t interfaces, a n d in the r e m a i n d e r of this thesis the t e r m "focus+context" is used to refer specif ical ly to these interfaces. T h e p a r t i c u l a r focus+context n a v i g a t i o n technique e x a m i n e d in this thesis is r u b b e r sheet n a v i g a t i o n , or ig ina l ly deve loped b y S a r k a r . S n i b b e . T v e r s k y . a n d Reiss [45]. T h e n a m e of this t echnique comes f rom its centra l m e t a p h o r of i n t e r a c t i n g wi th the datase t as t h o u g h it were la id ou t o n a r u b b e r sheet w i t h its b o r d e r s tacked d o w n . U s e r s can select a n d stretch or compress a r b i t r a r y areas of the r u b b e r sheet, wh i l e the rest of the r u b b e r sheet r e m a i n s vis ible , t h o u g h it m a y be compressed . T h i s approach has the a d v a n t a g e of preserv ing users' sense of locat ion in the datase t . which can easily be lost d u r i n g distort i on -based n a v i g a t i o n [30]. S a r k a r et a l . descr ibe two var iants o f r u b b e r sheet n a v i g a t i o n . W i t h o r thogona l s t r e t c h i n g , i l lus trated in F i g u r e 2.2. users are res tr ic ted to se lect ing vert ica l or h o r i z o n t a l slices of the dataset. areas , which are s t re tched out. w i t h o u t affecting the rest of the dataset. T h i s has the advantage of preserv-ing o r t h o g o n a l o r d e r i n g of points in the dataset . b u t suffers from d i s c o n t i n u i t y of scale at the b o u n d a r y between focus a n d context areas. P o l y g o n a l s t r e t c h -ing , s h o w n in F i g u r e 2.3. enables users to select a r b i t r a r y po lygons as areas of interest , and s m o o t h l y integrates s tre tched out focus areas w i th context re-gions in terms of scale. However , p o l y g o n a l s t re tch ing does not preserve datase t s y m m e t r y . Chapter 2. Related Work 11 F i g u r e 2.2: R u b b e r sheet n a v i g a t i o n wi th o r t h o g o n a l stretching as descr ibed in S a r k a r et a l . [45]. T h e symmetry o f the dataset is preserved at the expense of d i scont inui t ies of scale. F i g u r e 2.3: R u b b e r sheet n a v i g a t i o n wi th p o l y g o n a l s tre tch ing u s i n g a rec tan-gu lar selection area as descr ibed in Sarkar et a l . [45]. F o c u s a n d context, areas are s m o o t h l y in tegrated , b u t dataset s y m m e t r y is not preserved. Chapter 2. Related Work 12 T h e p a r t i c u l a r version of r u b b e r sheet nav igat ion discussed i n this thesis is based o n t h a t i m p l e m e n t e d i n the T r e e J u x t a p o s e r v i sua l i za t ion t o o l devel-o p e d b y M u n z n e r , G u i m b r e t i e r e , T a z i r a n , Z h a n g , a n d Z h o u [37] (see F i g u r e 2.4). T h e f o r m of r u b b e r sheet nav igat ion used i n T r e e J u x t a p o s e r represents a m i d d l e g r o u n d between o r t h o g o n a l a n d p o l y g o n a l s tre tching as descr ibed b y Sarkar et a l . [45], since it provides s m o o t h in tegrat ion of focus a n d context regions whi le preserv ing s y m m e t r y of tree s tructures . S i m i l a r l y to m a n y imple -mentat ions of p a n a n d z o o m nav igat ion , T r e e J u x t a p o s e r uses rect i l inear regions as selection areas, a n d provides a n i m a t e d transi t ions to m a i n t a i n user context d u r i n g s tre tching . T h e s e s imi lar i t ies enable a n easier c o m p a r i s o n between the two types of nav igat ion techniques. File Find Tools Help SKneliai arthlacanihuci'phala |li«t.rfcv«<t a m n i o n • Uiaryx hulot rntrui [_ acanthopterygit osteiththyes u t e i duloi M M U M lift y I itoi.iirs myriprislis acaaipaaa oltirkthys !ar|«(entraa -u i a t l i l i i i ret Mat i rat h i c h i hy o i d e • hppUftethai p aratratki c ht byt cvprinodontrTormes apticaeUafcari • broldel •,t™,e"l-M" Uy liailllrf annirllldar lialil f 1|aatarkyathai lat it m i c a aaiiiMaalit , 1 didp*y<hr alii pUiybrlnnr argalas lovii i i.ihuhthodt > afflaij aalaiyx u ' u t i t i i l i f l * j a» ia apercalajls haloiratiui i autfimaolatus •alaccMrai alaacMa balatcartrai inaau •alaccBiTis mttki haleceatrai rafai halaccairai ipinaii vstiaus halaceatras lartafM aryripriitis aduita -•aiyiiam*il ausuills Hyripriitis | i .obu. Myripriftil mai«|.in •ytlarlstls raadalii Myilprixtis vial acta Myriprinis wiiiatas acaaiphan aurolinr atti a im hi hyi plbvaxti -»iaraaccniraa alkie Ma idisairnliaa iainifciaai aaamalept kaptoatran anMmm nidi iraaialai baplaftclbai eloagaim Mi itrai kkhthyl iidilli ihinnileimri miii.lrpiUuliii iiiaiadl at aanirlla pakir* F i g u r e 2.4: R u b b e r sheet nav igat ion a n d guaranteed v i s ib i l i ty i n T r e e J u x t a -poser [37]. F o c u s a n d context regions are s m o o t h l y in tegrated whi le preserv ing the s y m m e t r y of the tree's topology. Chapter 2. Related Work 13 2.3 Guaranteed Visibility B o t h p a n a n d z o o m a n d focus+context , n a v i g a t i o n techniques have been s h o w n to be effective for e x p l o r i n g datasets o f u p to several h u n d r e d i tems. H o w e v e r , w i t h larger datasets . b o t h classes of nav iga t ion techniques often encounter vis-ib i l i ty issues, where m a r k e d areas of interest to the user, such as l a n d m a r k s or search results , c a n n o t be seen. In p a n a n d z o o m interfaces, this m a y o c c u r be-cause areas of interest move off-screen due to n a v i g a t i o n act ions . In its e x t r e m e f o r m , the m o v i n g of areas of interest off-screen c a n resu l t ing in a p h e n o m e n o n referred to as d e s e r t f o g [28]. where the user is faced w i t h a view of the dataset. d e v o i d of n a v i g a t i o n a l cues. T h e m a r k e d areas m a y also become s i m p l y too smal l to be d i sp layed at the set scale a n d reso lut ion. In focus+context interfaces, ar-eas of interest, m a y be rendered invis ible due to the effects of d i s t o r t i o n , w h i c h c a n lead t h e m to be cu l led or aggregated w i t h n o n - m a r k e d areas. A c o m m o n so lut ion to v i s ib i l i ty issues is to augment the v i s u a l i z a t i o n wi th an overv iew w i n d o w , which can enable users to see m a r k e d areas outs ide the deta i l v iew. However , overview windows have a n u m b e r of d r a w b a c k s . T h e y take up screen real estate, m a y not p r o v i d e e n o u g h reso lut ion to ensure that m a r k e d areas in large datasets are v is ible to users, a n d d iv ide users 1 a t t ent ion [51]. T o address this issue, M u n z n e r et a l . [37] i n t r o d u c e the concept of g u a r a n -t e e d v i s i b i l i t y , the p r o p e r t y that m a r k e d areas o f the dataset. are g u a r a n t e e d to be visible regardless of datase t size or n a v i g a t i o n act ions taken by users. M u n z n e r et a l . differentiate between three cases to cons ider when g u a r a n t e e i n g v i s ib i l i ty of m a r k e d areas: 1. O f f - s c r e e n : A m a r k e d a r e a m a y move off-screen due to user n a v i g a t i o n act ions or restr ict ions o n avai lable screen real estate. 2. S u b - p i x e l : T h e d i m e n s i o n s o f a m a r k e d area m a y shr ink to less t h a n a p ixe l . T h i s s i tuat ion is p a r t i c u l a r l y l ikely to o c c u r when the n u m b e r of i tems in a dataset is larger than the n u m b e r o f pixels avai lable to the v i s u a l i z a t i o n . Chapter 2. Related Work 14 3. O c c l u s i o n : A m a r k e d a r e a c a n be o c c l u d e d by o ther p a r t s of the dataset . such as labels in a t w o - d i m e n s i o n a l l ayout or other i t ems in a three-d i m e n s i o n a l v i s u a l i z a t i o n . M u n z n e r et a l . [37] i m p l e m e n t e d all three types of g u a r a n t e e d v is ib i l i ty in c o n j u n c t i o n w i t h r u b b e r sheet n a v i g a t i o n in the prev ious ly d iscussed T r e e J u x t a -poser sys tem, w h i c h serves as the basis for the r u b b e r sheet n a v i g a t i o n interfaces discussed in this thesis. G u a r a n t e e d v i s ib i l i ty has also been i m p l e m e n t e d in sev-eral p a n a n d z o o m interfaces, i n c l u d i n g C i t y L i g h t s [51] a n d H a l o [6]. i l lus trated in F i g u r e s 2.5 a n d 2.6, respect ively . B a s e d on the J a z z p a n a n d z o o m inter-face toolkit, [9], C i t y L i g h t s ind icates the d irec t ion of off-screen m a r k e d areas v ia c o m p a c t i n d i c a t o r s integrated a l o n g v iew borders . H a l o [6] deve lops the concept o f C i t y L i g h t s further to p r o v i d e a n i n d i c a t i o n of b o t h d i r e c t i o n a n d dis tance to off-screen m a r k e d areas in the context, of smal l - screen devices. M a r k e d areas are s u r r o u n d e d w i t h r ings that are jus t large e n o u g h to reach into the b o r d e r regions of the v i s u a l i z a t i o n . T h e r ings enable users to d i scern the a p p r o x i m a t e locat ion of the m a r k s , whi le us ing a re la t ive ly sma l l p r o p o r t i o n of the avai lable screen real estate. In a, contro l led e x p e r i m e n t , H a l o was found to i m p r o v e p e r f o r m a n c e on a n a v i g a t i o n task c o m p a r e d to a n arrow-based technique that , s imi lar ly to C i t y L i g h t s , o n l y ind icated d i r e c t i o n to off-screen m a r k e d areas. T h i s result m o -t ivated o u r use of Halo - l ike arcs to prov ide off-screen g u a r a n t e e d v is ib i l i ty in the p a n a n d z o o m nav iga t ion interfaces e x a m i n e d in o u r s tudies . 2.4 Evaluation of Navigation Techniques T o date , no e m p i r i c a l e v a l u a t i o n of r u b b e r sheet n a v i g a t i o n has appeared in the i n f o r m a t i o n v i sua l i za t ion l i terature . However , a n u m b e r o f eva luat ions have c o m p a r e d o ther focus+context n a v i g a t i o n techniques to p a n a n d z o o m inter-faces. O n e of the first such eva luat ions was presented by Schaf ler et a l . [46]. who c o m p a r e d a z o o m i n g interface to a fisheyc interface for p e r f o r m i n g a nav-igat ion a n d r o u t i n g task in a h i erarch ica l l y c lus tered network datase t . Resul t s Chapter 2. Related Work IS F i g u r e 2.6: G u a r a n t e e d v i s ib i l i ty i n H a l o [6]. D i r e c t i o n a n d d i s tance to off-screen m a r k e d areas are i n d i c a t e d by arcs that represent v i s ib le p o r t i o n s of r ings d r a w n a r o u n d the m a r k e d areas . Chapter 2. Related Work 16 showed that the fisheye interface resulted in s igni f icant ly lower c o m p l e t i o n t imes a n d was preferred by subjects . G u t w i n a n d S k o p i k [22] c o m p a r e d fisheye interfaces us ing three different dis-tor t ion models to two p a n n i n g interfaces w i t h overv iews for p e r f o r m i n g a large s teering task. Resu l t s i n d i c a t e d that all the fisheye interfaces enabled faster task c o m p l e t i o n t h a n ei ther o f the p a n n i n g interfaces, a l t h o u g h this f inding was p a r t i a l l y e x p l a i n e d by the i m p l e m e n t a t i o n s h o r t c o m i n g s of one of the p a n n i n g interfaces. G u t w i n followed u p this s tudy w i t h a n o t h e r invest igat ion [21], w h i c h c o m p a r e d fisheye, p a n n i n g , a n d two-level z o o m interfaces for e d i t i n g , web nav-igat ion , a n d m o n i t o r i n g tasks. T h a t s tudy found t h a t the fisheye interface was s ignif icantly faster for the web navigat ion. task , b u t the z o o m interface p e r f o r m e d better on the m o n i t o r i n g task a n d was strongly pre ferred by subjects . H o r n b a e k a n d F r o k j a e r [25] c o m p a r e d p a n n i n g interfaces w i th a n d without, overviews to a fisheye interface for read ing e l ec tron ic d o c u m e n t s . T h a t s tudy found that subjects read d o c u m e n t s faster w h e n u s i n g the fisheye interface t h a n when us ing e i ther of the p a n n i n g interfaces. H o w e v e r , the p a n n i n g interface wi th an overview p r o v i d e d better c o m p r e h e n s i o n a n d was preferred by subjects . B a u d i s c h , L e e , a n d H a n n a [5] p e r f o r m e d a c o m p a r a t i v e s t u d y o f three inter-faces s imi lar to those e x a m i n e d by H o r n b a e k a n d F r o k j a e r . B o t h the fisheye interface a n d the p a n n i n g interface wi th an overv iew p r o v i d e d guaranteed vis-ib i l i ty of m a r k e d areas, whi le the p a n n i n g interface w i thout an overview d i d not . A l t h o u g h results for b o t h per formance a n d subjec t preference were h igh ly dependent o n the task, the two interfaces w i th g u a r a n t e e d v i s ib i l i ty were faster t h a n the c o m p a r i s o n interface for most tasks a n d were preferred by all subjects , a finding that m o t i v a t e d o u r use of guaranteed v i s ib i l i ty in the interfaces ex-a m i n e d in o u r exper iment . B a u d i s c h et a l . also suggested that future studies invest igate the potent ia l benefits of c o m b i n i n g focus+context . n a v i g a t i o n w i t h overviews, which corresponds to one of the interfaces e x a m i n e d in o u r e x p e r i -ment . T h e eva luat ion presented in this thesis a t t e m p t s to overcome a n u m b e r of l imi ta t ions of scope shared by these studies. F i r s t , the n o n - d i s t o r t i o n inter-Chapter 2. Related Work 17 faces used in these studies p r o v i d e d either p a n n i n g or z o o m i n g capabi l i t ies , bu t not b o t h . P a n n i n g a n d z o o m i n g are increas ing ly used together in b o t h re-search a n d c o m m e r c i a l interfaces, a n d some eva luat ions of interfaces c o m b i n i n g these two techniques have a p p e a r e d in the l i terature , most n o t a b l y H o r n b a e k et a l . ; s c o m p a r i s o n of p a n a n d z o o m interfaces w i th a n d w i t h o u t overviews [24]. However , c o m b i n e d p a n a n d z o o m interfaces have not to date been e m p i r i c a l l y c o m p a r e d to focus+context interfaces . Second , a l l the a b o v e - m e n t i o n e d studies used v a r i a t i o n s on the fisheye interface p a r a d i g m . T h e work presented in this thesis seeks to e x p a n d the l i t e ra ture on eva luat ion of f ocus+contex t interfaces to inc lude r u b b e r sheet n a v i g a t i o n , w h i c h represents one p o t e n t i a l a l ternat ive to fisheye views. T h i r d , a l t h o u g h the fisheye interfaces used in G u t w i n a n d S k o p i k ; s s t u d y [22] p r o v i d e d context in different ways d e p e n d i n g on their dis-tor t ion mode l s , neither this s t u d y n o r any of the others discussed in this sect ion e x a m i n e d level of context as a factor. F o u r t h , none of the s tudies invest igated a u g m e n t i n g focus+context interfaces w i th overviews to d e t e r m i n e whether this would c o m p e n s a t e for the d r a w b a c k s o f d i s t o r t i o n , a l i m i t a t i o n discussed in m o r e de ta i l in B o d n a r s thesis [12]. F i n a l l y , a p a r t f rom the s t u d y p e r f o r m e d by Schaffer et a l . [46], the eva luat ions discussed above d i d not rely on tasks or datasets der ived from r e a l - w o r l d appl i ca t ions , a l i m i t a t i o n addressed in o u r s tudy t h r o u g h the choice o f a n ecological ly va l id task a n d datase t inspired by the needs of phylogenet ic b io logis ts . 2.5 Evaluation of Tree Visualizations T h e work descr ibed in this thesis is re lated to a n o t h e r b r a n c h of the l i terature on e v a l u a t i o n of in format ion v i sua l i za t ions , n a m e l y user s tudies of tree v i sua l -izat ion techniques . In p a r t i c u l a r , two recent studies have p e r f o r m e d contro l led exper iment s i n v o l v i n g interfaces for v i sua l i z ing large tree datasets . K o b s a [30] c o m p a r e d five tree v i sua l i za t ion interfaces, as well as W i n d o w s E x p l o r e r , w h i c h was used as a basel ine for c o m p a r i s o n . K o b s a : s s tudy used a h i erarch ica l tree dataset of m o r e t h a n 5,700 nodes a m i a var ie ty of tasks related Chapter 2. Related Work 18 to b o t h datase t topo logy a n d i t e m a t t r ibute s . W i n d o w s E x p l o r e r o u t p e r f o r m e d the c o m p a r i s o n interfaces a n d was also preferred by most subjec ts , h i g h l i g h t i n g the dif f iculty o f c o m p a r i n g a n interface wi t l i w h i c h users have experience to those they are e n c o u n t e r i n g for the first t ime. T h e c o m p a r i s o n was confounded by the fact t h a t some interfaces were m i s s i n g funct ional i ty r e q u i r e d to comple te some of the tasks. P l a i s a n t . G r o s j e a n . a n d B e d e r s o n [43] c o m p a r e d their S p a c e T r e e tool , w h i c h used a n o n - d i s t o r t i o n - b a s e d focus+contex t interface, to W i n d o w s E x p l o r e r a n d a h y p e r b o l i c tree browser based on that deve loped by L a m p i n g . R a o . a n d P i r o l l i [31]. T h e e x p e r i m e n t used a large tree datase t o f more t h a n 7,000 nodes a n d a variety o f search a n d topo log ica l tasks. T h e results of the s t u d y were m i x e d , re-vea l ing t h a t S p a c e T r e e p e r f o r m e d s igni f icant ly faster for some topo log ica l tasks, but not for others , w i th no significant differences in terms of subjec t preference. A c o m m o n l i m i t a t i o n to b o t h these s tudies is that the interfaces e x a m i n e d used w i d e l y different m e t h o d s of da ta presenta t ion a n d in terac t ion , m a k i n g the ir results diff icult to interpret . T h e e x p e r i m e n t s descr ibed in this thesis a i m to overcome this issue by c o m p a r i n g interfaces that share v i s u a l presentat ion a n d in terac t ion m e t a p h o r s a n d differ on ly in t erms of nav iga t ion technique . Chapter 3 Task and Dataset 19 In order to lend ecological va l id i ty to o u r e x p e r i m e n t , we der ived the task a n d dataset used in it f rom the d o m a i n of phy logene t i c biology. P h y l o g e n e t i c b i -ologists m o d e l e v o l u t i o n a r y re lat ionships as h i erarch ica l trees in a n effort to i m p r o v e the ir u n d e r s t a n d i n g of how different organ i sms evolve a n d co-evolve . T h e recent f lood of m o l e c u l a r da ta o b t a i n e d f r o m D N A a n d p r o t e i n s e q u e n c i n g has enab led the c o n s t r u c t i o n of phy logenet i c trees of ever- increas ing size. T o -day, some g r o u p s of phy logene t i c biologists have c o n s t r u c t e d trees c o n t a i n i n g t h o u s a n d s of nodes , a n d m a n y hope soon to be able to recons truc t the c o m -plete T r e e o f L i fe , e s t i m a t e d to conta in over ten mi l l i on species [37]. H o w e v e r , a recent survey [15] po in t s out that progress has been h a m p e r e d by a lack of tools s u p p o r t i n g e x p l o r a t i o n , v i sua l inspec t ion , a n d s t r u c t u r a l c o m p a r i s o n in such large datasets . T h i s c h a p t e r d o c u m e n t s the choice of o u r task a n d datase t based on the requ irements of this d o m a i n . 3.1 Task T o ga in an u n d e r s t a n d i n g of the tasks i n v o l v e d in phylogenet ic analys is of large tree datasets u s i n g i n f o r m a t i o n v i sua l i za t ion tools , we c o n d u c t e d interviews w i t h ten phy logenet i c biologists f rom univers i t ies in C a n a d a a n d the U n i t e d States . W e learned that p h y l o g e n e t i c biologists use interact ive v i sua l i za t ions of large e v o l u t i o n a r y trees to ga in a deeper u n d e r s t a n d i n g of the re la t ionships between a n d w i th in groups of o r g a n i s m s . T h r o u g h the process of topo log ica l ana lys i s , these researchers a i m to d e t e r m i n e how- species have evolved a n d co-evolved , a n d how character i s t i cs are passed from one species to the next in a n e v o l u t i o n a r y Chapter 3. Task and Dataset 20 lineage. B a s e d o n o u r discussions, we developed a set of four tasks, descr ibed a n d i l lus trated below, w h i c h were representative of the tree- topologica l tasks per-formed b y phylogenet ic biologists, b u t d i d not require spec ia l ized knowledge of evo lu t ionary trees. W e then val idated these tasks w i t h several of the biologists we h a d prev ious ly interviewed to ensure the tasks' ecological val idity . E a c h of the tasks was c o m p o s e d of several low-level tasks such as f ind, identify, a n d compare , as descr ibed i n the v i sua l task t a x o n o m y of W e h r e n d a n d L e w i s [49]. In the i l lustrat ions below, a co lored n o d e represents a species, whereas a co lored subtree represents a related g r o u p of species. 3.1.1 T a s k 1: D e t e r m i n i n g the lowest c o m m o n a n c e s t o r In a phylogenet ic tree, the l o w e s t c o m m o n a n c e s t o r of two nodes is a n or-gan i sm t h a t is a n ancestor of b o t h the species i n quest ion, a n d that has the greatest d e p t h i n the tree, as i l lus trated in F i g u r e 3.1. D e t e r m i n i n g the lowest c o m m o n ancestor is an i m p o r t a n t task in p h y l o g e n e t i c t a x o n o m y , a b r a n c h of phylogenet ic bio logy concerned w i t h classifying species based o n phylogenet ic data . 1 r-M F i g u r e 3.1: T a s k 1: D e t e r m i n i n g the lowest c o m m o n ancestor. In this case, node A is the lowest c o m m o n ancestor of nodes B a n d C . Chapter 3. Task and Datase t 21 3.1.2 Task 2: Comparing the topological distances between nodes T o p o l o g i c a l d is tance in a tree is the n u m b e r of hops between two nodes , a n d is not the same as geometric distance, which m a y change w i t h n a v i g a t i o n , as i l lus trated i n F i g u r e 3.2. In a phylogenet ic tree, the topo log ica l d i s tance be-tween two nodes is indicat ive of the n u m b e r of e v o l u t i o n a r y steps between the species they represent. M e a s u r i n g a n d c o m p a r i n g topo log ica l distances is one of the p r i m a r y tasks for w h i c h phylogenet ic biologists require v i sua l i za t ions of evo lu t ionary trees. F i g u r e 3.2: T a s k 2: C o m p a r i n g the topolog ica l distances between nodes . In this case, node A is 2 topolog ica l hops f rom node B a n d 3 topo log ica l h o p s f r o m node C , m a k i n g node B topolog ica l ly closer. 3.1.3 Task 3: Determining whether two subtrees are adjacent In a tree, two subtrees are adjacent if no other n o d e is between t h e m , as i l lus-t ra ted i n F i g u r e 3.3. In phylogenet ic biology, this task represents d e t e r m i n i n g whether the groups of species represented b y the subtrees are s i s t e r g r o u p s , or groups of organisms w h o are most closely re lated to one another i n t erms of their e v o l u t i o n a r y h i s tory (for instance, great apes a n d monkeys ) . C h a p t e r 3. Task and Dataset 22 F i g u r e 3.3: T a s k 3: D e t e r m i n i n g whether two subtrees are adjacent . In this case, the subtrees labeled A a n d B are not adjacent. 3.1.4 T a s k 4: D e t e r m i n i n g w h e t h e r a sub t r ee c o n t a i n s u n m a r k e d nodes In a phylogenet ic tree, m a r k e d nodes m a y indicate the presence of a u n i q u e feature or character . T h e presence of uncolored nodes or subtrees in a m a r k e d subtree, s h o w n in F i g u r e 3.4, m a y therefore indicate a characte r reversa l , a n event caus ing the loss of a character formerly present in a n evo lu t ionary l ine (for example , the loss of a ta i l in great apes a n d h u m a n s ) . 1 • 1 1 A i t 1 i i — * Z T F i g u r e 3.4: T a s k 4: D e t e r m i n i n g whether a subtree contains u n m a r k e d nodes. In this case, the subtree labe led A contains a n u n m a r k e d node, B . Chapter 3. Task and Dataset 23 A f t e r the deve lopment o f the set of tasks descr ibed above , we dec ided to focus o u r s tudy on a s ingle task in order to l imi t the effect of task as a factor . T a s k 2 was selected for fur ther invest igat ion due to its re la t ive complexi ty , h i g h i m p o r t a n c e to phy logenet i c analys is , a n d the fact that it w o u l d require subjec t s to p e r f o r m mul t ip le n a v i g a t i o n act ions a l o n g wel l -defined pa ths , thus r e d u c i n g p e r f o r m a n c e variabi l i ty . T a s k instances were assessed in p i lo t studies to ensure t h a t they were iso-m o r p h i c in difficulty. In p a r t i c u l a r , topo log ica l d is tances between nodes a l w a y s fell in a range of 7 to 10, a n d c o u l d not be d e t e r m i n e d w i t h o u t in teract ing w i t h the interface for any o f the task instances. A l s o , co lored nodes were not l oca ted in close p r o x i m i t y to each o ther in order to ensure that at least one in terac t ion h a d to be per formed to d e t e r m i n e each topo log ica l d i s tance . 3.2 Dataset T h e dataset used for in i t i a l p i l o t ing for o u r s t u d y was the animaliaA da tase t from the 2003 Infovis C o n t e s t [42]. a phylogenet ic tree o f a p p r o x i m a t e l y 190,000 nodes represent ing a h y p o t h e s i s about the evo lut ion of o r g a n i s m s in the k i n g d o m Animalia. Init ial p i lot results suggested that this dataset. was not an o p t i m a l choice for o u r exper iment . Its topo logy was not sufficiently deep to require s u b -jects to p e r f o r m a large amount, of n a v i g a t i o n , while its size necessitated s tar t t imes of u p to 45 seconds for our v i sua l i za t ion tools . F o r this reason, subse-quent, p i l o t i n g and the f o r m a l exper iment used the phylogenyMatchesTaxonomy dataset , a b i n a r y tree cons i s t ing of 5,918 nodes , which also represents e v o l u -t i o n a r y re lat ionships be tween species the k i n g d o m Animalia. T h i s dataset. was used courtesy of D a v i d Hi] 1 is of the U n i v e r s i t y o f T e x a s , a n d is avai lable f rom the O l d u v a i project, websi te [39]. T h i s dataset al lowed for complex topo log i -cal c o m p a r i s o n s r e q u i r i n g a signif icant a m o u n t of n a v i g a t i o n while r e d u c i n g the start t imes for our tools to u n d e r 5 seconds. A l t h o u g h we h a d o r i g i n a l l y assumed that, node labels were important to the way biologists in teract w i t h phylogenet ic tree v i sua l i za t ions , o u r discuss ions Chapter 3. Task and Dataset 24 w i t h b io log is t s revealed that the ir t y p i c a l use o f e v o l u t i o n a r y trees involved very l i tt le labe l read ing . W e therefore r e m o v e d n o d e labels f r o m the dataset for the p u r p o s e o f o u r studies. U s i n g no labels enab led us to a v o i d unnecessary n o d e o c c l u s i o n a n d potent ia l c o n f o u n d i n g of e x p e r i m e n t a l resul ts b y subjects ' p r i o r knowledge of e v o l u t i o n a r y re la t ionships between species. Chapter 4 25 Study 1 T h e goal of S t u d y 1 was to e x a m i n e the effect of level of context on perfor-m a n c e in interfaces w i t h p a n a n d z o o m a n d r u b b e r sheet n a v i g a t i o n . T h e s t u d y invo lved four different interfaces, r epresen t ing all c o m b i n a t i o n s of the two nav-igat ion techniques w i t h a n d w i t h o u t a n overview, as i l lus tra ted in T a b l e 4.1. S u b j e c t s used these interfaces wi th v a r y i n g levels of context to solve a t o p o l o g -ical task in a large tree dataset . T h e level of context t h a t resul ted in the best p e r f o r m a n c e for each interface was t h e n used to c o m p a r e the p e r f o r m a n c e of the four interfaces in S t u d y 2. d e s c r i b e d in C h a p t e r 5. T h i s c h a p t e r descr ibes the s tudy, presents the results re lated to the effects of n a v i g a t i o n technique , a n d discusses their i m p l i c a t i o n s a n d the w a y they affected S t u d y 2. T h e results re lated to the effects of presence or absence of overview a n d their i m p l i c a t i o n s are discussed in de ta i l in B o d n a r s thesis [12]. ^ N a v i g a t i o n O v e r v i e w — R u b b e r S h e e t P a n a n d Z o o m N o t p r e s e n t R S N - N o O V P Z N - N o O V P r e s e n t R S N + O V P Z N + O V T a b l e 4.1: Interfaces represent ing all c o m b i n a t i o n s of the n a v i g a t i o n a n d pres-ence o f overv iew factors . A l l four interfaces were used in b o t h S t u d y 1 a n d S t u d y 2. Chapter 4. Study 1 26 4.1 Hypotheses O u r p r i m a r y h y p o t h e s i s for t ins s t u d y was that p e r f o r m a n c e in each interface w o u l d vary wi th the level of context a c c o r d i n g to a U - s h a p e d curve , w i t h v e r y low a n d very h igh context levels r e su l t ing in p o o r p e r f o r m a n c e c o m p a r e d to v a l -ues between these two extremes . W e expec ted that low levels of context w o u l d not prov ide sufficient reso lut ion to enable users to o b t a i n contex tua l cues, whi le h i g h levels of context w o u l d c o n s t r a i n the a m o u n t o f screen real estate ava i lab le to show features o f interest to users a n d therefore adverse ly i m p a c t n a v i g a t i o n . S i m i l a r l y , we expec ted that , in interfaces w i t h overv iews , sma l l overv iew sizes w o u l d provide insufficient reso lut ion , wh i l e large overv iew sizes w o u l d d e t r a c t f r o m the nav iga t ion w i t h i n the de ta i l v iews. T h e re su l t ing hypotheses are pre-sented below. H I : F o r b o t h p a n a n d z o o m a n d r u b b e r sheet n a v i g a t i o n , m e d i u m levels of context w i t h i n the deta i l v iew wi l l per for in be t ter t h a n ei ther h i g h or low levels of contex t . H 2 : In interfaces w i t h overviews, m e d i u m sized overv iews wi l l p e r f o r m bet ter than either s m a l l or large overviews . T h e values for s m a l l , m e d i u m , a n d large levels o f context a n d overv iew sizes for each interface were chosen based on results of p i lo t studies, a n d are l isted in Sec t ion 4.6. D u e to the v a r i a t i o n of levels of context a n d overview sizes w i t h i n interfaces in this s tudy , no specific hypotheses were deve loped for the effects of n a v i g a t i o n technique o r presence of overv iew on p e r f o r m a n c e , factors t h a t were invest igated in m o r e de ta i l in S t u d y 2. 4.2 Interfaces T h e interfaces e x a m i n e d in this s tudy are i l lus tra ted in F i g u r e s 4.1 to 4.4. In o r d e r to prov ide cons i s tent v i sua l representa t ion , d r a w i n g p e r f o r m a n c e , a n d Chapter 4. Study 1 27 F i g u r e 4.2: P Z N - N o O V interface used i n S t u d y 1. A z o o m act ion has filled the extent of the top view. A r c s inspired by H a l o [6] ind icate d irec t ion a n d dis tance to off-screen m a r k e d nodes. Chapter 4. Study 1 28 F i g u r e 4.3: R S N + O V interface used i n S t u d y 1. A z o o m act ion has s tretched the region s h o w n b y the field-of-view b o x i n the overview to fill the top focus region of the de ta i l v iew. F i g u r e 4.4: P Z N + O V interface used i n S t u d y 1. A z o o m act ion has filled the extent of the top deta i l v iew w i t h the region shown b y the field-of-view b o x in the overview. Chapter 4. Study 1 29 in terac t ion m o d e l , a l l interfaces were bu i l t on the P R I S A D software infrastruc-ture [47], based o n the T r e e J u x t a p o s e r scalable tree v i s u a l i z a t i o n a p p l i c a t i o n [37]. T h e use o f the P R 1 S A D in fras t ruc ture also he lped reduce i m p l e m e n t a -t ion t ime for the interfaces. W h i l e T r e e J u x t a p o s e r was in i t ia l ly deve loped as a F o c u s + C o n t e x t v i s u a l i z a t i o n too l u s i n g r u b b e r sheet n a v i g a t i o n , the inherent s imi lar i t ies between r u b b e r sheet a n d p a n a n d z o o m n a v i g a t i o n a l lowed us to ex tend its b e h a v i o u r to s u p p o r t c o n v e n t i o n a l p a n a n d z o o m i n t e r a c t i o n , as well v a r y i n g levels o f context , overv iews , a n d mul t ip l e foci . T h i s sect ion discusses the i m p l e m e n t a t i o n of each of these interface c o m p o n e n t s a n d then examines the interfaces themselves in de ta i l . 4.2.1 N a v i g a t i o n T h e or ig ina l T r e e J u x t a p o s e r a p p l i c a t i o n [37] used r u b b e r sheet-style expans ions a n d c o n t r a c t i o n s o f a r b i t r a r y rec t i l inear regions for n a v i g a t i o n , a n d i n c l u d e d a d -vanced features such as l inked n a v i g a t i o n between m u l t i p l e trees. N a v i g a t i o n in T r e e J u x t a p o s e r enab led users to select rec tangu lar regions us ing mouse drags, a n d resize their select ion box to a r b i t r a r y size. W e rep laced this s ty le o f nav iga -t ion w i t h a uni f ied set of n a v i g a t i o n act ions i m p l e m e n t e d across a l l interfaces. A l l in terac t ion o c c u r r e d t h o u g h m o u s e drags , a n d in o u r subsequent analysis , a discrete n a v i g a t i o n act ion refers to a single mouse d r a g . A l l t rans i t ions were s m o o t h l y a n i m a t e d across 20 frames to ensure fluid in terac t ion w i t h the inter-faces. In each interface, n a v i g a t i o n was contro l l ed us ing a two b u t t o n mouse w i t h a scrol l whee l , w i th z o o m in m a p p e d to the left mouse b u t t o n , p a n n i n g m a p p e d to the r ight mouse b u t t o n , a n d z o o m out m a p p e d to the scro l l wheel. E a c h interface also s u p p o r t e d a reset funct ion , w h i c h was m a p p e d to the R key. S i m i l a r to r u b b e r sheet n a v i g a t i o n as i m p l e m e n t e d in T r e e J u x t a p o s e r , our i m p l e m e n t a t i o n o f r u b b e r sheet n a v i g a t i o n al lowed users to select a rec tangular r region us ing m o u s e drags . However , in o u r i m p l e m e n t a t i o n , se lect ion boxes were a lways e x p a n d e d into focus areas of fixed size a n d aspect rat io . T h e d imens ions of the focus areas were fixed in o r d e r to ensure that the level of contex t remained Chapter 4. Study 1 30 constant regardless of user interact ions . 4.2.2 O v e r v i e w s a n d F o c i O v e r v i e w s w i t h m o v a b l e field o f v iew boxes were present in two o f the interfaces. F o r cons is tency between interfaces; the v iew d i m e n s i o n s in each interface were chosen to equal ize the total screen real estate across t h e m , wi th each interface a lways p r o v i d i n g a tota l of 600,000 pixels o f i n f o r m a t i o n . B a s e d on the guide-lines deve loped by A h l b e r g a n d S h n e i d e r m a n [2], we ensured t h a t all n a v i g a t i o n act ions were t i g h t l y coup led between the overv iew a n d detai l v i ew. F o r this s tudy, two foci were i m p l e m e n t e d in b o t h p a n a n d z o o m a n d r u b b e r sheet n a v i g a t i o n interfaces to al low users to s imul taneous ly v i ew a n d interact w i t h mul t ip l e non-adjacent regions of the dataset . In the r u b b e r sheet n a v i -gat ion interfaces, users cou ld select one o f two focus regions as the target for rect i l inear z o o m i n g act ions , a l l o w i n g t h e m to exp lore two n o n - a d j a c e n t regions of the dataset at different levels o f c o m p r e s s i o n . In the p a n a n d z o o m inter-faces, users c o u l d navigate in two separate v iews, a l l owing t h e m to exp lore two different regions of the dataset at different scales. T h e decis ion to i m p l e m e n t mul t ip le foci was m o t i v a t e d by the scenar io where subjects w o u l d be requ ired to navigate to features located between the two foci in the process of c o m p l e t i n g o u r task. F o r these instances, we expected r u b b e r sheet n a v i g a t i o n to benefit from the context region between the two foci , w h i c h w o u l d be e i ther not v is ible or on ly v is ible in the overview in interfaces u s i n g p a n a n d z o o m n a v i g a t i o n , as i l lustrated in F i g u r e 4.5. 4.2.3 G u a r a n t e e d V i s i b i l i t y a n d L e v e l s o f C o n t e x t G u a r a n t e e d v i s ib i l i ty of m a r k e d areas was p r o v i d e d in b o t h deta i l v iews a n d overviews for b o t h p a n and z o o m a n d r u b b e r sheet, nav iga t ion interfaces , a n d addressed the three types of g u a r a n t e e d v i s i b i l i t y discussed in Sect ion 2.3. In p a r t i c u l a r , s u b - p i x e l guaranteed v i s ib i l i ty in a l l interfaces was p r o v i d e d by the u n d e r l y i n g P R I S A D v i sua l i za t ion f ramework [47]. which ensured t h a t i tems of Chapter 4. Study 1 31 A if ~ T F i g u r e 4.5: M o t i v a t i n g scenario for use of two focus areas i n S t u d y 1. M a r k e d node A , located between the two focus areas i n each interface, is vis ible in b o t h views i n the R S N + O V interface (above) , b u t on ly i n the overview the P Z N + O V interface(below). R u b b e r sheet nav igat ion was therefore expected to p e r f o r m better t h a n p a n a n d z o o m nav iga t ion i n this scenario. Chapter 4. Study 1 32 interest, in all views were v i s i b l y m a r k e d even w h e n they were compressed to sub-p ixe l size. O c c l u s i o n of m a r k e d areas by other p a r t s o f the dataset was a v o i d e d by us ing a 2 D r a t h e r t h a n a 3 D spat ia l l ayout a n d r e m o v i n g labels f r o m o u r datase t . Off -screen guaranteed v i s ib i l i ty was i m p l e m e n t e d in the e x p e r i m e n t a l i n -terfaces in different ways d e p e n d i n g on n a v i g a t i o n t echn ique . In r u b b e r sheet interfaces; nav iga t ion was c o n s t r a i n e d so that i tems outs ide the focus areas were c o m p r e s s e d in context areas a l o n g the p e r i p h e r y of the view. In p a n a n d z o o m interfaces, d irec t ion to a n d d i s tance from off-screen m a r k e d areas were e n c o d e d us ing o p a q u e e l l ipt ical arcs s i m i l a r to those i m p l e m e n t e d in B a u d i s c h a n d R o s e n -ho l t z : s H a l o [6]. A s in r u b b e r sheet n a v i g a t i o n interfaces, these arcs a p p e a r e d in p e r i p h e r a l context areas a l o n g the edges of a v iew. H o w e v e r , these context areas were not expl ic i t ly v i s u a l l y de l imi t ed , the ir reso lut ion d i d not change, a n d their s h a p e was oval r a t h e r t h a n rec tangu lar . A l t h o u g h the degree of c o n t e x t u a l i n f o r m a t i o n prov ided by contex t areas in b o t h n a v i g a t i o n techniques varied w i t h user in terac t ion , we used the to ta l extent o f these areas as an a p p r o x i m a t i o n for the a m o u n t of context w i t h i n each interface. In a d d i t i o n to p e r i p h e r a l context areas , c o n t e x t u a l i n f o r m a t i o n was also p r o v i d e d b y overviews in those interfaces that c o n t a i n e d t h e m . F o r the p u r p o s e of v a r y i n g the level of contex t in this s tudy, we therefore d i s t ingu i shed between two poss ib le levels of context in each interface, i l lu s tra ted in F i g u r e s 4.6 a n d 4.7: 1. L e v e l of n a v i g a t i o n a l contex t : F r a c t i o n of size o f navigat ion-spec i f i c context areas C to the to ta l size of focus a n d context areas in the de ta i l v iew F+C. 2. L e v e l of overv iew contex t : F r a c t i o n of size of the overv iew 0 to to ta l size of all views O+F+C (0 for interfaces w i t h o u t a n overv iew) . A s prev ious ly m e n t i o n e d , each interface a lways p r o v i d e d a tota l of 600.000 pixels of i n f o r m a t i o n in all v iews. In interfaces w i t h o u t overv iews , this a m o u n t Chapter 4. Study 1 33 0 c c F c c F c C F i g u r e 4.6: C a l c u l a t i o n of levels o f contex t i n S t u d y 1 R S N interfaces . L e v e l o f n a v i g a t i o n a l context is the f r a c t i o n o f the size o f the p e r i p h e r a l contex t areas C to the to ta l size of the de ta i l v i ew F+C. L e v e l of overv iew context is the f rac t ion o f the size of the o v e r v i e w O to the t o t a l size of a l l v iews O+F+C. 0 f F ) l F i g u r e 4.7: C a l c u l a t i o n o f levels of context i n S t u d y 1 P Z N interfaces . T h e d o t t e d l ine indicates the b o u n d a r y between focus a n d context reg ions , w h i c h is not v i s u a l l y d e m a r c a t e d i n the interfaces . Leve l s of n a v i g a t i o n a l a n d o v e r v i e w context are as i n F i g u r e 4.6. Chapter 4. Study 1 34 was equal to F+C. whi le in interfaces w i t h overviews, it was equa l to O+F+C. T h u s , as shown in F i g u r e s 4.1 to 4.4, interfaces w i t h o u t overviews p r o v i d e d larger de ta i l v iews t h a n their c o u n t e r p a r t s w i th overviews , but the level of nav-igat iona l context was kept cons tant for each nav iga t ion technique , regardless of the presence of a n overview. 4.2.4 R S N - N o O V In te r face A s i l lus trated in F i g u r e 4.1, this interface h a d no overv iew a n d a l lowed users to n a v i g a t e the datase t us ing the m e t a p h o r of e x p a n d i n g a n d c o m p r e s s i n g a r u b b e r sheet w i t h its borders na i led d o w n . U n l i k e in c o n v e n t i o n a l p a n a n d z o o m interfaces, nav iga t ion act ions d i d not, push context regions off-screen, but compressed t h e m in the p e r i p h e r y o f the view, where they r e m a i n e d v i sua l ly salient. F o c u s regions were d e m a r c a t e d by co lored boxes, w h i c h were a lways located in the center of the view. Users cou ld select, a r e c t a n g u l a r a r e a of interest for z o o m i n g in by d r a g g i n g out. a box wi th the left mouse b u t t o n . T h e contents of the selected area then e x p a n d e d to fill one of the two focus regions in a, s m o o t h t r a n s i t i o n . B y default , the sys tem filled the focus whose c e n t r o i d was closer to the c e n t r o i d of the selected a r e a of interest. User s c o u l d also specify w h i c h focus region was to be filled by press ing a modi f ier key. T h e Shif t key was used to specify the top focus a n d the C t r l key for the b o t t o m focus. T h e s e keys were chosen because of their pos i t i on one above the o ther on the k e y b o a r d , w h i c h he lped users associate t h e m w i t h the respect ive foci . A n a c t i o n ana logous to p a n n i n g was a c c o m p l i s h e d v ia h o r i z o n t a l a n d ver t i ca l d r a g m o t i o n s w i t h the r ight mouse b u t t o n , a l lowing users to f ine-tune focus region selections. Users c o u l d z o o m out. by d r a g g i n g out a rec t i l inear region larger t h a n the focus region, the contents of w h i c h were then c o m p r e s s e d to fill the focus reg ion . 4.2.5 P Z N - N o O V In te r f ace A s i l lus trated in F i g u r e 4.2, this interface had no overv iew a n d a l lowed users to navigate us ing convent iona l p a n a n d z o o m interact ions . Just, as w i t h the Chapter 4. Study 1 35 R S N - N o O V interface, users selected a r e c t a n g u l a r area of interest for z o o m i n g in w i th a left mouse d r a g , r e su l t ing in an a n i m a t e d t rans i t ion that c o m p l e t e l y filled the focus w i th the selected area. S i m i l a r l y to the R S N - N o O V interface, b y default the sys tem filled the focus whose c e n t r o i d was closer to the cen tro id of the selected area of interest , bu t users c o u l d also specify w h i c h focus was to be filled by press ing the Shif t or C t r l modi f i er key. Users c o u l d fine-tune the focus selection by p a n n i n g w i t h hor izonta l a n d v e r t i c a l r igh t -mouse drags , a n d g r a d u a l l y z o o m out w i t h v e r t i c a l m i d d l e - m o u s e drags . W h e n m a r k e d regions m o v e d off-screen due to n a v i g a t i o n act ions, c o l o r e d H a l o - l i k e arcs , represent ing the vis ible par t s of e l l ip t i ca l r ings centered on the m a r k e d regions, a p p e a r e d in context areas at the p e r i p h e r y o f the interface. T h e arcs ind i ca ted d i r e c t i o n a n d dis tance to m a r k e d regions, a n d d i sappeared once the m a r k e d regions b e c a m e vis ible on-screen. U n l i k e in the R S N - N o O V interface, the p e r i p h e r a l context areas in this interface were oval rather t h a n r e c t a n g u l a r in shape , were not v i sua l ly d e l i m i t e d , a n d d i d not change in t erms o f reso lut ion . 4.2.6 R S N + O V In te r f ace F i g u r e 4.3 i l lustrates that th is interface used the same n a v i g a t i o n contro ls as the R S N - N o O V interface. It also had an overv iew showing two co lored field o f v iew boxes c o r r e s p o n d i n g to the extent o f each focus in the de ta i l v iew. T h e size a n d locat ion of the field of view boxes were u p d a t e d d y n a m i c a l l y as nav igat ion took place in the deta i l view. Users c o u l d p e r f o r m the r u b b e r sheet nav igat ion equivalents of p a n n i n g and z o o m i n g , as i m p l e m e n t e d in the R S N -N o O V interface, d irec t ly in the overview by d r a g g i n g the field o f v iew boxes, which then u p d a t e d the a p p r o p r i a t e foci in the de ta i l v iew. 4.2.7 P Z N + O V In te r f ace F i g u r e 4.4 i l lustrates that, this interface h a d the same n a v i g a t i o n contro ls as the P Z N - N o O V interface, as well as an overv iew. J u s t as w i t h the R S N + O V interface, the field of v iew boxes in the overview d y n a m i c a l l y reflected n a v i g a t i o n Chapter 4. Study 1 36 in the de ta i l v iew a n d c o u l d b e m a n i p u l a t e d d i rec t ly to c o n t r o l the foci in the detai l v iew. 4.3 Task and Dataset T h e task used in the e x p e r i m e n t was a tree topo log ica l task t h a t required sub-jects to c o m p a r e topo log ica l d is tances between co lored nodes in a large tree dataset a n d determine w h i c h of the dis tances was smal ler . T a s k a n d dataset are b o t h d e s c r i b e d in detai l in C h a p t e r 3. 4.4 Apparatus T h e s t u d y was c o n d u c t e d on two systems r u n n i n g W i n d o w s X P wi th P e n t i u m 4 processors . 2.0 G B R A M . N v i d i a G e F o r c e 2 v ideo cards , a n d 19 inch m o n i -tors conf igured at a reso lut ion of 1280x1024 pixels . T h e e x p e r i m e n t a l software, i n c l u d i n g the interfaces, was ful ly a u t o m a t e d a n d was c o d e d in J a v a 1.4.2 a n d O p e n G L u s i n g the S w i n g a n d G L 4 J a v a l ibraries . 4.5 Participants F o r t y subjects (15 female) between 18 a n d 39 years of age successful ly c o m p l e t e d the s t u d y a n d were each c o m p e n s a t e d S10 for their p a r t i c i p a t i o n . A l l subjects were r i g h t - h a n d e d , h a d n o r m a l or correc ted to n o r m a l v i s i o n , a n d were not co lor b l i n d . T h e y were r e c r u i t e d t h r o u g h advert i sements pos t ed t h r o u g h o u t the univers i ty c a m p u s a n d t h r o u g h an onl ine p a r t i c i p a n t s c h e d u l i n g sys tem. O r i g i n a l l y . 45 subjects p a r t i c i p a t e d in the exper iment . O n e of the subjects was u n a b l e to follow the t r a i n i n g ins truct ions successfully, whi le another was not successful in l earn ing to use the mouse scrol l wheel , r equ ired to p e r f o r m z o o m i n g out in the p a n a n d z o o m interfaces. T h r e e o thers followed the ins truct ions b u t c o m m i t t e d three or more errors (an error rate greater t h a n 10%). T h e s e five Chapter 4. Study 1 37 O r d e r B l o c k 1 B l o c k 2 B l o c k 3 B l o c k 4 B l o c k 5 1 3 1 5 2 '4 2 4 3 1 5 2 3 2 4 3 1 5 4 5 2 4 3 1 5 1 5 2 4 3 T a b l e 4.2: L a t i n square used to c o u n t e r b a l a n c e the o r d e r of presentat ion for level o f context . subjects were treated as out l iers for the purpose of d a t a analys i s , l eaving a to ta l of 40 d a t a points . 4 . 6 D e s i g n T h e eva luat ion used a. 4 ( interface, between subjects) by 5 (level o f context , w i t h i n subjects) des ign, where each level of context c o r r e s p o n d e d to a b lock c o n t a i n i n g 5 tr ials . N a v i g a t i o n t echn ique a n d presence of overview were not e x a m i n e d as separate factors in this s t u d y due to the interface-specif ic v a r i a t i o n of levels of context a n d overv iew sizes. Interface was chosen to be a between-subjects factor due to the poss ib i l i ty o f transfer effects o f n a v i g a t i o n technique in a w i th in - subjec t s des ign, as well as the t ime r e q u i r e d to t ra in subjects on n a v i g a t i o n wi th each technique . S u b j e c t s were r a n d o m l y assigned to each of the four interfaces. L e v e l o f context was chosen to be a w i th in - subjec t s factor to al low c o m p a r i s o n between the different levels of context, for each interface. T o m i n i m i z e o r d e r i n g effects, we c o u n t e r b a l a n c e d the o l d e r of presentat ion for level o f context us ing a L a t i n square , as shown in T a b l e 4.2. In o r d e r to de termine the range of levels of context to be used in the s tudy, we p i lo ted each e x p e r i m e n t a l interface w i t h 5 to 9 subjects u s i n g the e x p e r i m e n t a l task a n d dataset . T h e levels o f context used in p i l o t i n g var i ed from 10% to 95% for n a v i g a t i o n a l context a n d f rom 5% to 50% for overv iew context . B a s e d on Chapter 4. Study 1 38 In ter face L e v e l o f L e v e l o f N a v i g a t i o n a l C o n t e x t O v e r v i e w C o n t e x t R S N - N o O V 4 0 ; 50, 60, 70, 80 0 P Z N - N o O V 30, 40, 50, 60, 70 0 R S N + O V 60 5, 10, 15, 20, 25 P Z N + O V 50 5, 10, 15, 20, 25 T a b l e 4.3: Leve l s o f context used for each interface, in percent . B o l d f a c e denotes values expected to result in o p t i m a l p e r f o r m a n c e for each interface based on p i lo t results. T h e level o f context in the de ta i l v iew of interfaces w i t h an overv iew was set to the m i d d l e level of context used for their c o u n t e r p a r t s w i t h o u t an overview. pi lot results , we selected a range of five levels of context for each interface that p e r f o r m e d best in t erms of c o m p l e t i o n t imes . T h e levels o f context used in each interface are l isted in T a b l e 4.3, wi th the levels o f context expected to p e r f o r m best for each interface based on pi lot results shown in b o l d . 4.7 Procedure T h e e x p e r i m e n t was des igned to fit into a s ingle 60 m i n u t e session. T h e e x p e r i -menter first i n s t r u c t e d subjects on the use o f the different n a v i g a t i o n techniques afforded by the interface to which they h a d been r a n d o m l y ass igned. S u b j e c t s were then shown the e x p e r i m e n t a l task a n d ins tructed that they were to take as m u c h t ime as necessary to solve it correc t ly whi le m a x i m i z i n g the ir efficiency. T h e exper imenter t h e n t ra ined subjects o n the use of long, t h i n h o r i z o n t a l se-lection areas to c o m p l e t e the task. T h i s s trategy h a d been found to i m p r o v e task c o m p l e t i o n t ime in a l l interfaces in p i l o t i n g . Since m a n y of the p a t h s be-tween co lored nodes were hor izonta l , this s trategy enabled subjects to b r i n g them r a p i d l y into focus. Subjects were t h e n given a t r a i n i n g b lock of 5 tr ia ls . D u r i n g these tr ials , subjec t s per formed the task on their o w n . a n d the exper i -Chapter 4. Study 1 39 menter r e m i n d e d subjects of the t r a i n i n g s trategy as n e e d e d . F o r the t r a i n i n g b lock , the m i d d l e level o f context va lue of those shown in T a b l e 4.3 was used for each interface. A t the e n d of the t r a i n i n g session, subjec t s were g iven a one m i n u t e b r e a k . A f t e r the b r e a k , the e x p e r i m e n t e r ex i ted the r o o m where the e x p e r i m e n t was c o n d u c t e d , a n d subjects proceeded w i t h the exper iment . T h e c o m p l e t e t r a i n i n g p r o t o c o l for this s t u d y c a n be found in A p p e n d i x A . In the e x p e r i m e n t , subjects were presented wi th 5 b locks , each c o n t a i n i n g 5 tr ials , for a to ta l of 25 tr ials . A l l subjects were presented w i t h an ident ica l set of quest ions , w i t h a p r e d e t e r m i n e d g r o u p i n g o f quest ions into b locks . T h e o r d e r of b locks was d e t e r m i n e d u s i n g the L a t i n s q u a r e shown in T a b l e 4.2. T h e b locks of quest ions were verified to be i s o m o r p h i c in diff iculty in p i l o t i n g . Subjec t s were g iven a one m i n u t e break between each b lock of quest ions . A t the e n d of the e x p e r i m e n t , subjects c o m p l e t e d ques t ionna ires , w h i c h c a n be f o u n d in A p p e n d i x C . T h e ques t ionna ires were used to col lect i n f o r m a t i o n a b o u t the subjects ' d e m o g r a p h i c b a c k g r o u n d a n d c o m p u t e r usage. T h e y also i n c l u d e d rat ings for ease of use, ease o f n a v i g a t i o n , a n d interface-specif ic features on 5 -po int L i k e r t scales. S p a c e was also p r o v i d e d for subjec t s to c o m m e n t on their exper iences w i t h the interfaces a n d p r o v i d e suggest ions for i m p r o v e m e n t . S h o r t i n f o r m a l interviews were c o n d u c t e d w i t h some o f the subjects based on their q u e s t i o n n a i r e responses. 4.8 Measures O u r p e r f o r m a n c e measures were based o n logged d a t a a n d i n c l u d e d task c o m -plet ion t imes a n d errors . A d d i t i o n a l l y , se l f -reported measures were col lected t h r o u g h the p o s t - e x p e r i m e n t ques t ionna ire , as descr ibed in the prev ious sec-t ion . T h e s t u d y was des igned to m i n i m i z e errors , w i t h task c o m p l e t i o n t imes used as the p r i m a r y measure of p e r f o r m a n c e . Chapter 4. Study 1 40 4 .9 R e s u l t s T h i s sect ion presents the e x p e r i m e n t a l results , r e p o r t i n g in de ta i l o n the results re lated to effects of l e a r n i n g a n d n a v i g a t i o n technique . T h e results for presence of overv iew are s u m m a r i z e d here, a n d deta i l s on these can be found i n B o d n a r ' s master's thesis [12]. P r i o r to analys is , ou t l i e r d a t a l y i n g m o r e t h a n 3 s t a n d a r d dev ia t i ons f r o m the means o f each e x p e r i m e n t a l cell were r e m o v e d from the analys i s . T h e G r e e n h o u s e - G e i s s e r a d j u s t m e n t was used for non- spher ica l d a t a , a n d the B o n -ferroni a d j u s t m e n t for p o s t - h o c c o m p a r i s o n s . A l o n g wi th s ta t i s t i ca l s ignif icance, we r e p o r t p a r t i a l e ta - squared (n2), a m e a s u r e o f effect, size, w h i c h is often m o r e in format ive t h a n s ta t i s t i ca l s ignif icance in a p p l i e d h u m a n - c o m p u t e r i n t e r a c t i o n research [32]. T o in terpre t this value, .01 is a sma l l effect size, .06 is m e d i u m , a n d .14 is large [17]. T h e overa l l results for m e a n c o m p l e t i o n t imes per tr ia l are i l lu s tra ted in F i g -ure 4.8. A one-way A N O V A was r u n to u n d e r s t a n d the effect of interface, level of context , a n d b lock o n c o m p l e t i o n t ime . A s expected , p e r f o r m a n c e i m p r o v e d as subjects progressed t h r o u g h the e x p e r i m e n t , w i th a s ignif icant m a i n effect of b lock (_F(4,144) = 12.309, p < .001, if = -255). T h e r e was also a s ignif icant m a i n effect of interface o n c o m p l e t i o n t i m e (F(3,36) = 2.924, p < .05, rj2 = .196), b u t pos t -hoc c o m p a r i s o n s revealed no s ignif icant pa irwise differences be-tween interfaces. N o significant, in teract ion effect between b lock a n d interface was present . O n e - w a y A N O V A s were per formed to invest igate the effect o f level o f context on c o m p l e t i o n t imes for each interface. A s s h o w n in T a b l e 4.4, there was no s ignif icant effect, of level o f context for any of the interfaces, despi te h igh effect, sizes for three of the four interfaces. T h e best m e a n p e r f o r m a n c e w i t h the R S N -N o O V a n d P Z N - N o O V interfaces o c c u r r e d w i t h context levels o f 50% a n d 40%,, respect ively , while the best, m e a n p e r f o r m a n c e wi th b o t h the R S N + O V a n d P Z N + O V interfaces o c c u r r e d wi th au overv iew size of 15%. M e a n c o m p l e t i o n times for each interface are shown in F i g u r e s 4.9 t h r o u g h 4.12. Separate figures Chapter 4. Study 1 41 RSN-NoON PZN-NoOV •Mr RSN+OV PZN+OV 20 10 0 1 2 3 4 5 Block F i g u r e 4.8: M e a n p e r - t r i a l c o m p l e t i o n t imes b y interface for b locks 1-5, i n sec-onds (JV=40) . I n t e r f a c e d f F S i g . P a r t i a l rj2 R S N - N o O V 1.826, 16.43 1.272 .303 .124 P Z N - N o O V 4,36 .456 .767 .048 R S N + O V 4,36 1.380 .260 .133 P Z N + O V 4,36 1.480 .229 .141 T a b l e 4.4: Resu l t s of four one-way A N O V A s invest igat ing the effect of level of context o n c o m p l e t i o n t ime, b y interface (JV=10). Resu l t s for the R S N - N o O V interface are adjus ted for sphericity. Chapter 4. Study 1 42 are used to repor t these results due to the differences between ranges of levels of context e x a m i n e d for each interface. C o u n t e r to o u r hypothes i s H I , there was no significant difference between per formance w i t h m i d d l e a n d extreme levels of context for either R S N or P Z N . F i g u r e s 4.9 t h r o u g h 4.10 i l lustrate that the expected U - s h a p e d per formance t r e n d for this hypothes is was not present i n d a t a for either nav igat ion tech-nique. S imi lar ly , F i g u r e s 4.11 a n d 4.12 show that , counter to our hypothes i s H 2 , there was no significant difference between per formance w i t h m i d d l e a n d extreme overview sizes for either R S N or P Z N interfaces w i t h overviews. T h e impl ica t ions of this f inding are discussed in deta i l i n B o d n a r ' s thesis [12]. O n average, subjects c o m m i t t e d 0.75 errors over the course of the exper i -ment , for a m e a n error rate of 3.0%. T h e r e were no significant m a i n or interac-t ion effects of interface or level of context o n error rate. 4 . 1 0 S u m m a r y o f R e s u l t s W e s u m m a r i z e our results a c c o r d i n g to the hypotheses s tated i n Sect ion 4.1: R I : F o r b o t h p a n a n d z o o m a n d r u b b e r sheet nav igat ion , m e d i u m levels of context d i d not p e r f o r m better t h a n low or h igh levels. R 2 : F o r b o t h p a n a n d z o o m a n d r u b b e r sheet nav iga t ion , i n interfaces w i t h overviews, m e d i u m overview sizes d i d not p e r f o r m better t h a n low or h igh ones. T h e best per formance in terms of m e a n c o m p l e t i o n t imes o c c u r r e d w i t h a context level of 50% i n the R S N - N o O V interface, a context level of 40% i n the P Z N - N o O V interface, a n d overview sizes of 15% i n b o t h the R S N + O V a n d P Z N + O V interfaces. Chapter 4. Study 1 43 30 20 10 -o\ 1 1 1— —. 1— 1 30 40 50 60 70 80 90 Level of context (%) F i g u r e 4.9: M e a n per - t r ia l comple t ion t imes (dark line) a n d q u a d r a t i c t r e n d line (light line) for the R S N - N o O V interface, in seconds (N=10). F i g u r e 4.10: M e a n p e r - t r i a l comple t ion t imes (dark line) a n d q u a d r a t i c t r e n d line (l ight line) for the P Z N - N o O V interface, i n seconds (JV=10) . Chapter 4. Study 1 44 100 30 20 10 0 0 5 10 15 20 25 30 Level of context (%) F i g u r e 4.11: M e a n p e r - t r i a l c o m p l e t i o n t imes (dark line) a n d q u a d r a t i c t r e n d line (light line) for the R S N + O V interface, i n seconds (N=W). F i g u r e 4.12: M e a n per - t r ia l comple t ion t imes (dark line) a n d q u a d r a t i c t r e n d line (l ight line) for the P Z N + O V interface, i n seconds (JV=10). Chapter 4. Study 1 45 4.11 Strategies In order to g a i n a bet ter u n d e r s t a n d i n g of w h y level of context d i d not s ignif i-cant ly affect c o m p l e t i o n t imes , we e x a m i n e d logged n a v i g a t i o n a n d reset ac t i on d a t a . A l t h o u g h the logg ing format used in this s tudy was not c o n d u c i v e to a s tat is t ical analys i s of these d a t a , a m a n u a l ana lys i s o f n a v i g a t i o n a n d reset, pat terns for i n d i v i d u a l subjec t s showed t h a t , for each interface, most subjec t s consistent ly used a sma l l n u m b e r of well-defined strategies t h r o u g h o u t the ex-per iment . S o m e of these strategies were re la ted to n a v i g a t i o n pat terns , whi le others invo lved different ways of us ing overv iews or c o m p e n s a t i n g for the a b -sence of overviews by means of us ing the reset funct ion . T h e strategies re la ted to n a v i g a t i o n are deta i led in T a b l e 4.5. while those re lated to overv iew a n d reset usage are discussed in B o d n a r s thesis [12]. B a s e d on the m a n u a l ana lys i s of log d a t a , we g r o u p e d subjects w h o used the R S N - N o O V , P Z N - N o O V , a n d R S N + O V interfaces a c c o r d i n g to the ir p r i m a r y n a v i g a t i o n strategy. O n e - w a y A N O V A s were t h e n p e r f o r m e d for each interface to u n d e r s t a n d the effect of nav iga t ion s t ra tegy on c o m p l e t i o n t ime . F o r b o t h of the R S N interfaces, the "zoom, then p a n " s trategy resul ted in faster m e a n c o m p l e t i o n t imes t h a n the "zoom only" s t ra tegy (67.9 seconds vs. 75.7 seconds for R S N - N o O V ; 63.8 seconds vs. 75.6 seconds for R S N + O V ) . A l t h o u g h this difference was not, s ta t i s t i ca l ly significant for R S N - N o O V ( F ( 2 . 4 7 ) = 1.616, p > .2, if = .033), a n d on ly border l ine s ignif icant for the R S N + O V ( F ( 2 , 4 7 ) = 3.008, p < .09. if = .059), these results suggested the i m p o r t a n c e of the use of the r u b b e r sheet equivalent, of p a n n i n g for effective n a v i g a t i o n , a n d led us to emphas ize the use of this o p e r a t i o n in the t r a i n i n g pro toco l of S t u d y 2. F o r the P Z N - N o O V interface, the three strategies were s igni f icant ly different in t erms of c o m p l e t i o n t ime (F(2 .47 ) = 9.226, p < .001, if = .282). P o s t - h o c pairwise c o m p a r i s o n s showed that the "zoom in , then pan" s trategy was signif-icant ly slower (p < .01) t h a n b o t h the "zoom in . then z o o m out" s trategy a n d the "overview 7" s trategy ( m e a n c o m p l e t i o n t imes: 104.2 seconds. 80.0 seconds, a n d 75.4 seconds, respect ive ly ) . T h i s f inding m o t i v a t e d us to p r o v i d e a de ta i l ed Chapter 4. Study 1 46 I n t e r f a c e S t r a t e g y D e s c r i p t i o n R S N - N o O V Z o o m o n l y (N=6) Z o o m , t h e n p a n (N=4) U s e R S N z o o m only. U s e R S N z o o m a n d p a n . P Z N - N o O V Z o o m i n . then p a n (N=4) Z o o m i n . then z o o m out (N—4) O v e r v i e w (N=2) Z o o m into n e i g h b o u r h o o d of a m a r k e d n o d e , then follow p a t h by p a n n i n g . Z o o m into n e i g h b o u r h o o d of a m a r k e d node , then z o o m out to reveal p a t h . Z o o m in w i t h one focus, then use the o ther as overview. R S N + O V Z o o m o n l y (N=6) Z o o m , t h e n p a n (N=4) Use R S N z o o m only . U s e R S N z o o m a n d p a n . P Z N + O V N o n e T a b l e 4.5: Strategies re la ted to n a v i g a t i o n deve loped by subjec t s in S t u d y 1. t r a i n i n g s trategy for the P Z N - N o O V interface in S t u d y 2. 4.12 Discussion Resu l t s o f this s tudy show t h a t level of context d i d not s igni f icant ly affect per -formance as measured by c o m p l e t i o n t imes w i t h e i ther p a n a n d z o o m or r u b b e r sheet n a v i g a t i o n . A l s o , the expected U - s h a p e of the c o m p l e t i o n t i m e . d a t a , i n d i c a t i n g super ior p e r f o r m a n c e on m e d i u m r a t h e r t h a n h i g h or low levels of context , was not present. T h i s f inding seems s u r p r i s i n g since, for b o t h n a v i g a -t ion techniques , we e x a m i n e d a range o f levels of context, s p a n n i n g a s ignif icant p r o p o r t i o n o f the avai lable screen real estate. It is poss ible t h a t o u r s tudy des ign d i d not p r o v i d e sufficient power to detect, differences w i t h i n the ranges for each interface. T h i s e x p l a n a t i o n w o u l d account for the m e d i u m to h igh effect sizes of level of context in all interfaces, as shown in T a b l e 4.4. Chapter 4. Study 1 47 A n o t h e r poss ibi l i ty is that the ranges we e x a m i n e d were st i l l not sufficiently large to s ignif icantly i m p a c t the way subjects nav igated . T h i s poss ibi l i ty is s u p p o r t e d b y qual i ta t ive feedback f r o m quest ionnaires a n d fol low-up interviews w i t h the subjects . O n l y one subject c o m m e n t e d o n the difference between smal l a n d large levels o f context , a n d a n u m b e r o f subjects s ta ted that they were not even aware of the changes i n level of context d u r i n g the exper iment . W e there-fore hypothes ize that , for b o t h p a n a n d z o o m a n d r u b b e r sheet nav iga t ion , the range of levels of context conduc ive to effective nav iga t ion is sufficiently b r o a d to encompass the range e x a m i n e d i n o u r exper iment , a n d t h a t differences between levels of context w i t h i n this range do not s ignif icantly i m p a c t perfor-mance . T h u s , for each nav igat ion technique, there m a y not exist an o p t i m a l level of context value for p e r f o r m i n g a given task, b u t rather u p p e r a n d lower b o u n d s b e y o n d w h i c h per formance deteriorates, s imi lar to the b o u n d s that have been proposed for overview sizes [41]. It is also possible that the effect of level of context on nav igat ion per formance can be represented b y a d iscont inuous value funct ion . In p a r t i c u l a r , a sma l l a m o u n t of context m a y i m p r o v e per formance c o m p a r e d to zero context , w i t h further increases p r o v i d i n g no a d d e d benefits. F o r three of the four interfaces e x a m i n e d in this exper iment , nav igat ion strategies used b y the subjects h a d a not iceable effect on per formance . In the R S N interfaces, the two strategies differed on ly in whether subjects chose to use the rubber sheet equivalent of p a n n i n g or to rely solely o n r u b b e r sheet z o o m i n g actions. T h i s difference l ikely resulted f rom the fact that the nav iga t ion s trategy used in t r a i n i n g (the use of long , t h i n hor izonta l selection areas) on ly involved z o o m i n g , a n d , a l though subjects were shown how to p a n , the use of p a n n i n g was left to their d iscret ion. P e r h a p s more surpr i s ing was the deve lopment of three d is t inct nav iga t ion strategies w i t h the P Z N - N o O V interface. T h i s f ind ing c a n b e exp la ined b y not-ing that this interface conta ined two focus areas of the same size a n d resolut ion, p r o v i d i n g subjects w i th the opt ions of us ing one v iew exclusively, a l t ernat ing between the focus areas d e p e n d i n g o n the locat ion of the m a r k e d nodes , or us ing one focus as a deta i l v iew a n d another as a n overview. F u r t h e r m o r e , the dis t inc-Chapter 4. Study 1 48 t i o n between the "zoom i n , then p a n " a n d "zoom i n , t h e n z o o m out" strategies c a n be a t t r i b u t e d to the fact that z o o m out , like p a n , was not specif ically covered by the t r a i n i n g strategy used i n the study. T h e s ignif icantly slower per formance of subjects us ing the "zoom i n , then pan" points to the inefficiency of p a n n i n g at a h igh magni f i ca t ion level, consistent w i t h results of prev ious studies [46]. T h e lack of significant difference i n per formance between the "overview" a n d "zoom i n , then z o o m out" strategies suggests that g r a d u a l z o o m i n g out m a y offer a n adequate subst i tute for the lack of a n overview i n nav igat ion tasks such as the one used in o u r s tudy, where users must follow a specific nav iga t ion p a t h . W e h a d in i t ia l ly be l ieved that the use of two focus areas i n this s t u d y w o u l d enable users to explore mul t ip l e areas of the same dataset s imultaneously , a n d hence prov ide per formance benefits. However , observat ion d a t a ind ica ted that this feature increased interface complexi ty , forcing subjects to invest more cog-nit ive effort into c o o r d i n a t i n g nav iga t ion between the mul t ip le focus regions t h a n they d i d into c o m p l e t i n g the task. A d d i t i o n a l l y , the presence of two focus areas was p a r t l y responsible for the deve lopment o f interface-specific strategies, i n p a r t i c u l a r the "overview" strategy i n the P Z N - N o O V interface, that repre-sented a confound ing factor. F i n a l l y , the H a l o - l i k e arcs as i m p l e m e n t e d for this s t u d y caused some subjects difficulties b y o c c l u d i n g regions of interest. T h e s e issues were addressed by m e r g i n g the focus areas in each interface a n d m a k i n g the Ha lo - l ike arcs translucent for the purposes of S t u d y 2. T h r e e other issues w i t h this s t u d y were no ted a n d addressed in S t u d y 2. F i r s t , qual i tat ive d a t a showed that m a n y subjects found it difficult to r e m e m -ber intermediate task results, such as the n u m b e r of nodes on the first p a t h ex-p lored for a g iven task instance. Second , the absence of the exper imenter in the exper iment r o o m after the t ra in ing p e r i o d h a m p e r e d o u r analysis of interface-specific strategies. Subjec t s in S t u d y 2 were therefore g iven a p e n a n d paper to record intermediate task results a n d observed t h r o u g h o u t the exper iment . F i n a l l y , nav iga t ion a n d reset ac t ion d a t a from this s t u d y were not logged in a format conduc ive to s tat is t ical analysis . T h i s capab i l i t y was implemented in S t u d y 2, a l lowing these d a t a to be ana lyzed as a d d i t i o n a l dependent variables. Chapter 5 49 Study 2 T h e goal of this s t u d y was to evaluate the p e r f o r m a n c e of p a n a n d z o o m a n d r u b b e r sheet n a v i g a t i o n techniques w i t h a n d w i t h o u t an overview. S u b j e c t s used interfaces based o n those used in S t u d y 1. each p r o v i d i n g the best, p e r f o r m i n g level of context as d i scussed in C h a p t e r 4 . T h i s c h a p t e r describes the s t u d y a n d presents the aspects o f its results re la ted to the effects of nav iga t ion t echn ique . 5.1 Hypotheses O u r hypotheses for this s t u d y were m o t i v a t e d by f indings reported in the l iter-a t u r e a n d the results o f S t u d y 1. F i r s t , we expected r u b b e r sheet n a v i g a t i o n to per for in better t h a n p a n a n d z o o m n a v i g a t i o n because , as discussed in C h a p t e r 2. focus+context a p p r o a c h e s have been shown to p e r f o r m better t h a n p a n a n d J z o o m interfaces for a var ie ty of n a v i g a t i o n tasks. S e c o n d , we d i d not expect a n overv iew to s igni f icant ly improve the p e r f o r m a n c e of r u b b e r sheet n a v i g a -t i on , because f o c u s + c o n t e x t approaches b y design a t t e m p t to prov ide the same c o n t e x t u a l i n f o r m a t i o n as a n overview, but in an integrated way.' F i n a l l y , we expected that an overv iew w o u l d s igni f icant ly i m p r o v e the pe r fo rmance o f p a n a n d z o o m nav iga t ion , because most prev ious studies h a v e shown that overv iews decrease nav igat ion t ime a n d help the user m a i n t a i n or i enta t ion w i t h i n a p a n a n d z o o m interface. O u r hypotheses were therefore as follows: H3: R u b b e r sheet, n a v i g a t i o n will p e r f o r m be t te r t h a n p a n a n d z o o m n a v i g a -t ion , i n d e p e n d e n t l y of the presence or absence o f a n overview. H4: F o r rubber sheet n a v i g a t i o n , the presence o f an overv iew will not result in Chapter 5. Study 2 50 bet ter p e r f o r m a n c e . H 5 : F o r p a n a n d z o o m n a v i g a t i o n , the presence of a n overv iew wi l l result in bet ter p e r f o r m a n c e . 5.2 Interfaces T h e interfaces e x a m i n e d in this s t u d y are i l lus trated in F i g u r e s 5.1 to 5.4. E a c h o f the interfaces was based on a n interface used in S t u d y 1. as d e s c r i b e d in detai l in Sec t ion 4.2. B a s e d on the results of S t u d y 1. m u l t i p l e foci were r e m o v e d from all interfaces, a n d Halo - l ike arcs in the P Z N interfaces were m a d e trans lucent , such that they were sti l l v i sua l ly sal ient but, d i d not ful ly o c c l u d e areas of the dataset . T h e des ign rat ionale for these changes is d iscussed in Sec t ion 4.12. T h e values for level of context a n d overview size in each interface were der ived f rom the values t h a t prov ided the best p e r f o r m a n c e in S t u d y 1, as out l ined in Sect ion 4.9. In p a r t i c u l a r , in R S N interfaces, the fract ion of the v iew occup ied by the p e r i p h e r a l context area was set to 50%. whi le in P Z N interfaces, the fract ion of the v iew where H a l o - l i k e arcs cou ld a p p e a r was set to 40%. In interfaces w i t h overviews, overv iews a n d detai l views c o m p r i s e d 15% a n d 85% of the to ta l n u m b e r of pixels ava i lab le , respectively. T h e c a l c u l a t i o n of levels of context for this s t u d y was s i m i l a r to that in S t u d y 1. but reflected the use of a single focus in each interface, as i l lus tra ted in F i g u r e s 5.5 a n d 5.6. A s in S t u d y 1. each interface a lways p r o v i d e d a to ta l of 600.000 pixels of i n f o r m a t i o n in all views. 5.3 Task and Dataset T h i s e x p e r i m e n t ut i l ized the same tree topo log ica l task a n d dataset as S t u d y 1. B o t h the task a n d the dataset are descr ibed in deta i l in C h a p t e r 3. Chapter 5. Study 2 51 F i g u r e 5.1: R S N - N o O V interface u s e d i n S t u d y 2. T h i s interface p r o v i d e d a s ingle focus r a t h e r t h a n two foci as i n S t u d y 1 (see F i g u r e 4.1). L e v e l o f contex t was set to 50%. F i g u r e 5.2: P Z N - N o O V interface u s e d i n S t u d y 2. T h i s interface cons i s t ed of a s ingle focus r a t h e r t h a n two foci as in S t u d y 1 (see F i g u r e 4.2), a n d p r o v i d e d t rans lucent r a t h e r t h a n o p a q u e arcs . L e v e l of contex t was set to 40%. Chapter 5. Study 2 52 F i g u r e 5.4: P Z N + O V interface used i n S t u d y 2. T h i s interface p r o v i d e d a s ingle focus r a t h e r t h a n two foci as in S t u d y 1 (see F i g u r e 4.4) a n d t rans lucent r a t h e r t h a n o p a q u e arcs . T h e o v e r v i e w c o m p r i s e d 15% of t o t a l ava i lab l e pixels . Chapter 5. Study 2 53 0 c c c F c c c F i g u r e 5.5: C a l c u l a t i o n of levels o f context i n S t u d y 2 R S N interfaces . L e v e l of n a v i g a t i o n a l context is the f rac t ion o f the size o f the p e r i p h e r a l contex t areas , C , to the t o t a l size of the d e t a i l v iew, F+C. L e v e l o f overv iew context is the f r a c t i o n of the size o f the overv iew , O, to the t o t a l size o f a l l v iews , O+F+C. 0 c 1 c ( F A F i g u r e 5.6: C a l c u l a t i o n of levels o f context i n S t u d y 2 P Z N interfaces . T h e d o t t e d line indicates the b o u n d a r y between focus a n d context reg ions , w h i c h is not v i s u a l l y d e m a r c a t e d in the interfaces . L e v e l s o f n a v i g a t i o n a l a n d overv iew context are as i n F i g u r e 5.5. Chapter 5. Study 2 54 5.4 Apparatus T h e e x p e r i m e n t was c o n d u c t e d o n the two sys tems used to conduct , S t u d y ] . conf igured as discussed in Sec t ion 4.4. 5.5 Participants F o r t y subjects (24 female) between 18 a n d 39 years of age successful ly comple ted the s tudy a n d were each c o m p e n s a t e d $15 for their p a r t i c i p a t i o n . A l l subjects were r i g h t - h a n d e d , h a d n o r m a l or correc ted to n o r m a l v i s i o n , a n d were not co lor b l i n d . P a r t i c i p a n t s were r e c r u i t e d t h r o u g h advert i sements pos ted throughout the un ivers i ty c a m p u s a n d t h r o u g h an onl ine p a r t i c i p a n t s c h e d u l i n g sys tem. O r i g i n a l l y . 44 subjects p a r t i c i p a t e d in the exper iment . T w o of the subjects were u n a b l e to follow the t r a i n i n g ins truct ions successfully, whi le two others fol-lowed the ins truct ions but c o m m i t t e d four or more errors (an error rate greater t h a n 10%). T h e s e subjects were treated as out l iers for the p u r p o s e of d a t a analys is , l eav ing a tota l of 40 d a t a points . 5.6 Design T h e e v a l u a t i o n used a 2 ( n a v i g a t i o n , between subjects ) by 2 (presence of overview between subjects ) by 7 (blocks, w i t h i n subjects) des ign, where each block c o n -ta ined 5 tr ia ls . Subjects were r a n d o m l y ass igned to each of the four c o m b i n a -t ions of n a v i g a t i o n a n d presence of overview. A s in S t u d y ) , a between-subjects design was chosen for the n a v i g a t i o n a n d presence of overv iew factors to avo id potent ia l transfer effects, as well due to the t ime required to t r a i n subjects on n a v i g a t i o n w i t h each technique . D u e to the s ignif icant l e a r n i n g effects found in S t u d y 1. the n u m b e r of b locks was increased from 5 to 7 to ensure that subject p e r f o r m a n c e reached a p la teau that w o u l d enable an accura te c o m p a r i s o n of the effects of nav iga t ion and presence of overview. P i l o t i n g showed that 7 blocks of 5 tr ials each were sufficient for p e r f o r m a n c e to p la t eau . Chapter 5. Study 2 55 5.7 Procedure T h e e x p e r i m e n t was des igned to fit into a single 90 m i n u t e session. T h e e x p e r i -menter first ins tructed subjects on the use of the nav iga t ion funct ions afforded by the interface to w h i c h they h a d been r a n d o m l y ass igned. S u b j e c t s were t h e n shown the e x p e r i m e n t a l task a n d ins truc ted that they were to take as m u c h t ime as necessary to solve it correc t ly whi le m a x i m i z i n g their efficiency. T h e e x p e r i -menter t h e n d e m o n s t r a t e d the t r a i n i n g s trategy to be used by the subject . T h e t r a i n i n g strategies var ied d e p e n d i n g o n the nav igat ion t echn ique a n d presence or absence of an overview in each interface, a n d can be found in A p p e n d i x B . A l l t r a i n i n g strategies s tar ted w i t h d r a g g i n g out a l o n g t h i n selection area a long the h o r i z o n t a l p a t h between two o f the m a r k e d nodes , as descr ibed in Sect ion 4.7. F o r the r u b b e r sheet interfaces, selecting a l o n g t h i n h o r i z o n t a l area h a d the effect of s t re t ch ing the datase t a l o n g the vert ica l axis , as i l lus trated in F i g u r e s 5.1 a n d 5.3. F o r the p a n a n d z o o m interfaces, se lect ing a long t h i n h o r i z o n t a l area, h a d the effect of z o o m i n g the contents of the focus box to fill the ent ire v iew, as i l lus trated in F i g u r e s 5.2 a n d 5.4. F o r a l l interfaces, after z o o m i n g into the p a t h between the m a r k e d nodes , subjects were ins tructed to count nodes that became v i s u a l l y salient. F o l l o w -ing this s tep, subjects were s h o w n how to d r a g out long t h i n hor i zonta l a n d ver t i ca l se lect ion areas to e x p a n d other compressed regions a l o n g the p a t h . In a c c o r d a n c e w i t h the results of S t u d y 1, w h i c h showed t h a t u s i n g the p a n n i n g funct ion i m p r o v e d p e r f o r m a n c e , subjects were ins tructed to use p a n n i n g in p a n a n d z o o m interfaces, or the equivalent o f p a n n i n g in r u b b e r sheet interfaces, to m a k e adjus tment s to the focus area if needed. In interfaces w i t h overviews, subjects were ins tructed how to use b o t h the overview a n d deta i l views for nav-igat ion a n d c o u n t i n g nodes, bu t were not exp l i c i t ly to ld to nav igate in e i ther view. F o r the P Z N - N o O V , a l t h o u g h the "overview" and "zoom in . then z o o m out" strategies p e r f o r m e d equa l ly well in S t u d y 1. the "overview" strategy was too s imi lar to usage pat terns in the P Z N + O V interface', a n d was e l i m i n a t e d as a Chapter 5. Study 2 56 poss ib i l i ty t h r o u g h the r e m o v a l of m u l t i p l e foci for this s tudy. Subjec t s were therefore i n s t r u c t e d to use the "zoom i n . then z o o m out : : s trategy. T h i s s tra tegy invo lved s lowly z o o m i n g out a n d a d d i n g nodes as they a p p e a r e d a l o n g the p a t h u p the tree, as shown in F i g u r e 5.2. F o l l o w i n g the d i scovery of one of the two topo log ica l d is tances , subjects were i n s t r u c t e d to reset the interface before d i scover ing the second d i s tance . T h i s s tra tegy was m o t i v a t e d by the f indings of S t u d y 1 t h a t showed that the use of reset between the d i scovery of the two topo log ica l d i s tances i m p r o v e d per formance . T h i s result a n d its impl i ca t ions are discussed in de ta i l in B o d n a r ' s thesis [12]. A f t e r b e i n g s h o w n the strategies, subjects were g iven a t r a i n i n g b lock o f 5 tr ials . F o r each o f the first 2 trials , the e x p e r i m e n t e r d e m o n s t r a t e d s o l v i n g the quest ion u s i n g the strategies a n d then asked subjects to repeat this so lut ion . F o r the last 3 tr ia ls of the session, subjects p e r f o r m e d the task on the ir own , a n d the e x p e r i m e n t e r r e m i n d e d t h e m of the t ra ined s t ra tegy as needed. A t the end of the t r a i n i n g session, subjects were given a one m i n u t e break before p r o c e e d i n g w i t h the exper iment . D u r i n g b o t h the t r a i n i n g b lock a n d the e x p e r i m e n t , subjec t s were p r o v i d e d w i t h a, pen a n d p a p e r to r e c o r d in termedia te task results , a n d the e x p e r i m e n t e r r e m a i n e d in the exper iment r o o m to observe the ir progress . T h e s e changes f r o m S t u d y 1 were i m p l e m e n t e d to address p r o c e d u r a l issues discussed in Sec t ion 4.12. In the e x p e r i m e n t , subjec t s were presented w i t h 7 b locks , each c o n t a i n i n g 5 tr ials , for a to ta l of 35 tr ia l s . A l l subjects were presented w i t h a n ident ica l set o f quest ions. T h e g r o u p i n g o f quest ions to b lock was p r e d e t e r m i n e d , but the o r d e r of b locks was r a n d o m l y generated for each subjec t . T h e blocks of quest ions were verified to be i s o m o r p h i c in difficulty in p i l o t i n g . T h e e x p e r i m e n t e r cont inued to observe the subject t h r o u g h o u t the s tudy, but never in tervened . S u b j e c t s were g iven a one m i n u t e b r e a k between each b lock of quest ions . A t the e n d of the e x p e r i m e n t , subjects c o m p l e t e d quest ionnaires , w h i c h c a n be found in A p p e n d i x D . A s in S t u d y 1, the ques t ionnaires were used to col lect i n f o r m a t i o n about the subjec t s ' d e m o g r a p h i c b a c k g r o u n d a n d c o m p u t e r usage as Chapter 5. Study 2 57 well as ra t ings for ease o f use, ease of n a v i g a t i o n , a n d interface-specif ic features on 5-point L i k e r t scales. F o r this s tudy, the ques t ionnaires also i n c l u d e d the N A S A - T L X scales [23], a s t a n d a r d i z e d i n s t r u m e n t for assessing various d i m e n -sions o f w o r k l o a d . T h i s i n s t r u m e n t was i n c l u d e d to i m p r o v e o u r u n d e r s t a n d i n g o f h o w d e m a n d i n g each interface was perce ived by subjects . S h o r t in formal in -terv iews were c o n d u c t e d w i t h some of the subjects based on their ques t ionnaire responses . 5.8 Measures O u r p e r f o r m a n c e measures were based o n logged d a t a a n d i n c l u d e d task c o m p l e -t ion t imes , errors , n a v i g a t i o n act ions (pan , z o o m i n , a n d z o o m out) , a n d reset act ions . T h e inc lus ion o f the two latter measures was m o t i v a t e d by the resul ts of S t u d y 1. where n a v i g a t i o n a n d reset pa t t erns were found to have a s t r o n g inf luence on per formance . Sel f repor ted measures were co l lected t h r o u g h the p o s t - e x p e r i m e n t ques t ionna ire , as descr ibed in the prev ious sect ion. L i k e S t u d y 1, this s t u d y was des igned to m i n i m i z e errors , w i t h task c o m p l e t i o n t imes a n d n u m b e r o f nav iga t ion a n d reset act ions used as the p r i m a r y measures of perfor-m a n c e . 5.9 Results T h i s sect ion presents the results for b o t h p e r f o r m a n c e a n d self -reported m e a -sures o f the exper iment . T h e results for l e a r n i n g effects a n d nav iga t ion technique are r e p o r t e d in detai l , wh i l e the results for presence o f overv iew are s u m m a r i z e d . D e t a i l e d r e p o r t i n g of the la t ter c a n be found in B o d n a r s master 's thesis [12]. A series of A N O V A s was r u n to u n d e r s t a n d the effect o f nav iga t ion a n d overv iew on the p e r f o r m a n c e a n d sel f -reported measures . P r i o r to these a n a l y -ses, out l i er da ta l y ing m o r e t h a n 3 s t a n d a r d dev ia t ions f rom the means of each cell were removed from the analys is . T h e G r e e n h o u s e - G e i s s e r adjustment was used for non-spher ica l d a t a , a n d the B o n f e r r o n i a d j u s t m e n t for pos t -hoc c o m -Chapter 5. Study 2 r>8 90 80 RSN-NoOV PZN-NoOV RSN+OV * - PZN+OV 20 10 0 2 3 4 5 6 7 Block F i g u r e 5.7: M e a n per - t r ia l comple t ion times b y interface for b locks 1-7, i n sec-onds (JV=40) . parisons. A s in the discussion of the results of S t u d y 1 in Sect ion 4.9, we repor t p a r t i a l e ta -squared (r/ 2 ) , a measure of effect size. T o interpret this value, .01 is a s m a l l effect size, .06 is m e d i u m , a n d .14 is large [17]. T h e overa l l results for m e a n comple t ion t imes per t r i a l are i l lu s tra ted in F i g u r e 5.7. A s expected , per formance i m p r o v e d as subjects progressed t h r o u g h the exper iment , a l though the rate of improvement d i d v a r y a m o n g the interfaces, w i t h a signif icant m a i n effect of b lock (F(3.174,114.26) = 44.568, p < .001, r / 2 = .553) a n d a significant interact ion between b lock a n d nav iga t ion (F(3 .176 ,114 .35 = 3.721), p < 0.02, if = .094). Separate one-way repeated measures A N O V A s were r u n for each of the inter-faces to de termine per formance plateaus. Pos t -hoc pairwise compar i sons showed no differences between blocks 5, 6, a n d 7 for any of the interfaces, i n d i c a t i n g that p e r f o r m a n c e h a d reached a p la teau b y the e n d of the exper iment i n a l l interfaces, a n d therefore t h a t the effect of b lock was successfully contro l l ed for. T h u s , for the analyses of c o m p l e t i o n t imes, nav iga t ion act ions, a n d reset ac-tions, we focus exclusively on blocks 1 a n d 7, which represent p e r f o r m a n c e at Chapter 5. Study 2 59 D e p e n d e n t v a r i a b l e M e a n M e a n F S i g . P a r t i a l ( R S N ) ( P Z N ) v2 C o m p l e t i o n t i m e (seconds) 59.58 47.49 13.744 .001 .276 N u m b e r of n a v i g a t i o n act ions 4.55 4.04 3.087 .090 .079 N u m b e r of resets 1.228 0.968 4.912 .014 .156 T a b l e 5.1: M e a n s a n d effects of n a v i g a t i o n technique for the c o m p l e t i o n t ime , n u m b e r o f n a v i g a t i o n act ions , a n d n u m b e r of resets dependent variables (N=40). Degrees of freedom are (1,36) for all d e p e n d e n t variables . the b e g i n n i n g a n d e n d o f the exper iment . F o r these analyses , 2 (nav igat ion) b y 2 (presence of overv iew) by 2 (block) A N O V A s were p e r f o r m e d . C o u n t e r to o u r h y p o t h e s i s H3, b o t h o u r logged a n d se l f -reported measures showed that p a n a n d z o o m n a v i g a t i o n o u t p e r f o r m e d r u b b e r sheet n a v i g a t i o n . T h e results for c o m p l e t i o n t imes by n a v i g a t i o n for b locks 1 a n d 7 are i l lus tra ted in F i g u r e 5.8. B o t h at the b e g i n n i n g a n d at the end of the e x p e r i m e n t , p a n a n d z o o m interfaces were s ignif icant ly faster than interfaces u s i n g r u b b e r sheet n a v i g a t i o n . S u b j e c t s also per formed b o r d e r l i n e s igni f icant ly fewer n a v i g a t i o n act ions a n d s igni f icant ly fewer resets u s i n g p a n a n d z o o m n a v i g a t i o n , as s h o w n in T a b l e 5.1. T w o - w a y ( n a v i g a t i o n by overview) A N O V A s were c o n d u c t e d on each o f the N A S A - T L X measures. . T h e s e analyses showed that m e n t a l d e m a n d was signif-i cant ly lower in the p a n a n d z o o m interfaces ( F ( ] , 3 6 ) = 4.214, p < .05, rf = .105). Subjec t s also r e p o r t e d that p a n a n d z o o m interfaces were s igni f icant ly easier to navigate ( F ( l , 3 6 ) = 10.385, p < .005, » / 2 = .224). B o t h of these self-repor ted measures s u p p o r t the results o b t a i n e d from the logged p e r f o r m a n c e measures . Presence of overv iew h a d no significant, effect, on any of the p e r f o r m a n c e measures . T h i s f i n d i n g suppor t s o u r hypothes i s H4, but, is counter to o u r hypothes i s H5. T h e self -reported measures d i d , however, favor an overview. S u b j e c t s r ep o r t ed a lower phys ica l d e m a n d for interfaces wi th an overv iew a n d F i g u r e 5.8: M e a n per - t r ia l comple t ion t imes i n seconds b y nav iga t ion technique for b locks 1 a n d 7, above, a n d by b lock , below ( N = 4 0 ) . P Z N was s igni f icant ly faster t h a n R S N i n b o t h blocks, regardless of presence or absence of overview. Chapter 5. Study 2 61 found t h e m m o r e enjoyable to use. B o d n a r s thesis [12] reports o n a n d discussed these results in m o r e de ta i l . O n average, subjects c o m m i t t e d 1.6 errors over the course o f the e x p e r i m e n t , for a m e a n error rate of 4.7%. T h e r e were no s ignif icant m a i n or in terac t ion effects of n a v i g a t i o n or presence of overview o n error rate. 5.10 Summary of Results W e s u m m a r i z e o u r results a c c o r d i n g to the hypotheses s tated in Sect ion 5.1: R 3 : P a n a n d z o o m na.viga.tion interfaces p e r f o r m e d better t h a n r u b b e r sheet n a v i g a t i o n interfaces in terms of c o m p l e t i o n t imes, n u m b e r o f n a v i g a t i o n act ions , a n d n u m b e r of reset, act ions . M e n t a l d e m a n d was also repor ted as lower in p a n a n d z o o m interfaces. R 4 : F o r r u b b e r sheet n a v i g a t i o n , h a v i n g an overv iew m a d e no significant, dif-ference in terms of c o m p l e t i o n times, n a v i g a t i o n act ions , or resets. H a v i n g an overv iew was, however , reported to r e d u c e phys i ca l d e m a n d . R 5 : S i m i l a r l y , for p a n a n d z o o m nav iga t ion , h a v i n g an overview m a d e no signif-icant, difference in t erms o f c o m p l e t i o n t imes , nav iga t ion act ions , or resets. H a v i n g an overv iew was, however, r e p o r t e d to reduce p h y s i c a l d e m a n d . 5.11 Discussion In S t u d y 2. p a n a n d z o o m n a v i g a t i o n resul ted in lower task c o m p l e t i o n t ime , n u m b e r of n a v i g a t i o n act ions , a n d n u m b e r of reset act ions t h a n r u b b e r sheet nav iga t ion . T h e s e results present convergent ev idence for p a n a n d z o o m nav-igation o u t p e r f o r m i n g r u b b e r sheet, nav iga t ion , a f inding that c a n be p a r t i a l l y exp la ined b y the na ture of the task used in o u r s tudy . F o l l o w i n g a topo log ica l p a t h in a large tree datase t u n d e r the d i s t o r t i o n inherent in r u b b e r sheet n a v i -gat ion m a y have caused a loss of or ientat ion . T h i s e x p l a n a t i o n is s u p p o r t e d by the fact that r u b b e r sheet interfaces required s igni f icant ly m o r e reset act ions . Chapter 5. Study 2 62 O u r observat ions showed t h a t , w h e n subjects seemed d i sor iented in terms o f their l o c a t i o n in the datase t . they frequent ly used the reset key to revert to the o r i g i n a l state of the v i s u a l i z a t i o n , w h i c h indicates t h a t the n u m b e r of reset act ions m a y be a good i n d i c a t o r of loss of o r i e n t a t i o n . T h e f ind ing that p a n a n d z o o m n a v i g a t i o n p e r f o r m e d bet ter than r u b b e r sheet n a v i g a t i o n is not cons i s tent w i t h G u t w i n a n d S k o p i k ' s f inding that fo-c u s + c o n t e x t interfaces c a n p e r f o r m as well as or bet ter on p a t h n a v i g a t i o n tasks as p a n n i n g interfaces w i t h overviews [22]. T o e x p l a i n this d i screpancy , it must be n o t e d that G u t w i n a n d Skopik ' s n o n - d i s t o r t i o n - b a s e d interfaces d i d not i m p l e m e n t z o o m i n g a n d featured an i n t e r a c t i o n m o d e l users found confus ing . F u r t h e r m o r e , the task used in their s t u d y requ ired users to move the m o u s e po in ter a l o n g a p a t h . G u t w i n a n d S k o p i k suggested that a task that s i m p l y re-quires users to v i sua l ly n a v i g a t e a p a t h , w h i c h was the case in o u r study, w o u l d be m u c h m o r e amenab le to interfaces w i t h o u t in tegrated focus a n d context . A l t h o u g h subjects c o u l d be expected to be m o r e f a m i l i a r w i t h p a n a n d z o o m n a v i g a t i o n f rom previous exper ience , the difference in p e r f o r m a n c e between p a n a n d z o o m a n d r u b b e r sheet interfaces d i d not decrease s igni f icant ly as subjects b e c a m e m o r e adept in the use o f the interfaces. W e therefore speculate that the exper ience subjects ga ined w i t h r u b b e r sheet n a v i g a t i o n d u r i n g the e x p e r i m e n t m a y not have been sufficient to overcome inherent diff iculties w i t h this f orm of n a v i g a t i o n . F u r t h e r inves t igat ion is r equ ired to d e t e r m i n e whether m o r e exper ience w i t h r u b b e r sheet nav iga t ion c a n enable subjects to improve the ir p e r f o r m a n c e b e y o n d the p l a t e a u observed in this s tudy. In t erms o f subject ive exper ience , p a n a n d z o o m interfaces were rated by sub- , jects as less menta l l y d e m a n d i n g a n d easier to nav igate t h a n their r u b b e r sheet equivalents , regardless of the presence of an overview. T h i s result is consistent w i t h results of prev ious s tudies [21. 25], w h i c h found that n o n - d i s t o r t i o n p a n n i n g interfaces w i t h overviews were preferred to the ir d i s t o r t i o n - b a s e d c o u n t e r p a r t s . W e p o s t u l a t e that , since the p a n a n d z o o m interface w i t h o u t an overview used in o u r s t u d y p r o v i d e d an a l t erna t ive form of c o n t e x t u a l i n f o r m a t i o n (namely H a l o -like arcs ) , subjects d i d not cons ider the lack of an overview to be a l iabi l i ty w h e n Chapter 5. Study 2 63 assessing their perceived menta l d e m a n d a n d ease of nav igat ion . Q u a l i t a t i v e feedback f rom quest ionnaires a n d interviews h igh l ighted three m a i n areas where subjects encountered diff iculty w i t h r u b b e r sheet nav igat ion . F i r s t , subjects r e p o r t e d that the g loba l effect of d is tort ions resul t ing f rom r u b -ber sheet nav igat ion prevented t h e m from forming a stable m e n t a l m o d e l of the dataset , even w h e n a n overview was present. Second , subjects r epor ted b e i n g confused b y the effects of the r u b b e r sheet nav igat ion equivalent o f p a n n i n g , w h i c h was also i m p l e m e n t e d using d i s tor t ion . S o m e subjects s tated that they w o u l d have preferred this o p e r a t i o n to feel like p a n n i n g i n p a n a n d z o o m in -terfaces they have prev ious ly exper ienced, such as G o o g l e M a p s [20]. F i n a l l y , a n u m b e r of subjects identif ied the use of a single fixed focus region in r u b b e r sheet nav igat ion interfaces as a h indrance , a n d suggested that they would have benefitted f rom movable , resizable focus areas, as i m p l e m e n t e d i n the or ig ina l T r e e J u x t a p o s e r a p p l i c a t i o n [37]. T h e s e findings suggest that the constraints imposed on the size a n d locat ion of focus areas in our studies adversely affected the usabi l i ty of the r u b b e r sheet nav igat ion interfaces. O u r or ig ina l rat ionale for i m p o s i n g these constraints was to contro l the level of context , a parameter that was shown to have no significant effect on performance in S t u d y 1. F o r this reason, we r e c o m m e n d that future studies of r u b b e r sheet nav igat ion re lax these constraints as necessary to prov ide usabi l i ty on a p a r wi th c o m p a r i s o n nav igat ion techniques. A possible c o n f o u n d i n g factor i n this s t u d y was the lack of adjus tment to the bes t -per forming levels of context der ived f r o m S t u d y 1 results to account for the m e r g i n g of the two focus regions used in that s t u d y into a single focus region. However , we postu late that the m a g n i t u d e of such a n adjus tment w o u l d have been m i n o r , a n d its i m p a c t negl igible g iven the lack o f significant effect o f level of context on the results of S t u d y 1. 64 Chapter 6 Conclusions and Future Work T h i s thesis presented the first e m p i r i c a l eva luat ions in the i n f o r m a t i o n v i s u a l i z a -t ion l i t erature c o m p a r i n g p a n a n d z o o m a n d r u b b e r sheet n a v i g a t i o n techniques . W e c o n d u c t e d two exper iments , each o f w h i c h involved interfaces i m p l e m e n t -ing J i l l c o m b i n a t i o n s o f these two n a v i g a t i o n techniques w i t h a n d without, an overview. O u r results indicate that p a n a n d z o o m n a v i g a t i o n was s igni f icant ly faster t h a n r u b b e r sheet nav iga t ion , r e q u i r e d fewer n a v i g a t i o n act ions , a n d de-m a n d e d less m e n t a l effort to comple te a topo log ica l c o m p a r i s o n task in a large tree dataset . regardless of the presence of a n overview. W e also found that level of context d i d not s igni f icant ly inf luence p e r f o r m a n c e w i t h e i ther n a v i g a t i o n technique , whi le interface-specific s trategies deve loped b y subjects d i d . A s contro l l ed l a b o r a t o r y exper iments , the studies d e s c r i b e d in this thesis are necessari ly l imi t ed in terms of rea l i sm a n d genera l i zab i l i ty [36]. T h e d e r i v a t i o n of the task e m p l o y e d in the s tudy f rom discuss ions w i th experts in the d o m a i n of phy logenet i c biology, c o m b i n e d wi th the use of a dataset from the same d o m a i n , render this work m o r e ecological ly va l id t h a n most of the studies that have preceded it in the l i terature . However , the degree of r ea l i sm p r o v i d e d b y these factors is t e m p e r e d by the use of n o n - e x p e r t subjects a n d a task that d i d not require a knowledge o f phylogenet ic b io logy . T h e genera l i zab i l i ty of the results of this s t u d y is l ikewise l imi ted by the use o f a single, a lbei t c o m p o u n d , task to represent the var iety of possible t o p o l o g i c a l tasks of interest to phy logene t i c biologists , as well as the usabi l i ty issues w i t h o u r i m p l e m e n t a t i o n of r u b b e r sheet Chapter 6. Conclusions and Future Work 65 nav iga t ion identif ied in the prev ious sect ion. G i v e n these l imi ta t ions , this work s h o u l d be regarded as a s tar t ing po int for further exp lorat ion into the propert ies of p a n a n d z o o m a n d r u b b e r sheet nav igat ion rather t h a n a definitive s tatement about their relative per formance . Several possibil it ies for future studies arise f r o m the results descr ibed in this thesis. C o m p a r i s o n s of p a n a n d z o o m a n d r u b b e r sheet n a v i g a t i o n m e t h o d s us ing other topo log ica l tasks such as those descr ibed in C h a p t e r 3, as well as more general non- topo log ica l tasks a n d different types of datasets , c o u l d y ie ld a n i m p r o v e d u n d e r s t a n d i n g of the relative strengths a n d weaknesses of each nav igat ion technique. F o r instance, a future exper iment c o u l d use a steering task s imi lar to that used b y G u t w i n a n d S k o p i k [22] to de termine whether the benefits of the integrat ion of focus a n d context regions prov ided b y the fisheye views i n that s t u d y ex tend to r u b b e r sheet nav igat ion . T h e c o m p a r i s o n cou ld also be extended to inc lude other variants o n the p a n a n d z o o m a n d focus+context nav igat ion metaphors , such as fisheye views a n d semant ic z o o m i n g . T h e findings f rom S t u d y 1 suggest that nav igat ion strategy c a n have a sig-nif icant i m p a c t on user performance , as well as the poss ibi l i ty of a n interact ion between strategies deve loped b y users a n d affordances p r o v i d e d b y nav igat ion techniques. There fore , another obvious next step w o u l d be to c o n d u c t a sys tem-atic exp lora t ion of the strategies users n a t u r a l l y a d o p t o n their own to complete different tasks w i t h each t y p e of nav iga t ion e x a m i n e d i n our s tudy. It w o u l d be equal ly va luable to examine possibi l i t ies for how the nav igat ion techniques e x a m i n e d i n this s t u d y might be t u n e d to bet ter a c c o m m o d a t e strategies users w o u l d n a t u r a l l y t end to develop w h e n us ing t h e m . F o r instance, future studies cou ld examine whether, as suggested by o u r findings, users can develop z o o m i n g strategies that provide per formance benefits equivalent to those of a n overview, a n d determine whether z o o m i n g enhancements such as speed-dependent auto-m a t i c z o o m i n g [26] c a n further improve per formance w i t h such strategies. T h e need for more t h a n one focus region is another issue that affected the results of the studies descr ibed i n this thesis a n d deserves further invest igat ion. O u r S t u d y 1 showed t h a t the presence of two focus areas i n the or ig ina l versions Chapter 6. Conclusions and Future Work 66 of o u r interfaces increased their complex i ty a n d required subjects to invest too m u c h cognit ive effort into c o o r d i n a t i n g nav igat ion between t h e m . However , a n u m b e r of subjects i n S t u d y 2 suggested that mul t ip l e focus areas w o u l d have p r o v e d beneficial , especial ly i n c o n j u n c t i o n w i t h r u b b e r sheet nav iga t ion . A n e m p i r i c a l e x a m i n a t i o n of the combinat ions of nav iga t ion techniques, tasks, a n d datasets for w h i c h mul t ip le focus areas m a y prove e i ther he lpfu l or a h indrance c o u l d prov ide a va luable a d d i t i o n to the i n f o r m a t i o n v i sua l i za t ion l i terature . F i n a l l y , a logical d irec t ion for cont inu ing this work is to explore how the nav igat ion m e t h o d s e x a m i n e d i n it c o m p a r e when used b y d o m a i n experts in phylogenet ic b io logy i n their work. A s prev ious ly ment ioned , a l though subjects in our s t u d y b e c a m e more proficient in their use of r u b b e r sheet nav igat ion over the course of the exper iment , a n d their l earn ing o c c u r r e d at a faster rate t h a n that of subjects us ing p a n a n d z o o m nav iga t ion , this i m p r o v e m e n t was not sufficient to s ignif icantly decrease the per formance differential between t h e m . It w o u l d therefore be va luable to investigate whether the use of r u b b e r sheet nav igat ion by d o m a i n expert users over an extended p e r i o d of t ime w o u l d result i n this gap decreasing to the po int where per formance w i t h r u b b e r sheet nav iga-t ion is not s ignif icantly different f rom that w i th p a n a n d z o o m nav iga t ion . S u c h an invest igat ion c o u l d also b e extended into other aspects of potent ia l future work we have discussed here, such as task-specific strategies a n d mul t ip l e focus areas, to determine how the usage pat terns of d o m a i n experts differ f r o m those of non-expert users. T h i s work represents a n in i t ia l step i n invest igat ing the propert ies of r u b b e r sheet nav igat ion a n d how it compares to the more establ i shed p a n a n d z o o m nav igat ion p a r a d i g m . T h e eventual goal of these invest igations w o u l d b e the deve lopment of des ign guidel ines of relevance to a wide range of in format ion v i sua l i za t ion appl icat ions t h a t cou ld benefit f r o m i m p r o v e d m e t h o d s for n a v i -ga t ing t h r o u g h large a m o u n t s of d a t a . 67 Bibliography [1] A d o b e C o r p o r a t i o n . A d o b e P h o t o s h o p website , h t t p : / / w w w . a d o b e . c o m / p r o d u c t s / p h o t o s h o p . R e t r i e v e d O c t o b e r 15, 2005. [2] A h l b e r g . C . a n d S h n e i d e r m a n , B . (1994). V i s u a l i n f o r m a t i o n seeking: t ight c o u p l i n g of d y n a m i c query filters w i t h starfield d i sp lays . In CHI '94: Pro-ceedings of the SIGCH1 conference on Human factors in computing systems, pages 313-317, N e w Y o r k , N Y . U S A . A C M Press . [3] B a r t r a m , L . . H o , A . , D i l l . J . , a n d H e n i g m a n , F . (1995). T h e c o n t i n u o u s z o o m : a c o n s t r a i n e d fisheye t echnique for v i e w i n g a n d n a v i g a t i n g large infor-m a t i o n spaces. In UIST '95: Proceedings of the 8th annual ACM symposium, on User interface and software technology, pages 207-215, N e w Y o r k , N Y , U S A . A C M P r e s s . [4] B a u d i s c h , P . , G o o d , N . , a n d S tewart , P . (2001). F o c u s plus context screens: c o m b i n i n g d i s p l a y technology w i t h v i s u a l i z a t i o n techniques. In UIST '01: Proceedings of the 14th annual ACM symposium on User interface software and technology, pages 31-40 , N e w Y o r k , N Y . U S A . A C M Press . [5] B a u d i s c h , P . , L e e , B., a n d H a n n a , L . (2004). F i s h n e t , a fisheye w e b browser w i t h search t e r m p o p o u t s : a c o m p a r a t i v e e v a l u a t i o n w i t h overv iew a n d l inear v iew. In AVI '04: Proceedings of the working conference on Advanced visual interfaces, pages 133-140 , N e w Y o r k , N Y , U S A . A C M Press . [6] B a u d i s c h . P . a n d R o s e n h o l t z . R . (2003). H a l o : a technique for v i s u a l i z i n g off-screen objec t s . In CHI '03: Proceedings of the S1GCH1 conference, on Bibliography 68 Human factors in computing systems, pages 481-488, N e w Y o r k . N Y . U S A . A C M Press . [7] B e d e r s o n , B . B . a n d B o l t m a n , A . (1999). D o e s a n i m a t i o n he lp users b u i l d m e n t a l m a p s of spat ia l i n f o r m a t i o n ? In InfoVis '99: Proceedings of the 1999 IEEE Symposium, on Information Visualization, page 28, W a s h i n g t o n , D C , U S A . I E E E C o m p u t e r Society. [8] B e d e r s o n , B . B . a n d H o l l a n , J . D . (1994). P a d + + : a z o o m i n g g r a p h i c a l interface for e x p l o r i n g a l ternate interface physics . In UIST  !94'- Proceedings of the 7th annual A CM symposium on User interface software and technology, pages 17-26 , New Y o r k , N Y , U S A . A C M Press . [9] B e d e r s o n , B . B . , M e y e r , J . , a n d G o o d , L . (2000). J a z z : an extensible z o o m a b l e user interface g r a p h i c s too lk i t in Java. In UIST '00: Proceedings of the 13th annual ACM symposium on User interface software and technology, pages 171-180 , N e w Y o r k , N Y , U S A . A C M Press . [10] B i e r , E . A . , Stone, M . C , P i e r , K . , B u x t o n , W . , a n d D e R o s e , T . D . (1993). T o o l g l a s s a n d m a g i c lenses: the see - through interface. In S1GGRAPH '93: Proceedings of the 20th annual conference on Computer graphics and inter-active techniques, pages 73-80 , N e w Y o r k , N Y , U S A . ACM Press . [11] B j o r k , S., H o l m q u i s t , L . E . . a n d R e d s t r o m , J . (1999). A framework for f o c u s + c o n t e x t v i sua l i za t ion . In InfoVis '99: Proceedings of the 1999 IEEE Symposium, on Information Visualization, page 53, W a s h i n g t o n , D C , U S A . I E E E C o m p u t e r Society. [12] B o d n a r . A . (2005). A n eva lua t ion o f overviews for large tree n a v i g a t i o n . M a s t e r ' s thesis, T h e U n i v e r s i t y of B r i t i s h C o l u m b i a . [13] C a r d , S. K . , M a c k i n l a y , J . D . , a n d S h n e i d e r m a n , B . , ed i tors (1999). Read-ings in information visualization: using vision to think. M o r g a n K a u f m a n n P u b l i s h e r s Inc.. San F r a n c i s c o . C A . U S A . Bibliography 69 [14] C a r d , S. K . a n d N a t i o n . D . (2002). Degree-of- interest trees: A c o m p o n e n t of a n a t tent ion-reac t ive user interface. In M a r s i c o , M . D . . L e v i a l d i . S . ; a n d P a n i z z i . E . ; ed i tors . Proc. AVI 2002: Advanced Visual Interfaces., T r e n t o , Italy. I S B N = l - 5 8 1 1 3 - 5 3 7 - 8 . [15] C a r r i z o . S. F . (2004). Phy logene t i c trees: an i n f o r m a t i o n v i s u a l i s a t i o n perspect ive . In CRPIT '0J,: Proceedings of the second conference on Asia-Pacific bioinformnlics, pages 315-320 . D a r l i n g h u r s t . A u s t r a l i a . A u s t r a l i a n C o m p u t e r Society. Inc. [16] C h e n . C . a n d C z e r w i n s k i . M . (2000). E m p i r i c a l eva lua t ion of i n f o r m a t i o n v i sua l i za t ions : A n i n t r o d u c t i o n . International Journal of Human-Computer Studies. 53:631-635. [17] C o h e n . J . (1973). E t a - s q u a r e d a n d p a r t i a l e ta - squared ii i c o m m u n i c a t i o n science. Human Communication Research, 28:473-490. [18] F u r n a s . G . W . (1986). G e n e r a l i z e d fisheye views. In CHI '86: Proceedings of the SIGCHI conference on Human factors in computing systems, pages 16-23, N e w Y o r k , N Y , U S A . A C M Press . [19] F u r n a s . G . W . (1997). Effect ive v iew n a v i g a t i o n . In CHI '97: Proceedings of the SIGCHI conference on Human factors in computing systems, pages 367-374, N e w Y o r k , N Y , U S A . A C M P r e s s . [20] G o o g l e , Inc. G o o g l e M a p s website, h t t p : / / m a p s . g o o g l e . c o m . R e t r i e v e d O c t o b e r 15, 2005. [21] G u t w i i i ; C . a n d F e d a k . C . (2004). I n t e r a c t i n g wi th b ig interfaces on s m a l l screens: a c o m p a r i s o n of fisheye. z o o m , a n d p a n n i n g techniques . In Gl '04: Proceedings of the 2004 conference on Graphics interface, pages 1 4 5 -152. S c h o o l of C o m p u t e r Sc ience. U n i v e r s i t y of W a t e r l o o ; W a t e r l o o ; O n t a r i o . C a n a d a . C a n a d i a n H u m a n - C o m p u t e r C o m m u n i c a t i o n s Society. Bibliography 70 [22] G u t w i n . C . a n d S k o p i k , A . (2003). Fisheyes are g o o d for large s teer ing tasks . In CHI '03: Proceedings of the SICCHI conference on Human factors in computing systems, pages 201-208 , N e w Y o r k , N Y , U S A . A C M Press . [23] H a r t , S. G . a n d S t a v e l a n d , L . E . (1988). D e v e l o p m e n t o f N A S A - T L X ( T a s k L o a d Index): results o f e m p i r i c a l a n d theoret ical research. In H a n c o c k ; P . a n d M e s h k a t i , N . , ed i tors , Advances in psychology: human mental workload, pages 139-183 . E l s e v i e r Sc ience , A m s t e r d a m . [24] H o r n b a e k , K . , B e d e r s o n , B . B . , a n d P la i sant , C . (2002). N a v i g a t i o n pat -terns a n d usabi l i ty of z o o m a b l e user interfaces w i th a n d w i t h o u t an overv iew. ACM Trans. Comput.-Hum. Interact.. 9(4):362-389. [25] H o r n b a e k , K . a n d F r o k j a e r , E . (2001). R e a d i n g o f e l ec tronic d o c u m e n t s : the usab i l i ty of l inear, fisheye, a n d overv i ew+deta i l interfaces. In CHI '01: Proceedings of the SIGCHI conference on Human factors in computing sys-tems. [26] Igarash i , T . a n d H i n c k l e y , K . (2000). Speed-dependent a u t o m a t i c z o o m i n g for b r o w s i n g large d o c u m e n t s . In UIST '00: Proceedings of the 13th annual ACM symposium on User interface software and technology, pages 139-148, N e w Y o r k , N Y , U S A . A C M Press . [27] J o h n s o n , J . A . (1995). A c o m p a r i s o n of user interfaces for p a n n i n g on a t o u c h - c o n t r o l l e d display . In CHI '95: Proceedings of the SIGCHI conference on Human factors in computing systems, pages 218-225 , N e w Y o r k . N Y , U S A . A C M P r e s s / A d d i s o n - W e s l e y P u b l i s h i n g C o . [28] J u l , S. a n d F u r n a s , G . W . (1998). C r i t i c a l zones in desert fog: a ids to mul t i s ca l e n a v i g a t i o n . In UIST '98: Proceedings of the 11th annual ACM symposium, on User interface software and technology, pages 97-106 . New-Y o r k , N Y , U S A . A C M Press . Bibliography 71 [29] K e a h e y , T . A . a n d R o b e r t s o n . E . L. (1997). N o n l i n e a r magn i f i ca t ion fields. In InfoVis '97: Proceedings of the 1991 IEEE Symposium on Information Visualization, page 51. W a s h i n g t o n . D C , U S A . I E E E C o m p u t e r Society . [30] K o b s a , A . (2004). U s e r e x p e r i m e n t s w i th tree v i s u a l i z a t i o n sys tems. In In-foVis '04-' Proceedings of the IEEE Symposium on Information Visualization. pages 9 - 1 6 ; W a s h i n g t o n ; D C ; U S A . I E E E C o m p u t e r Society. [31] L a m p i n g . J . , R a o , R... a n d P i r o l l i . P . (1995). A focus+context technique based on h y p e r b o l i c g e o m e t r y for v i sua l i z ing large hierarchies . In CHI '95: Proceedings of the SIGCHI conference on Human factors in computing sys-tems, pages 401-408; N e w Y o r k , N Y , U S A . A C M P r e s s / A d d i s o n - W e s l e y P u b -l i sh ing C o . [32] L a n d a u e r , T . (1997). Behavioral research methods in human-computer in-teraction, pages 203-227 . E l s e v i e r Science, A m s t e r d a m . [33] L a u , K . , R e n s i n k , R . A . , a n d M u n z n e r , T . (2004). P e r c e p t u a l invar iance of nonl inear F o c u s + C o n t e x t t rans format ions . In APGV '04: Proceedings of the 1st Symposium on Applied perception in graphics and visualization, pages 65-72 , N e w Y o r k , N Y , U S A . A C M Press. [34] Lee , 13., P a r r , C . S., C a m p b e l l , D . , a n d B e d e r s o n , B . B . (2004). H o w users interact w i t h b i o d i v e r s i t y in format ion u s i n g T a x o n T r e e . In AVI '04: Proceedings of the working conference on Advanced visual interfaces, pages 320-327, N e w Y o r k , N Y , U S A . A C M Press . [35] L e u n g , Y . K . a n d A p p e r l e y , M . D . (1994). A review a n d t a x o n o m y o f d i s tor t ion-or i ented presentat ion techniques. ACM Transactions on. Computer-Human Interaction. 1 (2):126-160. [36] M c G r a t h , J . E . (1995). M e t h o d o l o g y m a t t e r s : d o i n g research in the be-h a v i o r a l a n d soc ia l sciences, pages 152-169. Bibliography 72 [37] M u n z n e r , T . . G u i m b r e t i e r e , F . , T a s i r a n , S., Z h a n g . L . . a n d Z h o u , Y . (2003). T r e e J u x t a p o s e r : sca lable tree c o m p a r i s o n u s i n g F o c u s + C o n t e x t w i t h g u a r a n -teed v i s ib i l i ty . ACM Transactions on Graphics, 22(3):453-462. [38] N e k r a s o v s k i . D . , B o d n a r . A . , M c G r e n e r e , J . ; G u i m b r e t i e r e , F . , a n d M u n -zner, T . (2005). A n e v a l u a t i o n of p a n & z o o m a n d r u b b e r sheet n a v i g a t i o n wi th a n d wi thout an overv iew. In s u b m i s s i o n . [39] O l d u v a i pro jec t website , h t t p : / / o l d u v a i . s o u r c e f o r g e . n e t / t j / i n d e x . s h t m l . R e t r i e v e d O c t o b e r 15, 2005. [40] P e r l i n , K . a n d F o x , D . (1993). P a d : a n a l t ernat ive a p p r o a c h to the c o m -puter interface. In SIGGRAPH '93: Proceedings of the 20th annual con-ference on Computer graphics and interactive techniques, pages 57-64, N e w Y o r k , N Y , U S A . A C M Press . [41] P l a i s a n t , C . C a r r . D . , a n d S l m e i d e r m a n , B . (1995). Image-browser t a x o n -o m y a n d guide l ines for designers. IEEE Software, 12(2):21-32. [42] P l a i s a n t , C . a n d Fekete , J . - D . Infovis 2003 contest , h t t p : / / w w w . c s . u m d . e d u / h c i l / i v 0 3 c o n t e s t / . R e t r i e v e d O c t o b e r 15, 2005. [43] P l a i s a n t , C , G r o s j e a n , J . , a n d B e d e r s o n , B . B . (2002): Spacetree: S u p -p o r t i n g e x p l o r a t i o n in large node l ink tree, des ign evo lu t ion a n d e m p i r i c a l eva lua t ion . In InfoVis '02: Proceedings of the IEEE Symposium on Infor-mation Visualization, pages 57-64, W a s h i n g t o n . D C , U S A . I E E E C o m p u t e r Society. [44] S a r k a r . M . a n d B r o w n . M . H . (1994). G r a p h i c a l fisheye views. Communi-cations of the ACM, 37(12):73-83. [45] S a r k a r , M . , S n i b b e , S. S., Tversky. O . J . , a n d Reiss. S. P . (1993). S t r e t c h i n g the r u b b e r sheet: a m e t a p h o r for v i e w i n g large layouts on s m a l l screens. In UIST '93: Proceedings of the 6th annual ACM symposium on User interface software and technology, pages 81-91 , N e w Y o r k , N Y , U S A . A C M Press . Bibliography- lS [46] Schaffer, D . . Z u o . Z . . G r e e n b e r g . S.. B a r t r a m , L . , D i l l , J . , D u b s , S., a n d R o s e n i a n , M . (1996). N a v i g a t i n g h ierarch ica l ly c lustered ne tworks t h r o u g h fisheye a n d fu l l - zoom m e t h o d s . A CM Trans. Comput.-Hum. Inter-act... 3(2):162-188. [47] S lack , J . , H i l d e b r a n d , K . , a n d M u n z n e r , T . (2005). P r i s a d : P a r t i t i o n e d r e n d e r i n g i n f r a s t r u c t u r e for stable a c c o r d i o n d r a w i n g . In Info Vis '05: Pro-ceedings of the IEEE Symposium on Information Visualization, pages 4 1 - 4 8 , W a s h i n g t o n , D C , U S A . I E E E C o m p u t e r Society. [48] Spence , R . a n d A p p e r l e y , M . (1982). D a t a base n a v i g a t i o n : A n office e n v i r o n m e n t for the profess ional . Behaviour and Information Technology, l ( l ) : 4 3 - 5 4 . [49] W e h r e n d . S. a n d L e w i s , C . (1990). A p r o b l e m - o r i e n t e d c lass i f icat ion of v i s u a l i z a t i o n techniques . In Vis '90: Proceedings of the 1st conference on Visualization '90, pages 139-143. L o s A l a m i t o s , C A , U S A . I E E E C o m p u t e r Soc iety Press . [50] W o o d r u f f , A . , L a n d a y . J . , a n d S t o n e b r a k e r . M . (1998). C o n s t a n t i n f o r m a -t ion dens i ty in z o o m a b l e interfaces. In AVI '98: Proceedings of the working conference on Advanced visual interfaces, pages 57-65 , N e w Y o r k , N Y , U S A . A C M Press . [51] Zelhveger: P . T . , M a c k i n l a y , J . D . , Good, L . , Stefik, M . , a n d B a u d i s c h ; P. (2003). C i t y l ights: contex tua l views in m i n i m a l space. In CHI '03: Extended abstracts of the SIGCHI conference on Human factors in computing systems. pages 838-839; N e w Y o r k , N Y , U S A . A C M Press . Appendix A Study 1 Training Protocol Appendix A. Study 1 Training Protocol 75 A l l interfaces T h a n k y o u for y o u r wi l l ingness to p a r t i c i p a t e in o u r e x p e r i m e n t . Y o u wi l l be h e l p i n g us eva luate different techniques for v i s u a l i z i n g large datasets . Y o u wi l l be asked to comple te a series o f tasks that invo lve d e t e r m i n i n g re lat ive d is tances in large trees. F i r s t , let's rev iew some concept s that wil l he lp y o u to c o m p l e t e the tasks. Present subjects with paper tests. T h e task y o u wi l l p e r f o r m in this e x p e r i m e n t consists o f d e t e r m i n i n g the topo log ica l d i s tance between a series of m a r k e d nodes in the d i sp layed tree, where topo log ica l d i s tances are m e a s u r e d by the n u m b e r o f b l a c k squares be-tween m a r k e d nodes. R e m e m b e r from the tests that y o u jus t c o m p l e t e d t h a t topo log ica l d i s tance wi l l no t equa l geometr i c d i s tance . W e wi l l now explore the features of the interface y o u wi l l use. R S N - N o O V T h i s interface enables y o u to explore the datase t u s i n g a series o f z o o m i n g a n d p a n n i n g act ions that use the m e t a p h o r of s t re t ch ing a r u b b e r sheet w i t h its borders tacked d o w n . T h e left mouse b u t t o n wi l l al low y o u to d r a g out a box . the contents of w h i c h wil l fill one of the R E D or O R A N G E focus boxes . T h e rest of the tree wi l l t h e n be squished a r o u n d the focus b o x but wi l l r e m a i n vis ible at a l l t imes . Ask participant to try dragging out a box A s y o u are d r a g g i n g out a box . you m a y h o l d d o w n the S H I F T key to ind ica te that y o u w o u l d like this b o x to be the new R E D focus box, or h o l d d o w n the C T R L key to indicate t h a t y o u would like th is box to be the new O R A N G E focus box . N o t e also that the S H I F T key is a b o v e the C T R L key. just like the R E D focus box is above the O R A N G E focus box . Ask participant to try dragging out a box using SHIFT and. CTRL If y o u do not select e i ther the S H I F T or C T R L key, the tool wil l choose which focus box to p lace the contents of y o u r new box based on the p r o x i m a t e l y Appendix A. Study 1 Training Protocol 76 of y o u r n e w l y dragged out box to the ex i s t ing R E D a n d O R A N G E focus boxes. Y o u can z o o m out by d r a g g i n g out a b o x w h i c h is larger t h a n e i ther of the colored focus boxes. Ask participant to try zooming out T h e r ight mouse b u t t o n wi l l al low y o u to p a n h o r i z o n t a l l y a n d ver t i ca l ly wi th in the datase t us ing e i ther a h o r i z o n t a l or ver t i ca l d r a g m o t i o n s , w h i c h wi l l let y o u fine tune y o u r select ion. Ask participant to try panning P Z N - N o O V T h i s interface enables y o u to explore the datase t us ing two v iews w h i c h y o u c a n navigate t h r o u g h a series of p a n a n d z o o m act ions . Show subject paper illustration of two views T h e two deta i l views are i n d e p e n d e n t of one another , so y o u c a n nav igate in one w i t h o u t affecting the o ther . It is also possible to o v e r l a p the two deta i l views, a n d even have one ins ide the o ther . T h e left mouse b u t t o n wi l l al low y o u to d r a g ou t a box which wi l l become the new extent of y o u r deta i l view. Ask participant to try zooming in O n c e y o u are z o o m e d i n , y o u m a y h o l d d o w n the right, m o u s e b u t t o n a n d pan in any d i r e c t i o n . Y o u c a n n o t p a n if y o u are zoomed out ent irely . .447c participant to try panning H o l d i n g d o w n the m i d d l e mouse b u t t o n a n d d r a g g i n g the m o u s e toward y o u will a l low y o u to z o o m out . A s y o u z o o m out . y o u m a y also d r a g the mouse in the o p p o s i t e d irec t ion to z o o m back in . bu t on ly to the extent that y o u first began to z o o m out . Ask participant to try zooming out If a m a r k e d node is not c u r r e n t l y in v iew, an arc wi l l a p p e a r at the b o r d e r of the de ta i l v iew, i n d i c a t i n g the d i rec t ion a n d dis tance f rom y o u r current focus box to the m a r k e d node. T h e arc is par t of a c i r c u l a r r ing that s u r r o u n d s one of the nodes w h i c h is current ly off-screen. T h i s r i n g is just, large e n o u g h to reach Appendix A. Study 1 Training Protocol 77 the b o r d e r region o f the display. T h e co lour of the arc indicates the c o l o u r o f the m a r k e d node it represents . O n c e a m a r k e d node is v i s ib le on screen, the arc w i l l d i s a p p e a r . N o m a r k s wi l l a p p e a r in the overv iew w i n d o w since m a r k e d n o d e s are a lways v is ib le . A r c s are v iew dependent . R S N + O V T h i s interface enables y o u to explore the dataset us ing a series of z o o m i n g a n d p a n n i n g act ions t h a t use the m e t a p h o r of s tre tch ing a r u b b e r sheet w i th its b o r d e r s tacked clown. A separate w i n d o w wil l prov ide y o u wi th an overv iew of the datase t . a n d wi l l not be d i s t o r t e d . T h e left mouse b u t t o n wi l l a l low y o u to d r a g out a b o x , the contents o f w h i c h wi l l fill one of the R E D or O R A N G E focus boxes. T h e rest o f the tree wi l l then be squ i shed a r o u n d the focus box b u t wi l l r emain v is ib le at al l t imes. Ask participant to try dragging out a box A s y o u are d r a g g i n g out a. box . y o u m a y hold d o w n the S H I F T key to ind ica te that y o u w o u l d like this box to be the new R E D focus b o x . or ho ld d o w n the C T R L key to indicate that y o u w o u l d like this box to be the new O R A N G E focus box . N o t e also that, the S H I F T key is above the C T R L key, jus t like the R E D focus box is above the O R A N G E focus box. Ask participant to try dragging out a box using SHIFT and CTRL If y o u do not select either the S H I F T or C T R L key, the tool wil l choose w h i c h focus box to p lace the contents of y o u r new box based on the p r o x i m a t e l y o f y o u r newly dragged out box to the ex i s t ing R E D a n d O R A N G E focus boxes. Y o u c a n z o o m out by d r a g g i n g out a box which is larger t h a n ei ther of the co lored focus boxes. Ask participant to try zooming out T h e r ight mouse b u t t o n wi l l a l low y o u to pan h o r i z o n t a l l y and ver t i ca l ly w i t h i n the dataset u s i n g e i ther a h o r i z o n t a l or vert ica l d r a g mot ions , w h i c h wi l l let. y o u fine tune y o u r select ion. Ask participant to try panning Appendix A. Study 1 Training Protocol 78 A separate smal ler w i n d o w wi l l p r o v i d e y o u wi th an overv iew of the dataset . a n d ind ica te where in the datase t y o u r c u r r e n t focus boxes are. In the overview, the left m o u s e b u t t o n wi l l a l low y o u to d r a g out a box, the contents o f w h i c h wi l l fill one o f the R E D or O R A N G E focus boxes . A s y o u are d r a g g i n g out a box: y o u m a y h o l d d o w n the S H I F T key to indicate that y o u w o u l d like this box to be the new R E D focus box , or h o l d d o w n the C T R L key to indicate t h a t y o u w o u l d l ike this box to be the new O R A N G E focus box . Ask participant to try dragging out a box in the overview using SHIFT and CTRL Y o u m a y also ho ld d o w n the r ight m o u s e b u t t o n while ins ide one of the boxes represen t ing the l oca t ion o f y o u r focus b o x a n d move it to wherever y o u like w i t h i n the b o u n d s of the overv iew u s i n g a series of d r a g ac t ions . Ask participant to try panning in the overview P Z N + O V ' T h i s interface enables y o u to exp lore the datase t us ing two deta i l views w h i c h you can n a v i g a t e t h r o u g h a series of p a n a n d z o o m act ions . Show subject, paper illustration of two views T h e two de ta i l views are i n d e p e n d e n t of one another , so y o u c a n navigate in one w i t h o u t affecting the o ther . It is also possible to o v e r l a p the two deta i l views, a n d even have one inside the o ther . T h e left mouse b u t t o n wi l l al low y o u to d r a g ou t a box which wi l l b e c o m e the new extent of y o u r de ta i l v iew. Ask participant to try zooming in a detail view O n c e y o u are z o o m e d in . y o u m a y h o l d d o w n the right m o u s e b u t t o n a n d p a n in any d i r e c t i o n . Y o u c a n n o t p a n if y o u are zoomed out ent irely . Ask participant to try panning in a detail view H o l d i n g d o w n the m i d d l e m o u s e b u t t o n a n d d r a g g i n g the m o u s e t o w a r d you wi l l al low y o u to z o o m out . A s y o u z o o m out , you may also d r a g the mouse in the o p p o s i t e d irec t ion to z o o m back in , b u t on ly to the extent that you first began to z o o m out . Appendix A. Study 1 Training Protocol 79 Ask participant to try zooming out in a detail view A separate smal l er window wi l l p r o v i d e y o u w i t h a n overv iew o f the dataset . a n d indicate where in the datase t y o u r current de ta i l views are. In the overview, the left mouse b u t t o n wil l allow- y o u to d r a g out a b o x w h i c h wi l l become the new extent of y o u r detai l v iew. A s y o u are d r a g g i n g out a box . y o u m a y ho ld d o w n the S H I F T key to ind ica te t h a t y o u w o u l d like this box to be the new R E D detai l v iew, or ho ld d o w n the C T R L key to ind ica te that y o u w o u l d like this box to be the new O R A N G E deta i l v iew. Ask participant to try zooming in, overview using SHIFT and CTRL N o t e also that the S H I F T key is above the C T R L key. jus t l ike the R E D deta i l view a b o v e the O R A N G E de ta i l view. Y o u m a y also ho ld d o w n the r ight mouse b u t t o n whi le inside one of the boxes represent ing the l o c a t i o n of y o u r deta i l v iew a n d move it to wherever y o u like w i t h i n the b o u n d s o f the overview us ing a series of d r a g act ions . T h e modi f ier keys on ly w o r k in the overview window. Ask participant to try panning in overview using SHIFT and CTRL If a m a r k e d n o d e is not c u r r e n t l y in view, an arc wi l l a p p e a r at the b o r d e r of the deta i l v i ew , ind i ca t ing the d i r e c t i o n a n d d i s tance from y o u r c u r r e n t focus box to the m a r k e d node. T h e arc is p a r t of a c i r c u l a r r i n g that s u r r o u n d s one of the nodes w h i c h is current ly off-screen. T h i s r i n g is jus t large e n o u g h to reach the b o r d e r region of the display . T h e co lour o f the arc indicates the co lour o f the m a r k e d n o d e it represents. O n c e a ' m a r k e d n o d e is v i s ib le o n screen, the arc wil l d i s a p p e a r . N o m a r k s wi l l a p p e a r in the overv iew w i n d o w since m a r k e d nodes are a lways visible. A r c s are v iew dependent . A l l i n te r faces D o y o u have any questions a b o u t this interface? T h e R key c a n be pressed to reset y o u r c u r r e n t view to its in i t ia l s tar tup state. T h e E S C key can be pressed d u r i n g a box d r a g ac t ion to cance l y o u r current Appendix A. Study 1 Training Protocol 80 d r a g . A ques t ion p a n e l at the top o f the screen wi l l d i sp lay a ques t ion w h i c h wi l l require y o u to use the interface to solve. T h e quest ion wi l l ask y o u to c o m p a r e the topo log ica l d i s tances between m a r k e d nodes in the tree. T h e t o p o l o g i c a l d i s tance between m a r k e d nodes wi l l never be equa l . T h e ques t ion wi l l never change , b u t the l o c a t i o n of the m a r k e d nodes wi l l , thus y o u wi l l be requ ired to nav igate a n d exp lore different areas w i t h i n the large tree to answer the ques t ion correct ly . W h e n y o u have discovered the answer , we ask that y o u select the a p p r o p r i a t e check box a n d c l ick o n the s u b m i t b u t t o n . T h i s wil l a l low y o u to m o v e o n t o the next ques t ion . A n ins t ruc t ion pane l at the left of the interface wi l l serve as a r e m i n d e r of interface specific contro ls W e wi l l now ask y o u to comple te a series of t r a i n i n g tasks u s i n g this interface. T h e r e is no t ime l i m i t for c o m p l e t i n g these tasks - we want y o u to take as m u c h t ime as y o u need to ensure that y o u r answer is correct . W e want to e m p h a s i z e that we are e v a l u a t i n g the sys tem a n d not y o u r ab i l i ty to use it. F o r this reason , y o u wi l l receive no feedback as to w h e t h e r y o u r answers for the tasks were correct . A g o o d s trategy for so lv ing the tasks is to d r a w out long h o r i z o n t a l th in boxes. T h i s wi l l h e l p y o u to see the larger tree in more deta i l . Appendix B Study 2 Training Protocol Appendix B. Study 2 Tiaining Protocol 82 A l l i n t e r faces T h a n k y o u for y o u r wi l l ingness to p a r t i c i p a t e in o u r e x p e r i m e n t . Y o u wi l l be h e l p i n g us eva luate different techniques for v i s u a l i z i n g large datasets . Y o u wi l l be asked to c o m p l e t e a series o f tasks that involve d e t e r m i n i n g re lat ive distances in large trees. F i r s t , let's rev iew some concepts that wi l l he lp y o u to complete the tasks. Present subjects with paper tests. T h e task y o u wi l l p e r f o r m in this e x p e r i m e n t consists o f d e t e r m i n i n g the topo log ica l d i s tance between a series of m a r k e d nodes in the d i sp layed tree, where t o p o l o g i c a l distances are m e a s u r e d by the n u m b e r o f b l a c k squares be-tween m a r k e d nodes. R e m e m b e r f rom the tests that y o u jus t c o m p l e t e d that topo log ica l d i s tance wil l not e q u a l geometr ic d i s tance . W e wi l l now explore the features of the interface y o u wi l l use. R S N - N o O V T h i s interface enables y o u to e x p l o r e the datase t us ing a v iew w h i c h you can nav iga te u s i n g p a n a n d z o o m act ions . T h e v iew uses the m e t a p h o r o f s tre tch ing a n d s q u i s h i n g a r u b b e r sheet w i t h its borders tacked d o w n . N o t e that the co lored nodes are vis ible at a l l t imes , even if they are squi shed to the edges of the view. T h e left, mouse b u t t o n wi l l a l low you to d r a g out a box , the contents of which wil l (ill the red box . T h e rest o f the tree wi l l then be squi shed a r o u n d the red box but wil l r e m a i n vis ible at a l l t imes. Ask participant to try dragging out a box. Y o u c a n z o o m out by d r a g g i n g out. a box w h i c h is larger t h a n the red box. Ask participant to try zooming out. T h e r ight mouse b u t t o n wi l l al low you to p a n h o r i z o n t a l l y a n d vert ica l ly w i t h i n the v iew us ing either h o r i z o n t a l or ver t i ca l d r a g m o t i o n s , which wil l let y o u fine tune y o u r selection. Ask participant to try panning. Appendix B. Study 2 Training Protocol 83 Y o u c a n use the co lored nodes as v i s u a l anchors to he lp m a i n t a i n or i en ta t ion whi le p e r f o r m i n g nav iga t ion act ions . A s y o u z o o m or p a n . y o u c a n m o n i t o r the l o c a t i o n a n d size of the co lored nodes , w h i c h wil l give y o u a n idea o f w h a t p a t h to follow a n d h o w m u c h farther y o u have to go. P Z N - N o O V T h i s interface enables y o u to explore the dataset u s i n g a v iew w h i c h y o u can nav igate u s i n g p a n a n d z o o m act ions . T h e left m o u s e b u t t o n will a l low y o u to d r a g out a box . the contents of which wi l l then z o o m to fill the view c o m p l e t e l y . Ask participant to try zooming in. O n c e y o u are z o o m e d in . y o u m a y h o l d d o w n the r ight m o u s e b u t t o n a n d p a n in any d i r e c t i o n . Y o u cannot, p a n if y o u are z o o m e d out entirely. Ask participant to try panning. H o l d i n g d o w n the m i d d l e mouse b u t t o n a n d d r a g g i n g the mouse t o w a r d y o u wi l l a l low y o u to z o o m out . A s y o u z o o m out . y o u m a y also d r a g the mouse in the oppos i t e d i rec t ion to z o o m b a c k in . bu t on ly to the extent that y o u first began to z o o m o u t . Ask participant to try zooming out. If a m a r k e d n o d e is not c u r r e n t l y in v iew, a co lored arc wi l l a p p e a r at the b o r d e r o f the de ta i l v iew, i n d i c a t i n g the d i rec t ion a n d d i s tance f rom y o u r current, focus box to the m a r k e d node. T h e a r c is part, of a c i r c u l a r r i n g t h a t s u r r o u n d s , any m a r k e d n o d e w h i c h is c u r r e n t l y off-screen. T h e co lor of the arc indicates the co lor of the m a r k e d node it represents . O n c e a m a r k e d n o d e is v i s ib le on screen, the arc wi l l d i s a p p e a r . Y o u c a n use the arcs as v isual a n c h o r s to help m a i n t a i n or i en ta t ion o f m a r k e d nodes whi le p e r f o r m i n g nav igat ion ac t ions . A s y o u z o o m out or p a n . y o u can m o n i t o r the shape a n d size of the arc . w h i c h wi l l give y o u an idea of w h a t pa th to follow a n d how m u c h farther y o u have to go. Appendix B. Study 2 Training Protocol 84 R S N + O V T h i s interface enables y o n to explore the dataset u s i n g two views w h i c h y o u can navigate t h r o u g h u s i n g p a n a n d z o o m act ions . T h e larger v i ew wi l l d i sp lay deta i led i n f o r m a t i o n a b o u t parts o f the dataset . T h i s view uses the m e t a p h o r of s t re t ch ing a n d s q u i s h i n g a r u b b e r sheet w i t h its borders tacked d o w n . N o t e that the co lored nodes are vis ible at al l t imes , even if they are squi shed to the edges of this v iew. T h e smal ler v iew wil l p r o v i d e y o u w i t h a n overv iew of the dataset , a n d indicate where in the dataset the deta i l v iew is at a n y given t ime . T h i s view does not use the r u b b e r sheet m e t a p h o r . T h e left m o u s e b u t t o n wil l allow- y o u to d r a g ou t a box, the contents of w h i c h wi l l fill the red b o x . T h e rest of the tree wil l t h e n be squished a r o u n d the red b o x but wi l l r e m a i n vis ible at al l t imes . Ask participant to try zooming in in the detail view. Y o u can z o o m out by d r a g g i n g out a box w h i c h is larger t h a n the red box . Ask participant to try zooming out in the detail view. T h e right m o u s e b u t t o n wi l l al low y o u to p a n hor izonta l ly a n d ver t i ca l ly w i t h i n the v iew u s i n g either a h o r i z o n t a l or ver t i ca l d r a g mot ions , w h i c h wi l l let y o u fine tune y o u r selection. Ask participant to try panning in the detail view. In the smal l er v iew, the left mouse b u t t o n wi l l a l low y o u z o o m into an area b y d r a g g i n g out a box which wil l b e c o m e the new contents of the red box in the larger view. Ask participant to try zooming in overview. Y o u can also h o l d d o w n the r ight mouse b u t t o n whi le inside the red box in the smal ler v iew, a n d move it w i t h i n the v iew u s i n g a series of d r a g act ions . Ask participant to try panning in overview. Y o u can use the co lored nodes as v i sua l anchors to help m a i n t a i n or i en ta t ion whi le p e r f o r m i n g n a v i g a t i o n act ions . A s y o u z o o m or p a n . y o u can m o n i t o r the locat ion a n d size o f the colored nodes , which wi l l give y o u an idea o f what p a t h Appendix B. Study 2 Training Protocol 85 to follow a n d how m u c h farther y o u have to go. P Z N + O V T h i s interface enables y o u to exp lore the datase t us ing two v iews w h i c h y o u c a n navigate t h r o u g h us ing p a n a n d z o o m act ions . T h e larger v i ew wi l l d i sp lay detai led i n f o r m a t i o n a b o u t p a r t s of the dataset . T h e smal ler v iew wi l l p r o v i d e y o u wi th an o v e r v i e w of the dataset . a n d ind ica te where in the datase t the detai l view is at any g iven t ime. T h e left m o u s e b u t t o n wi l l a l low y o u to d r a g ou t a box . the contents of which wi l l then z o o m to fill the larger v iew complete ly . Ask participant to try zooming in the detail view. O n c e y o u are z o o m e d in . y o u m a y h o l d d o w n the right mouse b u t t o n a n d p a n in any d i r e c t i o n . Y o u c a n n o t p a n if y o u are z o o m e d out ent irely . Ask participant to try panning in tl\e detail view. H o l d i n g d o w n the m i d d l e mouse b u t t o n a n d d r a g g i n g the m o u s e t o w a r d y o u wi l l allow y o u to z o o m out . A s y o u z o o m o u t . y o u m a y also d r a g the mouse in the oppos i te d i r e c t i o n to z o o m back in . b u t o n l y to the extent t h a t y o u first began to z o o m o u t . Ask participant to try zooming out in the detail view. In the smal l er v iew, the left m o u s e b u t t o n wi l l al low you z o o m into an area by d r a g g i n g out a b o x w h i c h wi l l b e c o m e the new extent of y o u r de ta i l v iew. Ask participant lo try zooming in overview. Y o u can also h o l d d o w n the r ight mouse b u t t o n while inside the red box in the smal ler v iew, a n d move it w i t h i n the v iew u s i n g a series of d r a g act ions . Ask participant to try panning in overview. If a m a r k e d n o d e is not c u r r e n t l y in v iew, a co lored arc wi l l appea l ' at the b o r d e r of the de l a i l v iew, i n d i c a t i n g the d i r e c t i o n a n d distance f rom y o u r current focus box to the m a r k e d node. T h e arc is p a r t of a c i rcu lar r ing that s u r r o u n d s any m a r k e d n o d e which is c u r r e n t l y off-screen. T h e color of the arc indicates the color of the m a r k e d node it represents . O n c e a m a r k e d node is v is ible on Appendix B. Study 2 Training Protocol 86 screen, the arc wi l l d i s a p p e a r . . Y o u c a n use the arcs as v i sual anchors to he lp m a i n t a i n or i enta t ion of m a r k e d nodes whi l e p e r f o r m i n g nav igat ion act ions . A s y o u z o o m out or p a n , y o u c a n m o n i t o r the shape a n d size of the arc . w h i c h wi l l give y o u an idea of w h a t p a t h to follow a n d how m u c h farther y o u have to go. A l l i n te r faces D o y o u have any quest ions a b o u t this interface? T h e R key can b e pressed to reset y o u r current view to its in i t ia l s t a r t u p state. T h e E S C key c a n be pressed d u r i n g a box d r a g ac t ion to cance l y o u r c u r r e n t d r a g . A l l the contro ls I jus t showed y o u are also l isted at the left of the w i n d o w in case y o u need a r e m i n d e r . A t the top of the w i n d o w is the task y o u wi l l p e r f o r m in this e x p e r i m e n t . Y o u wi l l need to d e t e r m i n e whether the p u r p l e node is topological!}' closer to the b lue node or the green node in the tree. T h e task wi l l never change , but the locat ion of the m a r k e d nodes wil l w i t h each task. Y o u cannot sk ip or go b a c k to p r e v i o u s l y answered quest ions. N o t e that the topo log ica l distances to the b lue n o d e a n d the green n o d e wi l l never be equal , b u t they m a y be close. If it seems as t h o u g h they are e q u a l , p e r f o r m m o r e n a v i g a t i o n , a n d you wi l l d i scover that they are different f r o m each other . N o t e that there is on ly one p a t h between any two nodes in the tree. Y o u c a n use this p e n / p e n c i l a n d sheet of p a p e r to wri te d o w n topo log i -cal d is tances between nodes so that y o u don ' t have to r e m e m b e r t h e m as y o u p e r f o r m i n g the task. W h e n you are ready , select the a p p r o p r i a t e answer a n d cl ick on the s u b m i t b u t t o n . T h i s wi l l a l low y o u to move o n t o the next ques t ion . W e want, to e m p h a s i z e that we are e v a l u a t i n g the s y s t e m a n d not y o u r ab i l i ty Appendix B. Study 2 Training Protocol 87 to use it. F o r this r e a s o n , y o u wil l receive no i n d i c a t i o n of whether y o u r answer is correc t . T h e r e is no t ime l i m i t for c o m p l e t i n g these tasks . T a k e as m u c h t ime as y o u need to ensure t h a t y o u r answer is correc t , b u t do w o r k as efficiently as y o u c a n . R S N - N o O V A g o o d strategy for u s i n g this interface is to d r a w out long th in boxes . T h i s wi l l he lp y o u to see the larger tree in m o r e deta i l . It's often helpful to d r a w l o n g hor izonta l boxes to z o o m into the detai ls o f the dataset , a n d to d r a w l o n g ver t i ca l boxes to e x p a n d areas that are squished ver t i ca l ly . Demonstrate this, then ask participant to do it. A n o t h e r useful s t ra tegy is to reset the interface w h e n y o u have found one of the topolog ica l d i s tances before y o u m o v e onto a n o t h e r dis tance . Demonstrate this, then ask participant to do it. P Z N - N o O V A g o o d strategy for u s i n g this interface is to d r a w o u t long th in boxes . T h i s wi l l he lp y o u to see the larger tree in m o r e deta i l . Demonstrate this. then, ask participant to do it. ( O n c e you have z o o m e d in to the area a r o u n d e i ther the blue or the green node , y o u can count the n u m b e r of nodes on the p a t h t h a t are close to it. T h e n y o u c a n slowly z o o m out a n d , as y o u see m o r e nodes o n the p a t h to the p u r p l e node, a d d them to y o u r count . Demonstrate this, then ask participant to do it. A d d i t i o n a l l y , y o u c a n reset the interface w h e n y o u have found one o f the t o p o l o g i c a l d is tances before y o u move o n t o a n o t h e r d i s tance . R S N + O V A g o o d strategy for u s i n g this interface is to d r a w o u t long th in boxes . T h i s wi l l he lp you to see the larger tree in m o r e de ta i l . It's often helpful to d r a w Appendix B. Study 2 Training Protocol 88 long h o r i z o n t a l boxes to z o o m into the deta i l s o f the dataset . a n d to d r a w l o n g vert ica l boxes to e x p a n d areas that are squ i shed vert ical ly . Demonstrate this, then ask participant to do it. A n o t h e r useful s tra tegy is to first z o o m in to the area a r o u n d ei ther the b lue or the green n o d e us ing the s m a l l view. T h e n y o u can'use e i ther v iew to explore the pa th to the p u r p l e node . N o t e that, y o u c a n count nodes a l o n g the p a t h in either view. If y o u need to m a k e smal l ad jus tment s , you c a n p a n : for larger movements , y o u c a n z o o m in e i ther v iew. Demonstrate this, then ask participant to do it: Y o u c a n also reset the interface when y o u h a v e found one of the t o p o l o g i c a l distances before y o u m o v e o n t o another d i s tance . Demonstrate this, then ask participant to do it. W e s t rong ly suggest y o u use these strategies as y o n are a n s w e r i n g the ques-tions. P Z N + O V A g o o d s trategy for u s i n g this interface is to d r a w out long th in boxes. T h i s wi l l help y o u to see the larger tree in more de ta i l . . Demonstrate this, then ask participant to do it. A n o t h e r useful s trategy is to first, z o o m in to the area a r o u n d ei ther the b lue or the green node us ing the s m a l l view. T h e n y o u can use e i ther v iew to exp lore the p a t h to the p u r p l e node . N o t e that y o u c a n count nodes a l o n g the p a t h in either view. If you need to make smal l ad jus tment s , y o u c a n p a n : for larger movements , y o u can z o o m in e i ther view. Demonstrate this, then ask participant to do it. A d d i t i o n a l l y , y o u can reset the interface w h e n you have found one of the topolog ica l distances before y o u move onto a n o t h e r distance. Demonstrate this, then ask participant to do it. Appendix B. Study 2 Training Protocol 89 A l l i n te r faces W e s trong ly suggest y o u use these strategies as y o u are answer ing the quest ions. Appendix C Study 1 Questionnaires Appendix C. Study 1 Questionnaires 91 UBC T H E UNIVERSITY OF BRITISH COLUMBIA E x p e r i m e n t a l Q u e s t i o n n a i r e E v a l u a t i o n of I n f o r m a t i o n V i s u a l i z a t i o n T e c h n i q u e s Inter face # 1 S u b j e c t # Appendix C. Study 1 Questionnaires 92 P a r t 1 1. A g e G r o u p O 19 a n d u n d e r O 20 - 29 O 30 - 39 O 40 - 49 O 5 0 + 2. G e n d e r O M a l e O F e m a l e I 3 . E d u c a t i o n O S o m e h igh school O C o m p l e t e d h igh school O S o m e pos t - secondary e d u c a t i o n O C o m p l e t e d u n d e r g r a d u a t e degree O S o m e g r a d u a t e or profess ional school O C o m p l e t e d p o s t g r a d u a t e degree . 4. C o m p u t e r U s a g e ( h o u r s p e r w e e k ) : Oo-to O 10 - 20 O 20 - 30 O 30 - 40 O 40 - 50 Q 5 0 + Appendix C. Study 1 Questionnaires 93 P a r t 2 W i t h respec t to the v i s u a l i z a t i o n y o u w o r k e d w i t h , a) p lease i n d i c a t e the extent t o w h i c h y o u agree or d isagree w i t h the f o l l o w i n g s t a t e m e n t s : S D = = S t r o n g l y D isagree D = D i s a g r e e N = N e u t r a l A = A g r e e S A = = S t r o n g l y A g r e e Appendix C. Study 1 Questionnaires 94 I f ound this v i s u a l i z a t i o n to be ef-ficient for c o m p l e t i n g the tasks. O S D O D O N O A O S A N a v i g a t i n g t h r o u g h the d a t a was easy to do . O SD O D O N O A O S A L o c a t i n g c o l o u r e d nodes was easy. O SD O D O N O A O SA 1 found this v i s u a l i z a t i o n to be f rus tra t ing . O SD O D O N O A O S A C o m p a r i n g t o p o l o g i c a l distances between nodes was easy. O S D O D O N O A O S A I f ound it easy to get lost in this v i s u a l i z a t i o n . O SD O D O N O A O S A U s i n g two co loured focus boxes he lped m e to c o m p l e t e the task. O SD O D O N O A O S A B e i n g able to see compressed c o l o u r e d nodes a r o u n d the edges o f the v iew m a d e the task easier. O SD O D O N O A O S A I enjoyed us ing this v i sua l i za t ion . 6 SD O D O N O A O S A b) W h a t p a r t i c u l a r aspect (s ) of th i s v i s u a l i z a t i o n d i d y o u like'! c) W h a t p a r t i c u l a r aspect (s ) of th i s v i s u a l i z a t i o n d i d y o u dislike? Appendix C. Study 1 Questionnaires 95 d) P l e a s e use th i s space t o d e s c r i b e / i l l u s t r a t e any a l te rna t i ve s t rateg ies (other t h a n those y o u were s h o w n at t h e b e g i n n i n g o f t h e e x p e r i m e n t ) tha t y o u be l ieve w o u l d have w o r k e d b e t t e r for y o u . e) P l e a s e use th i s space t o m a k e a n y o t h e r c o m m e n t s about t h e e x p e r i m e n t or the v i s u a l i z a t i o n . T h a n k you for y o u r t i m e ! Appendix C. Study 1 Questionnaires 96 UBC T H E UNIVERSITY OF BRITISH COLUMBIA E x p e r i m e n t a l Q u e s t i o n n a i r e E v a l u a t i o n of I n f o r m a t i o n V i s u a l i z a t i o n T e c h n i q u e s s In ter face # 2 S u b j e c t # Appendix C. Study 1 Questionnaires 97 P a r t 1 1. A g e G r o u p O 19 a n d u n d e r O 20 - 29 O 30 - 39 0 4 0 - 4 9 O50 + 2. G e n d e r O M a l e O F e m a l e 3. E d u c a t i o n O Some h i g h school O C o m p l e t e d h i g h school O S o m e p o s t - s e c o n d a r y e d u c a t i o n O C o m p l e t e d u n d e r g r a d u a t e degree O S o m e g r a d u a t e or profess ional school O C o m p l e t e d p o s t g r a d u a t e degree 4. C o m p u t e r U s a g e (hours p e r week) : O 0 - 1 Q O 1 0 - 20 O 20 - 30 O 30 - 40 O 40 - 50 Q 5 0 + Appendix C. Study 1 Questionnaires 98 P a r t 2 W i t h respect to t h e v i s u a l i z a t i o n y o u w o r k e d w i t h , a) p lease i n d i c a t e t h e extent t o w h i c h y o u agree or d isagree w i t h the f o l l o w i n g s t a t e m e n t s : S D = = S t r o n g l y D isagree D = D i s a g r e e N = N e u t r a l A = A g r e e S A = = S t r o n g l y A g r e e Appendix C. Study 1 Questionnaires 99 I found this visualization to be ef-ficient for completing the tasks. O S D O D O N O A O S A Navigating through the data was easy to do. O S D O D O N O A O S A Locating coloured nodes was easy. 0 SD O D O N O A O S A 1 found this visualization to be frustrating. O S D O D O N O A O S A Comparing topological distances between nodes was easy. O SD O D O N O A O S A I found it easy to get lost in this visualization. O SD O D O N O A O S A Using two coloured focus boxes helped me to complete the task. O SD O D O N O A O S A The coloured arcs made naviga-tion easier. O S D O D 0 N O A O S A I enjoyed using this visualization. O S D O D O N O A O S A b) W h a t p a r t i c u l a r aspect (s ) of th i s v i s u a l i z a t i o n d i d y o u like? c) W h a t p a r t i c u l a r aspect (s ) of th i s v i s u a l i z a t i o n d i d y o u dislike? Appendix C. Study 1 Questionnaires 100 d) P l e a s e use t h i s space to d e s c r i b e / i l l u s t r a t e any a l t e r n a t i v e s t rateg ies (o ther t h a n those y o u were s h o w n at the b e g i n n i n g of t h e e x p e r i m e n t ) tha t y o u be l ieve w o u l d have w o r k e d b e t t e r for y o u . e) P l e a s e use th i s space to m a k e a n y o t h e r c o m m e n t s a b o u t the e x p e r i m e n t or the v i s u a l i z a t i o n . T h a n k y o u for y o u r t i m e ! Appendix C. Study 1 Questionnaires 101 UBC | T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A E x p e r i m e n t a l Q u e s t i o n n a i r e E v a l u a t i o n of I n f o r m a t i o n V i s u a l i z a t i o n T e c h n i q u e s Inter face # 3 S u b j e c t # Appendix C. Study 1 Questionnaires 102 P a r t 1 1. A g e G r o u p O 19 a n d u n d e r O 20 - 29 O 30 - 39 O 40 - 49 O50 + 2. G e n d e r O M a l e O F e m a l e 3 . E d u c a t i o n O S o m e h igh school O C o m p l e t e d h igh school O S o m e pos t - secondary e d u c a t i o n O C o m p l e t e d u n d e r g r a d u a t e degree O S o m e g r a d u a t e or profess ional school O C o m p l e t e d p o s t g r a d u a t e degree 4. C o m p u t e r U s a g e ( h o u r s p e r w e e k ) : O 0 - 10 O 10 - 20 O 20 - 30 O 30 - 40 O 40 - 50 O 50 + Appendix C. Study 1 Questionnaires 103 P a r t 2 W i t h respect to the v i s u a l i z a t i o n y o u w o r k e d w i t h , a) p lease ind ica te the ex tent to w h i c h y o u agree o r d isagree w i t h the f o l l o w i n g s ta tements : S D = = S t r o n g l y D i sag ree D = D isagree N = N e u t r a l A = A g r e e S A = = S t r o n g l y A g r e e Appendix C. Study 1 Questionnaires 104 I f o u n d this v i s u a l i z a t i o n to be ef-ficient for c o m p l e t i n g the tasks. 0 SD O D O N O A O S A N a v i g a t i n g t h r o u g h the d a t a was easy to do . O S D O D O N O A O S A L o c a t i n g co loured nodes was easy. O S D O D O N O A O S A I f o u n d this v i s u a l i z a t i o n to be f r u s t r a t i n g . O S D O D O N O A O S A C o m p a r i n g topo log ica l d is tances between nodes was easy. O S D O D O N O A O S A I f ound it easy to get lost in this v i s u a l i z a t i o n . O SD O D O N O A O S A T h e presence of the smal ler v iew m a d e the task easier. 0 SD O D O N O A O S A U s i n g two co loured focus boxes h e l p e d me to cortiplete the task. O S D O D O N O A O S A B e i n g able to see compressed c o l o u r e d nodes a r o u n d the edges o f the v iew m a d e the task easier. O SD O D O N O A O S A I enjoyed us ing this v i s u a l i z a t i o n . O S D O D O N O A O S A b) W h a t p a r t i c u l a r aspect(s ) o f t h i s v i s u a l i z a t i o n d i d y o u like? c) W h a t p a r t i c u l a r aspect(s ) o f th i s v i s u a l i z a t i o n d i d y o u dislike? Appendix C. Study 1 Questionnaires 105 d) P l e a s e use t h i s space to d e s c r i b e / i l l u s t r a t e any a l t e r n a t i v e s t rateg ies (o ther t h a n those y o u were s h o w n at the b e g i n n i n g of the e x p e r i m e n t ) t h a t y o u bel ieve w o u l d have w o r k e d b e t t e r for y o u . e) P l e a s e use t h i s space to m a k e a n y o ther c o m m e n t s a b o u t the e x p e r i m e n t or t h e v i s u a l i z a t i o n . T h a n k y o u for y o u r t i m e ! Appendix C. Study 1 Questionnaires 106 UBC I T H E UNIVERSITY OF BRITISH COLUMBIA E x p e r i m e n t a l Q u e s t i o n n a i r e E v a l u a t i o n of I n f o r m a t i o n V i s u a l i z a t i o n T e c h n i q u e s Inter face # 4 S u b j e c t # Appendix C. Study 1 Questionnaires 107 P a r t 1 1. A g e G r o u p O 19 and u n d e r O 20 - 29 O 30 - 39 O 40 - 49 O 50 + 2. G e n d e r O M a l e O Female 3 . E d u c a t i o n O S o m e h igh school O C o m p l e t e d h i g h school O S o m e p o s t - s e c o n d a r y e d u c a t i o n O C o m p l e t e d u n d e r g r a d u a t e degree O Some g r a d u a t e or profess ional school O C o m p l e t e d p o s t g r a d u a t e degree 4. C o m p u t e r U s a g e ( h o u r s p e r w e e k ) : Oo - io O 10 - 20 O 20 - 30 O 30 - 40 O 40 - 50 O 5 0 + Appendix C. Study 1 Questionnaires 108 P a r t 2 W i t h respect to the v i s u a l i z a t i o n y o u w o r k e d w i t h , a) please i n d i c a t e the ex tent to w h i c h y o u agree or d isagree w i t h t h e f o l l o w i n g s t a t e m e n t s : S D = = S t r o n g l y D i s a g r e e D = D isagree N = N e u t r a l A = A g r e e S A = = S t r o n g l y A g r e e Appendix C. Study 1 Questionnaires 109 I f ound this v i sua l i za t ion to be ef-ficient for c o m p l e t i n g the tasks. O S D O D O N O A O'SA N a v i g a t i n g t h r o u g h the da ta was easy to do . O SD O D O N O A O S A L o c a t i n g co loured nodes was easy. O S D O D O N O A O S A I f ound this v i s u a l i z a t i o n to be f rus tra t ing . O S D O D O N O A O S A C o m p a r i n g topo log ica l distances between nodes was easy. O S D O D O N O A - O S A I f o u n d it easy to get lost in this v i s u a l i z a t i o n . O S D O D O N O A O S A T h e presence of the smal ler v iew m a d e the task easier. O S D O D O N O A O S A U s i n g two co loured focus boxes h e l p e d me to comple te the task. O S D O D O N O A O S A T h e co loured arcs m a d e n a v i g a -t ion easier. O S D O D O N O A O S A I enjoyed us ing this v i s u a l i z a t i o n . O S D O D O N O A O S A Appendix D Study 2 Questionnaires \ Appendix D. Study 2 Questionnaires 111 UBC T H E UNIVERSITY OF BRITISH COLUMBIA E x p e r i m e n t a l Q u e s t i o n n a i r e E v a l u a t i o n of I n f o r m a t i o n V i s u a l i z a t i o n T e c h n i q u e s In ter face 1 S u b j e c t # Appendix D. Study 2 Questionnaires 112 P a r t 1 1. A g e G r o u p O 19 a n d u n d e r O 20 - 29 O 30 - 39 O 40 - 49 O50 + 2. G e n d e r O M a l e O F e m a l e 3 . E d u c a t i o n O S o m e h igh school O C o m p l e t e d h igh school O S o m e pos t - secondary e d u c a t i o n O C o m p l e t e d u n d e r g r a d u a t e degree O S o m e g r a d u a t e or profess ional school O C o m p l e t e d p o s t g r a d u a t e degree 4 . C o m p u t e r U s a g e ( h o u r s p e r w e e k ) : Oo - I O O 1 0 - 20 O 20 - 30 O 30 - 40 O 40 - 50 O50 + Appendix D. Study 2 Questionnaires 113 P a r t 2 W i t h respect to the v i s u a l i z a t i o n y o u w o r k e d w i t h , a) p lease ind ica te the ex tent t o w h i c h y o u agree o r disagree w i t h t h e f o l l o w i n g s ta tements : S D = = S t r o n g l y D i sag ree D = D isagree N = N e u t r a l A A g r e e S A : = S t r o n g l y A g r e e I found this v i sua l i za t ion to be ef-ficient for c o m p l e t i n g the tasks. O S D O D O N O A O S A N a v i g a t i n g t h r o u g h the d a t a was easy to do . O SD O D O N O A O S A L o c a t i n g co loured nodes was easy. O S D O D O N O A O SA I found this v i s u a l i z a t i o n to be f r u s t r a t i n g . O S D O D O N O A 0 SA C o m p a r i n g topo log ica l distances between nodes was easy. O S D O D O N O A O S A I found it easy to get lost. O S D O D O N O A O S A B e i n g able to see compressed c o l o u r e d nodes a r o u n d the edges o f the view m a d e the task easier. O SD O D O N O A O S A 1 enjoyed us ing this v i s u a l i z a t i o n . O S D O D O N O A O S A Appendix D. Study 2 Questionnaires 114 b) W i t h respect t o the v i s u a l i z a t i o n y o u worked w i t h , please answer the f o l l o w i n g quest ions by m a r k i n g a n ' X ' a long t h e scale bes ide the c o r r e s p o n d i n g q u e s t i o n . How much mental andperceptual activity was MENTAL DEMAM) ~ ~ 1 required to complete the task (e.g., looking, 1 > 1 i 1 j 1 . 1 1 | , i ! i 1 i 1 i 1 i searching, thinking, deciding, calculating, i t i i i i t s i i i i 1 1 1 1 ( 1 1 1 i Hit* ! remembering, etc.)? How much physical activity was required to PHYSICAL CeMAND 1 i 1 i 1 i 1 i I i 1 i I 1 i 1 I 1 I 1 complete the task(e.g., moving the mouse, L I 1 ! ! ,l„,l I i 1 ! ! E ! 1 M 1 I t ] dragging, clicking, pressing keys, etc.)? Lor Hie). | How much time pressure did y o u feel due to the TEMPORAL DEMAND rate or pace at which the tasks or task elements 1 i 1 I 1 i I i ! i 1 i 111 i 1 1 1 1 1 ! occurred? l>ow Hit). | jH ow hard did y o u have to work (m entally and EFFORT physically) to accomplish your level o f 1 i 1 i I i I i 1 i 1 i 1 1 1 1 1 1 1 1 1 performance? Lew llich j How successful do you think y o u were in PERFORMANCE accomplishing the goals of the task set by the 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I I 1 experimenter (or yourself)? I I 1 ! I 1 1 1 1 1 1 1 Cood Poor | How insecure, discouraged, irritated, stressed and annoyed versus secure, gratified, content, relaxed FRUSTRATION l t i l l 1 i I ! i • and complacent did you feel during the task? 1 1 1 ! 1 1 1 1 1 1 1 1 Low 1 1 1 1 1 1 1 I : H*>> | Appendix D. Study 2 Questionnaires c) W h a t p a r t i c u l a r aspect(s) o f t h i s v i s u a l i z a t i o n d i d y o u like? 115 d) W h a t p a r t i c u l a r aspect(s) o f t h i s v i s u a l i z a t i o n d i d y o u dislike? e) P l e a s e use th i s space to d e s c r i b e / i l l u s t r a t e a n y a l t e r n a t i v e s t rateg ies (o ther t h a n those y o u were shown at t h e b e g i n n i n g of the e x p e r i m e n t ) t h a t y o u bel ieve w o u l d have w o r k e d b e t t e r for y o u . f) P l e a s e use th i s space to m a k e any o t h e r c o m m e n t s a b o u t the e x p e r i m e n t or t h e v i s u a l i z a t i o n . T h a n k y o u for y o u r t i m e ! Appendix D. Study 2 Questionnaires 116 UBC T H E UNIVERSITY OF BRITISH COLUMBIA E x p e r i m e n t a l Q u e s t i o n n a i r e E v a l u a t i o n of I n f o r m a t i o n V i s u a l i z a t i o n T e c h n i q u e s Inter face # 2 S u b j e c t # Appendix D. Study 2 Questionnaires 117 P a r t 1 1. A g e G r o u p O 19 a n d u n d e r O 20 - 29 O 30 - 39 O 40 - 49 O50 + 2. G e n d e r O M a l e O Female 3 . E d u c a t i o n O Some h igh school O C o m p l e t e d h igh school O Some p o s t - s e c o n d a r y e d u c a t i o n O C o m p l e t e d u n d e r g r a d u a t e degree O Some g r a d u a t e or profess ional school O C o m p l e t e d p o s t g r a d u a t e degree 4 . C o m p u t e r U s a g e ( h o u r s p e r w e e k ) : Oo- io O 1 0 - 2 0 O 20 - 30 O 30 - 40 O 40 - 50 O50 + Appendix D. Study 2 Questionnaires 118 P a r t 2 W i t h respect to the v i s u a l i z a t i o n y o u w o r k e d w i t h , a) please i n d i c a t e the ex tent to w h i c h y o u agree or d isagree w i t h the fo l low ing s t a t e m e n t s : S D = = S t r o n g l y D i s a g r e e D = D isagree N = N e u t r a l A = A g r e e S A = = S t r o n g l y A g r e e 1 found this v i s u a l i z a t i o n to be ef-ficient for c o m p l e t i n g the tasks. 0 SD O D O N O A O S A N a v i g a t i n g t h r o u g h the d a t a was easy to do . 0 SD O D 0 N O A O S A L o c a t i n g co loured nodes was easy. O SD O D O N O A O S A I found this v i s u a l i z a t i o n to be frus trat ing . 0 SD O D O N O A O S A C o m p a r i n g topo log ica l distances between nodes was easy. O S D O D O N O A O S A I found it easy to get lost. O S D O D O N O A O S A T h e co loured arcs m a d e n a v i g a -t ion easier. O SD O D O N O A O S A 1 enjoyed us ing this v i s u a l i z a t i o n . O S D O D O N O A O S A Appendix D. Study 2 Questionnaires 119 b) W i t h respect to the v i s u a l i z a t i o n y o u w o r k e d w i t h , p lease answer the f o l l o w i n g quest ions b y m a r k i n g an ' X ' a l o n g the scale beside t h e c o r r e s p o n d i n g q u e s t i o n . How much mental and perceptual activity was I ME NT AL DE MAM) requiredto complete the task (e.g., looking, j 1 l I I 1 i t 1 ! ! 1 l 1 i I i ! i 1 i 1 1 searching, thinking, deciding, calculating, \ 1 1 I 1 I 1 I 1 1 1 I F Low ) t i i i i i i t remembering, etc.)? | i PHYSICAL DEMAND 1 1 1 1 1 1 1 1 I 1 1 ! How much physical activity was required to j i 1 i 1 i ! i 1 complete the task(e.g., moving the mouse, I I 1 „ I.J 1 i i LJ...I L.I.. 1 1..I J J H i * dragging, clicking, pressing keys, etc.)? \ Low How much time pressure did y o u feel due to the | TEMPORAL DEMAND rate or pace at which the tasks or task elements j 1 1 1 I 1 1 1 1 1 1 1 I 1 t 1 i 1 i ! i 1 occurred? j i Low H i * ( How hard did y o u have to work (mentally and [ EFFORT physically) to accomplish your level o f j 1 1 1 1 t 1 1 1 1 1 1 1 h h h l i l performance? ] Low How successful do you think y o u were in S PERFORMANCE accomplishing the goals of the task set by the 1 1 1 1 1 1 i 1 1 1 1 1 1 1 1 I 1 1 1 e xp erim enter (or yours elf)? j 11 1 ( 1 1 ! I l l 11 Cood Poor i |How insecure, discouraged, irritated, stressed and! jannoyed versus secure, gratified, content^ relaxed j FRUSTRATION t i l l l 1 t 1 1 i i F (and complacent did y o u feel during the task? j | i 1 1 1 1 1 ) 1 1 1 1 1 1 Low I I 1 1 1 1 1 1 I 1 Hie* i Appendix D. Study 2 Questionnaires c) W h a t p a r t i c u l a r aspect (s ) o f t h i s v i s u a l i z a t i o n d i d y o u like? 120 d) W h a t p a r t i c u l a r aspect (s ) of t h i s v i s u a l i z a t i o n d i d y o u dislike? e) P l e a s e use th i s space t o d e s c r i b e / i l l u s t r a t e any a l t e r n a t i v e s t rateg ies (o ther t h a n those y o u were s h o w n at the b e g i n n i n g of t h e e x p e r i m e n t ) tha t y o u be l ieve w o u l d have w o r k e d b e t t e r for y o u . f) P l e a s e use th is space t o m a k e a n y o t h e r c o m m e n t s a b o u t the e x p e r i m e n t o r the v i s u a l i z a t i o n . T h a n k y o u for your t i m e ! Appendix D. Study 2 Questionnaires 121 \VBC\ T H E UNIVERSITY OF BRITISH COLUMBIA E x p e r i m e n t a l Q u e s t i o n n a i r e E v a l u a t i o n of I n f o r m a t i o n V i s u a l i z a t i o n Techn iques In ter face # 3 S u b j e c t # s Appendix D. Study 2 Questionnaires 122 P a r t 1 1. A g e G r o u p O 19 a n d u n d e r O 20 - 29 O 30 - 39 O 40 - 49 O&0 + 2. G e n d e r O M a l e O F e m a l e 3 . E d u c a t i o n O S o m e h i g h school O C o m p l e t e d h igh school O S o m e pos t - s econdary e d u c a t i o n O C o m p l e t e d u n d e r g r a d u a t e degree O S o m e g r a d u a t e or profess ional school O C o m p l e t e d p o s t g r a d u a t e degree 4. C o m p u t e r U s a g e ( h o u r s p e r w e e k ) : O o - io O 10 - 20 O 20 - 30 O 30 - 40 O 40 - 50 O 5 0 + Appendix D. Study 2 Questionnaires 123 Part 2 W i t h respect to the visualization you worked with, a) please indicate the extent to which you agree or disagree with the following statements: SD = = Strongly Disagree D = Disagree N = Neutral A = Agree S A = = Strongly Agree I found this visualization to be ef-ficient for completing the tasks. O SD O D O N O A O S A Navigating through the data was easy to do. O S D O D O N O A O S A Locating coloured nodes was easy. O S D O D O N O A O S A I found this visualization to be frustrating. O S D O D O N O A O S A Comparing topological distances between nodes was easy. O S D O D O N O A O S A I found it easy to get lost. O S D O D O N O A O S A The presence of the smaller view made the task easier. O SD O D O N O A O S A Being able to see compressed coloured nodes around the edges of the view made the task easier. O SD - O D 0 N O A O S A I enjoyed using this visualization. O S D O D O N O A O S A Appendix D. Study 2 Questionnaires 124 b) W i t h respect to t h e v i s u a l i z a t i o n y o u w o r k e d w i t h , please answer the f o l l o w i n g quest ions b y m a r k i n g an ' X ' a l o n g t h e scale bes ide t h e c o r r e s p o n d i n g q u e s t i o n . 'How much mental and perceptual activity was jrequired to complete the task(e.g.7 looking, MENTAL DEMAND 1 1 1 1 1 f 1 1 I 1 i 1 i I i I isearching, thinking, deciding, calculating, remembering, etc.)? Vow Hist. How much physical activity was required to complete the task(e.g., moving the mouse, dragging, clicking, pressing keys, etc.)? PHYSICAL DEMAND 1 . 1 . I i 1 i 1 i M i l l ! ! ! 1 Urn Hi<* 'How much time pressure did you feel due to the Irate or pace at which the tasks or task elements TEMPORAL DEMAND 1 1 1 I 1 ! 1 I I 1 i 1 i 1 i 1 ! i 1 Joccurred? Low Hid. IHow hard did you have to work (mentally and physically) to accomplish your level of EFFORT 1 i 1 i 1 i I i ! i i 1 i 1 i 1 1 ! 1 performance? Lor Hfeh How successful do you think you were in accomplishing the goals of the task set by the experimenter (or yourself)? PERFORMANCE 1 i 1 i 1 i ! i i i 1 i 1 1 1 Cood Poor iHow insecure, discouraged, irritated, stressed and ]annoyed versus secure, gratified, contend relaxed Sand complacent did you feel during the task? | FRUSTRATION 1 1 1 1 1 1 1 1 1 1 i 1 i 1 i 1 1 1 1 Low Rich Appendix D. Study 2 Questionnaires c) W h a t p a r t i c u l a r aspect (s ) of th is v i s u a l i z a t i o n d i d y o u like? 125 d) W h a t p a r t i c u l a r aspect (s ) o f th is v i s u a l i z a t i o n d i d y o u dislike? e) P l e a s e use t h i s space t o d e s c r i b e / i l l u s t r a t e any a l t e r n a t i v e strategies (o ther t h a n those y o u were s h o w n at the b e g i n n i n g of the e x p e r i m e n t ) t h a t y o u be l ieve w o u l d have w o r k e d b e t t e r for y o u . f ) P lease use th is space t o m a k e any o t h e r c o m m e n t s a b o u t the e x p e r i m e n t or the v i s u a l i z a t i o n . T h a n k y o u for y o u r t i m e ! Appendix D. Study 2 Questionnaires 126 UBC T H E UNIVERSITY OF BRITISH COLUMBIA E x p e r i m e n t a l Q u e s t i o n n a i r e E v a l u a t i o n of I n f o r m a t i o n V i s u a l i z a t i o n T e c h n i q u e s In te r face # 4 S u b j e c t # Appendix D. Study 2 Questionnaires 127 P a r t 1 1. A g e G r o u p O 19 a n d u n d e r O 20 - 29 O 30 - 39 O 40 - 49 O 50 + 2. G e n d e r O M a l e O F e m a l e 3 . E d u c a t i o n O S o m e h igh schoo l O C o m p l e t e d h i g h school "O Some p o s t - s e c o n d a r y e d u c a t i o n O C o m p l e t e d u n d e r g r a d u a t e degree O Some g r a d u a t e or profess ional school O C o m p l e t e d p o s t g r a d u a t e degree 4 . C o m p u t e r U s a g e ( h o u r s p e r w e e k ) : O o - io O 10 - 20 O 20 - 30 O 30 - 40 O 40 - 50 O50 + Appendix D. Study 2 Questionnaires 128 P a r t 2 W i t h respect to the v i s u a l i z a t i o n y o u w o r k e d w i t h , a) p lease i n d i c a t e the ex tent t o w h i c h y o u agree or d isagree w i t h the f o l l o w i n g s t a t e m e n t s : S D z - S t r o n g l y D i s a g r e e D = D isagree N N e u t r a l A = A g r e e S A = = S t r o n g l y A g r e e I found this v i s u a l i z a t i o n to be ef-ficient for c o m p l e t i n g the tasks. O S D O D O N O A O S A N a v i g a t i n g t h r o u g h the d a t a was easy to do . O S D O D O N O A O S A L o c a t i n g co loured nodes was easy. O S D O D O N O A O S A I found this v i s u a l i z a t i o n to be frus trat ing . O S D O D O N O A O S A C o m p a r i n g topo log ica l d i s tances between nodes was easy. O S D O D O N O A O S A I found it easy to get lost. O S D O D O N O A O SA T h e presence of the smal l er view-m a d e the task easier. O S D O D O N O A O S A T h e co loured arcs m a d e n a v i g a -t ion easier. O S D O D 0 N O A O S A 1 enjoyed us ing this v i s u a l i z a t i o n . O S D O D 0 N O A O S A Appendix D. Study 2 Questionnaires 129 b) W i t h respect to the v i s u a l i z a t i o n y o u w o r k e d w i t h , p lease answer the f o l l o w i n g quest ions b y m a r k i n g a n ' X ' a l o n g the scale beside t h e c o r r e s p o n d i n g q u e s t i o n . IHow much mental and perceptual activity was •required to complete the task(e.g., looking, [searching, thinking, deciding, calculating, [remembering, etc)? [How much physical activity was required to Icomplete the task(e.g., moving the mouse, dragging, clicking, pressing keys, etc.)? \ MENTAL DEMAND I 1 t I i 1 1 1 i 1 L How much time pressure did you feel due to the [rate orpace at which the tasks or task elements occurred? \ IHow hard did you have to work (mentally and jphysically) to accomplish your level of iperformance? How successfi.il do you think you were in accomplishing the goals of the task set by the experimenter (or yourself)? u P H l ' a O U . DEMAND 1 I I I I I I I I I TEMPORAL DEMAND 1 I I I I I I I I I ! PERFORMANCE ! i i 1 1 1 1 jHow insecure, discouraged, irritated, stressed and! FRUSTRATION [annoyed versus secure, gratified, content, relaxed j [and complacent did you feel during the task? j L ! ' i 11 I i 11 1 i I M I N I M I M I I I I M I I I I M l M _ L Hid> Appendix D. Study 2 Questionnaires c) W h a t p a r t i c u l a r aspect(s) o f this v i sua l izat ion d i d y o u like? 130 d) W h a t p a r t i c u l a r aspect(s) of this v i sua l izat ion d i d y o u dislike? e) Please use this space to d e s c r i b e / i l l u s t r a t e any a l ternat ive strategies (other t h a n those y o u were shown at the b e g i n n i n g o f the exper iment ) that y o u bel ieve w o u l d have worked b e t t e r for y o u . f) P lease use this space to make any o ther comments a b o u t the e x p e r i m e n t or the v i sua l i za t ion . T h a n k y o u for your t ime! 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0051176/manifest

Comment

Related Items