- From: Al Gilman <asgilman@iamdigex.net>
- Date: Thu, 12 Jul 2001 13:22:06 -0400
- To: w3c-wai-ua@w3.org
At 03:24 PM 2001-07-10 , Jon Gunderson wrote: > 8.Issue 517: Need clarification of issue (Al Gilman or SVG people) > AG:: I can't replace the SVG group or make any definitive statements. So let's make this Al and SVG people, not 'or.' But here is my impression of what is going on and a possible next step toward resolving this one. I think that this may involve a miscommunication about the granularity of content scopes that the required control of play is supposed to apply to. I suspect that the SVG group interpreted 'elements' as XML elements such as 'g' elements which take on the role of dependend time containers in an animated SVG. It is in this context that they objected. I think the UA working group meant that these controls should be available for separately playable media objects, not for every 'g' element involved in the construction of an SVG animation, even if in the timing structure that element is a time container. It may well be that there are no independently playable components below the level of the "root time container" in this case. It would seem to be a misunderstanding of the scopes of content over which these controls would be exercised as a unit. See my remarks on issue 516 -- that may be clearer. Subissue a. In 2.3 I believe we set the precedent that user on/off control was only required for the substructures that actually play a role in the content control functions of the format. Each chunk of conditional content that is suject to one condition is treated as a whole. Finer subdivisions unconditionally included in that substructure are not items to be separately, interactively controlled under 2.3. In 4.4 and 4.5 I think that what the user needs to have this control over are at least the independently playable [multi-] media presentations as grouped into synchronous presentations by the author, with the possible extension to independently playable substructures that may be isolated pursuant to 2.3. The latter takes care of playing just one audio track, or just one video track in isolation as a relatively extreme but feasible recourse to try to understand a SMIL presentation when operating under duress of some sort or another. I think we should informally approach the SVG people with this perception of a possible miscommunication, and if this is in fact an important clarification, then we take a fresh look at do we perhaps have some feasible path to a mutually agreeable resolution. Al PS: Ian has repeatedly asked, what is the technicality that the UA group didn't understand. I don't really think it is a 'technicality,' quite. It is more like a significant difference in performance. In SMIL proper, one has a relatively small number of relatively strongly-encapsulated constituents. Many of these things it makes sense to play individually. In an SVG animation, although it uses the same language constructs to build its time relationships, one finds a relatively large number of relatively weakly-encapsulated constituents. One doesn't that much know where in the 'g' tree to stop. And 'playing' just one makes vanishingly little sense. We had thought a lot about the SMIL case and hadn't thought much about the animated SVG case. So what we said can be interpreted (and they did) as an overly-demanding descent into the bowels of an animation to perform play actions that don't come naturally. That isn't really what we meant.
Received on Thursday, 12 July 2001 13:11:48 UTC