SYMM Working Group
- SYMM 2.0 Working Group chartered late 1998 =96 shortly after original WG=
finished SMIL 1.0
- SMIL: Synchronized Multimedia Integration Language -- language for=
choreographing multiple media objects into multimedia presentation
- Work items:
- Enhance SMIL (Animation, Transitions, improved Timing model)
- Modularize SMIL
- Produce HTML integration of timing
- Predates Candidate Recommendation, but original charter includes=
requirement for two interoperable implementations
Branding of W3C Specifications
- Early discussion: What does "SMIL Conformant" mean?
- No guideline to follow at the time: tried to follow XHTML working group=
precedent, but that was a moving target
- Came up with two levels of conformance: "SMIL Integration Set=
Conformance" and "SMIL Host Language Conformance"
Manageability of the Specification
- SMIL 1.0 was ~30 page spec
- Criticism: too terse
- SMIL 2.0 is currently ~300-500 spec
- Not terse anymore :-)
- Specification often difficult to track and scope changes while under=
development
- Consistency difficult to ensure
- Normative vs. Informative Guidelines evolved
SMIL 2.0 Interop testing
- Test suite includes hundreds of test cases
- Some module areas (e.g. Animation) include valid tests for three=
different integrations (SVG, SMIL Language, XHTML+SMIL)
- Initial test cases mainly authored by module authors, with QA=
departments filling in holes
- QA <-> Working Group interaction
What went well/what=92s going well
- Quality and quantity of tests coming together
- Case update process moves pretty swiftly
- Test results largely complete on time
- Outstanding effort from external resources (clearly, more than=
"20%" being contributed by many working group members and even=
non-members)
- Testing is resulting in important specification clarifications
Issues with testing compliance (slide 1 of 2)
- Lack of official infrastructure (SMIL-specific infrastructure provided=
by Wo Chang at NIST)
- Result tracking
- Shared development of test cases (CVS? Web-interface)
- Issue resolution (bug tracking system)
- Authoring standards/tools
- Numbering/naming inconsistencies
- Commenting standards
- Lack of public domain media
Issues with testing compliance (slide 2 of 2)
- Granularity of testing
- We decided tests not geared toward rigorous conformance, so much as=
proof of implemenation experience
- Explicitly strove to test one feature at a time (excluded complex=
interactions) ("Basic Effectivity")
- Result: test suite meets charter goals, but probably doesn=92t=
make a comprehensive conformance test
- Conformance of modules vs. profiles
- Lots of jargon from WG members sailing over heads of non-WG=
contributors