Evaluating Scalable Vector Graphics for use in Software Visualisation

1
(No Transcript)
2
Evaluating Scalable Vector Graphics for use in
Software Visualisation

• Matthew Duignan, Robert Biddle, Ewan Tempero
• Victoria University of Wellington
• University of Auckland
• Matthew_at_duignan.net

3
Software Visualisation over the Web with VARE

• The Visualisation Architecture for REuse
• Tool support for software reuse
• Visualisation as documentation
• Allow programmers to find, test drive, and
  visualise code via the Web

4
(No Transcript)
5
Deploying visualisations?

• VARE is in need of a visualisation medium
• Complex visualisations need to be deployed over
  the web
• Need to be created programmatically
• Must integrate with Web standards
• HTML
• XML SOAP, PAL

6
What is SVG?

• Standard for Scalable Vector Graphics
• Similar to Macromedia Flash
• W3C Recommendation
• Open standard
• XML Based
• Interactive (Scriptable)

7

• ltcircle
• cx100
• cy100
• r50
• stylestrokeblue
• fillred
• stroke-width5/gt

8
Other features of SVG

• Delivered via Web-browsers
• Plug-ins (Adobe, KDE, Corel)
• Native (Mozilla)
• Works well with other web technologies
• HTML, ECMAScript, XML
• High quality graphics
• Filters, Fonts, Colours, Gradients, Patterns,
  Curves, Animation and more
• Structured
• Symbols, Groups

9
(No Transcript)
10
Software visualisation evaluation model

• But will SVG meet VAREs needs?
• SVG must be evaluated
• Build a model as basis of evauation
• To help insure evaluation is complete (as can be)
• To allow comparison of other media (Flash, Java
  etc.)

11
Evaluation model foundation

• Information Visualisation
• Build from past work cataloguing visualisation
  techniques (Bertin, Card, Mackinlay, Shneiderman,
  Ware)
• Software Visualisation
• Taxonomies of Software Visualisation What are
  the types of visualisations used in software
  visualisation?
• Higher level capabilities
• How well does the technology support the
  developer?

12
Evaluation model structure

• Basic information visualisation capabilities
• Graphical capability
• The spatial substrate
• Marks and their properties
• Temporal encoding
• Interaction
• Performance
• Software visualisation-specific and higher level
  capabilities
• Integration
• Higher level capabilities

13
SVG evaluation 1

• Graphical capabilities are good
• Although not so good on the spatial substrate
• Capable of complex interaction via scripting
• But not without scripting
• Acceptable performance
• But large complex interactive displays do slow
  down in current implementations

14
(No Transcript)
15
SVG evaluation 2

• Excellent integration and leverage of Web
  technologies
• Disapointing lack of higher level capabilities
• Low level graphics descriptions

16
Points of interest 1

• SVG Creation
• Hand authoring
• Time consuming and painful.
• XSLT
• limited use in implementing complex layout
  algorithms, can be used in visualisation creation
  pipeline
• Programmatic generation
• Use standard XML, or tailored SVG libraries
• Drawing tools
• Professional vector graphics tools already allow
  the creation of complex SVG (e.g. Adobe
  Illustrator)
• CASE tools

17
(No Transcript)
18
Points of interest 2.

• Streaming SVG
• Not designed for streaming so no real-time
  visualisations
• The importance of scripting
• Entirely dependent on scripting for non-trivial
  applications
• No standardised scripting language
• No layout constraints
• Simple tasks made complex

19
Alternatives to SVG

• Macromedia Flash
• VRML and X3D
• Java

20
Conclusions

• SVG is a valid medium for deploying simple
  software visualisations over the Web
• XML and W3C standards base insures good tool
  support
• Complex software visualisation applications are
  not well supported by SVG

21
Current and future work

• Current
• Development of a ECMAScript library to support
  SVG software visualisation development
• Future
• Extend Model
• Evaluate other media with Model

22
Questions?
Contact Matthew Duignan, Matthew_at_duignan.net