Modelling Languages: (mostly) Concrete (Visual) Syntax Hans - - PowerPoint PPT Presentation

http://msdl.cs.mcgill.ca/

Hans Vangheluwe

Modelling Languages: (mostly) Concrete (Visual) Syntax

Modelling Languages/Formalisms Syntax and Semantics

3

Model of Graph

Modelling Languages/Formalisms Syntax and Semantics

Textual Languages

Textual Languages

“this sentence is very short”

• Individual letters in an alphabet
• Combined into words
• Combined into sentences in a language
• Valid letters in words specified by reg

egular exp xpressions

• Valid words in a language specified by a grammar
• letters/words are combined by “is to the right of”

Textual Languages

syntax-directed editor (textual concrete syntax)

Visual Languages

syntax-directed editor (visual concrete syntax)

Visual Languages

Plex

Visual Languages

Grap aph

Visual Languages

Co Connecti tion T Type pes

Visual Languages

Iconic

Visual Languages

Box

Visual Languages

Visu sual L Langu guage age Cl Class sses

Visual Languages

Hy Hybrid id L Langu guages

Visual Languages

Syntax-dir irected V d Visu sual E al Editors: s: model l be behaviour

Visual Languages

Syntax-dir irected V d Visu sual E al Editors: s: model l be behaviour

Visual Languages

Generate te S Syntax tax-di directe ted V Vis isual l Edi dito tors

Visual Languages

Syntax-dir irected V d Visu sual E al Editors: s: freehand d (early stages of multi-domain project)

Visual Languages

20

Different M t Medi dia: : Gest stural I l Inte terac acti tion, S Sound, ...

Visual Languages

Introducti tion

• Visual notations pre-date textual ones
• Visual notations are important for

Modelling and Software Engineering

• Humans are excellent pattern

recognizers

• Need cognitively efficient and effective

notations. Cognitive effectiveness = speed, ease and accuracy with which a representation can be processed by the human mind

``Physics'' of Notations

a DSVL @ Lascaux

Introducti tion/R /Rati ationale le

Visual notations are often introduced without underlying theory or rationale

``Physics'' of Notations

Many visual notations for same concepts. No rigorous way to co comp mpare re effectiveness and hence no clear design goal.

Maryam M. Maleki, Robert F. Woodbury, Rhys Goldstein, Simon Breslav, Azam Khan. Designing DEVS visual interfaces for end-user programmers. Simulation 91(8): 715-734 (2015)

Co Communicati tion T Theory

``Physics'' of Notations

Encodi ding: 8 g: 8 v vis isual v variabl bles t s to (gr (graph phically) ) encode in information

``Physics'' of Notations

Decodin ing

automatic, fast, parallel slow, large effort, sequential Appropriate notations »

• ffload some of the burden from cognitive to perceptual

Note: “dual channel theory”: auditory/verbal channel and visual/pictorial channel are processed in parallel

Richard E. Mayer, Roxana Moreno. Nine Ways to Reduce Cognitive Load in Multimedia Learning. Educational Psychologist, 38(1), 43–5. 2003. ``Physics'' of Notations

Princip iple les f for D Desi sign gning g Effic fficient an t and E d Effecti tive V Visu sual N Nota tations

``Physics'' of Notations

Semioti tic Cla Clarity (semiotics = study of signs and sign processes)

1- 1-to-1

``Physics'' of Notations

Percept ptual D Disc scrimin inabi bili lity

``Physics'' of Notations

``Physics'' of Notations

Junaed Sattar, Gregory Dudek. Reducing Uncertainty in Human-Robot Interaction: A Cost Analysis Approach. ISER 2010: 81-95.

Percept ptual D Disc scrimin inabi bili lity

should be easy to distinguish visual symbols ability to distinguish is determined by visual distance larger visual distance » faster, more accurate recognition

• number of visual variables on which they differ and

the magnitude of the differences

• shape is the main visual variable

13

``Physics'' of Notations

Percept ptual D Disc scrimin inabi bili lity

Software Engineering notations mostly use rectangle variants Use re redundant visual encoding to in incr crease dis istance ce (e.g., textual + visual)

``Physics'' of Notations

Semantic T Tran anspa parency

The me meanin ing of a symbol can be in inferre rred from its appearance (intuitive) Symbols can be:

``Physics'' of Notations

Semantic T Tran anspa parency: : semantically immediate symbols

``Physics'' of Notations

Semantic T Tran anspa parency

``Physics'' of Notations

Semantic T Tran anspa parency

The me meanin ing of a symbol can be in inferre rred from its appearance (intuitive) Symbols can be:

• Semantically Immediate
• Semantically Opaque

Software Engineering notations are usually abstract (non-intuitive)

``Physics'' of Notations

``Physics'' of Notations

Seman anti tic T Tran anspar sparency: : semantically perverse symbols

Semantic T Tran anspa parency

The me meanin ing of a symbol can be in inferre rred from its appearance (intuitive) Symbols can be:

• Semantically Immediate
• Semantically Opaque
• Semantically Perverse

Domain-specific icons and visual arrangement should be intuitive

``Physics'' of Notations

Co Compl plexity m management (# (# ele lements ts in in di diagr gram » c cognit itiv ive o

• verlo

load) d)

``Physics'' of Notations

Mo Modu dulariz izati tion/Hie /Hierarchy

``Physics'' of Notations

Co Cognit itiv ive I Inte tegrati tion (d (different t notatio ions) s)

• Conceptual integration (cohere

rent mental model)

• Enable navig

igatio ion and transit itio ion between notations

``Physics'' of Notations

Visu sual E Expr pressi siveness

Number of visual variables used (UML, mostly shape, no colour) 8 degrees of visual freedom (0 = non-visual – 8 = visually saturated)

``Physics'' of Notations

Visu sual E Expr pressi siveness

Different visual variables have dif iffere rent ca capaci city to encode information

``Physics'' of Notations

Dual l Encodin ing

Combine Textual and Vis isual Supple lement rather than duplicate (e.g., multiplicity values)

``Physics'' of Notations

Rein inforce rce meaning

Grap aphic ic E Economy

• Not too many symbols. If many, provide le

legend

• Limit on human discrimination capability (6 levels per variable)
• Upper limit on graphic complexity

How?

``Physics'' of Notations

Co Cognit itiv ive F Fit

Adapt choice of visual notation to

• Task
• Audience (novices vs. experts)

Adaptation may be dynamic (“learn” about Task/User proficiency) Representation medium matters

``Physics'' of Notations

Interac actio ions a among g pr prin incipl ples

``Physics'' of Notations