Dialog Design

1
Dialog Design

• Categories of Dialogs

2
Agenda

• Dialog design
• Command Language
• WIMP - Window, Icon, Menu, Pointer
• Direct manipulation
• Speech/Natural language
• Gesture, pen, multi-touch, VR

3
Dialog Design

• How does a user interact with the interface?

4
Command Languages

• Earliest UI interaction paradigms
• Examples
• MS-DOS shell
• UNIX shell

5
CL Attributes

• Work primarily by recall, not recognition
• Heavy memory load
• Little or nothing is visibleso
• Poor choice for novicesbut...

6
CL Attributes

• Specify commands to operate on current data
  collection
• User only controls initiation
• Single thread of control
• Some other display area needed

7
CL Advantages

• Advantages for experts
• Speed, conciseness
• ls ltretgt
• Can express actions beyond a limited set
• Flags, pipes
• Repetition, extensibility
• Scripts, macros
• Easier implementation, less overhead
• Power
• Wild cards

8
CL Dangers

• With added power, comes added responsibility and
  danger
• UNIX
• rm -r
• Deletes every file that you have, and you cant
  get them back

9
CL Reflection

• Command languages are often maligned (for good
  reason)
• But increased functionality can win out over bad
  UI (e.g., UNIX)
• Try to get both
• Avoid excess functionality (comes at cost)

10
CL Design Goals

• Consistency
• Syntax, order
• Good naming and abbreviations
• Doing your homework in design can help alleviate
  some of the negatives

11
Consistency

• Provide a consistent syntax
• In general Have options and arguments expressed
  the same way everywhere
• UNIX fails here because commands were developed
  by lots of different people at different
  organizations
• No guidelines provided

12
Dialog Order

• English SVO subject verb object
• CL S assumed (you)
• Is VO or OV better? rm file
• file rm
• V dO iO vs. V iO dO
• print file thePrinter
• lpr -PthePrinter file

13
Dialog Order

• Technical issues dictate the choice
• V iO dO
• lpr -PthePrinter file
• The command must parse the arguments
• So the command comes first
• Flags control how to act on the file
• Want to parse all flags before checking files
• e.g. -o outputFile

14
Syntax

• Pick a consistent syntax strategy
• Simple command list
• e.g, vi, minimize keystrokes
• Commands plus arguments
• realistic, can provide keyword parameters
• cp fromfoo tobar
• Commands plus options plus arguments
• what you usually see

15
Terminology

• Keep terminology consistent
• Same concept expressed with same options
• Useful to provide symmetric (congruent) pairings
• forward/backward
• next/prev
• control/meta

16
Example

• vi text editor
• w - forward word
• b - backward word
• Wouldnt f be better for forward?
• f already used
• How about fw and bw?
• Extra keystrokes

17
Ordering

• Keep ordering consistent
• VO seems to be the most natural
• Typically need to pick where options go
• Example
• ln -s file1 file2 (I can never remember)
• Think of cp file1 file2

18
Names and Abbreviations

• Specificity versus Generality
• General words
• More familiar, easier to accept
• Specific (typically better)
• More descriptive, meaningful, distinctive
• (Nonsense does surprisingly well in small set)

19
Abbreviations
20
Abbreviations

• Abbrevs. allow for faster actions
• Expert performance begins to be dominated by
  motor times such as of keystrokes
• Not good idea for novices
• (Allow but dont require)
• IBM CMS and DEC VMS had shortened commands- first
  uniqe characters were enough
• filelist -- f, fi, fil, file, filel, fileli,
  filelis, filelist

21
Picking Good Abbrevs.

• Strategies
• Simple truncation (works best, but conflicts)
• Vowel drop plus truncation (avoid conflicts)
• First and last letters
• First letters of words in a phrase
• Standard abbrev from other contexts
• qty, rm, bldg
• Phonics
• xqt

22
Abbrv Gdlns

• Use single primary rule (with single fallback for
  conflicts)
• Use fallback as little as possible
• Mark use of fallback in documentation
• Let user know primary and secondary rules
• Truncation is good but generates conflicts
• Fixed length is better than variable length
• Dont use abbrevs in system output

23
Agenda

• Dialog design
• Command Language
• WIMP - Window, Icon, Menu, Pointer
• Direct manipulation
• Speech/Natural language
• Gesture, pen, multi-touch, VR

24
WIMP

• Focus Menus, Buttons, Forms
• Predominant interface paradigm now (with some
  direct manipulation added)
• Advantages
• ?

25
Menus

• Key advantages
• 1 keystroke or mouse operation vs. many
• No memorization of commands
• Limited input set

26
Menus

• Many different types
• pop-up
• pull-down
• radio buttons
• pie buttons
• hierarchies

27
Menu Items

• Organization strategies
• Create groups of logically similar items
• Cover all possibilities
• Ensure that items are non-overlapping
• Keep wording concise, understandable

28
Bad Example

• Travel web page links
• Flight page
• 3 Best Itineraries
• Flights Prices
• Timetables
• Fares
• Which do you choose for reservations?

29
Presentation Sequence

• Use natural if available
• Time
• e.g. Breakfast, Lunch, Dinner
• Numeric ordering
• e.g. Point sizes for font
• Size
• Canada-gt BC -gt Surrey

30
Presentation Sequence

• Choices
• Alphabetical
• Group related items
• Frequently used first
• Most important first
• Conventional order (MTWRF)
• Dont change the order on the fly!

31
Presentation Sequence

• User studies
• Novices alpha gt functional gt random
• Experts categorization
• How would you do it in general?

32
Presentation Sequence

• One possible methodology (first-gtlast)
• Natural order (if exists)
• Frequency of use
• Order of use
• Categorical
• Alphabetical
• Dont change dynamically!

33
Agenda

• Dialog design
• Command Language
• WIMP - Window, Icon, Menu, Pointer
• Direct manipulation
• Speech/Natural language
• Gesture, pen, multi-touch, VR

34
Direct Manipulation

• Continuous visibility of the objects and actions
  of interest
• Rapid, incremental actions
• Reversibility of all actions to encourage
  experimentation
• Syntactic correctness of all actionsevery action
  is syntactically legal
• Replacement of command language syntax by direct
  manipulation of object of interest

35
Direct Manipulation

• Examples
• WYSIWYG editors and word processors
• VISICALC - 1st electronic spreadsheet
• CAD
• Desktop metaphor
• Video games

36
DM Syntax

• Typical DM syntax is postfix
• DirectObjects first, Verb second
• In this case, the command completes the utterance
• Enables separate selection syntax
• Indirect objects typically specified before
  direct objects
• e.g. brush size before painting in Photoshop

37
DM Essence

• Representation of reality that can be manipulated
• The user is able to apply intellect directly to
  the task
• Dont have to name things, just touch them
• The tool itself seems to disappear

38
DM Advantages

• Easier to learn remember
• Direct WYSIWYG
• Flexible, easily reversible actions helps reduce
  anxiety in users
• Provides context instant visual feedback
• Exploits human use of visual spatial cues
• Limits types of errors

39
DM Problems

• Waste of screen space (info not all that dense)
• Need to learn meaning of components of visual
  representation
• Visual representation may be misleading
• Mouse ops may be slower than typing
• Not self-explanatory (no prompts)

40
DM Problems

• Not good at
• Repetition
• History keeping (harder)
• Certain tasks (Change all italics to bold)
• Macros harder
• Abstract elements (variables)

41
Direct Manipulation is Locality

• DM Relies on a primary geometric organization
• Items located nearby frequently edited together
• The words in a sentence
• A column of numbers in a spreadsheet
• Less related -gt Less local -gt Less DM!

42
DM Problems

• Not good at
• History keeping (where to localize history?)
• Italics to bold -gt Italicize words all over
• Macros harder
• Abstract elements (variables)

43
What is Directness?

• Related to two things
• Distance
• Gulfs between users goals system image
• Engagement

Hutchins, Hollan, Norman 86
44
Distance Two Gulfs

• Gulf of execution
• Distance between users goals and means of
  achieving them in system
• Does the system allow the user to do what the
  user wants to do?
• Gulf of evaluation
• Amount of effort person must expend to interpret
  system state and judge if intention was achieved
• Can user perceive if progressing favorably?

45
Directness and Distance

• Two types
• Semantic - Relation between what user wants to
  express and what is available in interface
• Can I say what I want (concisely)?
• Articulatory - Relation between meanings of
  expressions and their physical form(s)
• Is the way to perform an action expected and
  clear (appropriate)?

46
Engagement

• Feeling that you are directly manipulating the
  objects of interest
• Promoted by
• Unobtrusive interface
• Minimizing gulfs of execution and evaluation
• Appropriately responsive system

47
What is DM?

• Word? Emacs?
• In end, must characterize direct manipulation by
  feeling of directness and illusion of
  manipulating objects at hand

48
Agenda

• Dialog design
• Command Language
• WIMP - Window, Icon, Menu, Pointer
• Direct manipulation
• Speech/Natural language
• Gesture, pen, multi-touch, VR

49
Speech Input

• Speech synthesis
• Speaker recognition
• Speech recognition
• Natural language understanding

50
English Speech

• Made up of 40 phonemes, 24 consonants and 16
  vowels

51
Speech Synthesis

• Often hear robotic voice
• Store tones, then put them together
• The transition is the difficult thing to do

52
Speaker Recognition

• Tell which person it is (voice print)
• Could be important for monitoring meetings

53
Speech Recognition

• Primarily identifying words
• Improving all the time
• Commercial systems
• IBM/Nuance ViaVoice, Ford Sync ...

54
Recognition Dimensions

• Speaker dependent/independent
• Parametric patters are sensitive to speaker
• With training (dept) can get better
• Vocabulary
• Some are getting 50,000 words
• Isolated word vs. continuous speech
• Continuous where words stop begin
• Typically a pattern match, no context used

55
Recognition Systems

• Typical system has 5 components
• Speech capture device - Has analog -gt digital
  converter
• Digital Signal Processor - Gets word boundaries,
  scales, filters, cuts out extra stuff
• Preprocessed signal storage - Processed speech
  buffered for recognition algorithm
• Reference speech patterns - Stored templates or
  generative speech models for comparisons
• Pattern matching algorithm - Goodness of fit from
  templates/model to users speech

56
Errors

• Systems make four types of errors
• Substitution - one for another
• Rejection - detected, but not recognized
• Insertion - added
• Deletion - not detected
• Which is more common, dangerous?

57
Natural Language Understanding

• Putting meaning to the words
• Input might be speech or could be typed
• Holy grail of Artificial Intelligence problems

58
NL Factors/Terms

• Syntactic
• Grammar or structure
• Prosodic
• Inflection, stress, pitch, timing
• Pragmatic
• Situated context of utterance, location, time
• Semantic
• Meaning of words

59
SR/NLU Advantages

• Easy to learn and remember
• Powerful
• Fast, efficient (not always)
• Little screen real estate

60
SR/NLU Disadvantages

• Doesnt work well enough yet
• Assumes knowledge of problem domain
• Not prompted, like menus
• Requires typing skill (if keyboard)
• Enhancements are invisible
• Expensive to implement

61
Good in Situations

• Hands busy
• Mobility required
• Eyes occupied
• Conditions preclude use of keyboard
• Visual impairment
• Physical limitation