- From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
- Date: Fri, 22 Jan 2010 11:55:24 +0100
- To: www-svg@w3.org
The relevance of attributeType has a long discussion history ;o) In doubt for details SVG animation refers to SMIL and this explains in the sandwich model, what applies. Historically this was maybe intended to be simpler, however when I was a member of the SVG group, I tried to clarify this and with some discussion of the issue with the SYMM group they confirmed again, that it is of some relevance in the sandwich model and the interaction of animation of XML presentation attributes, CSS stylesheet properties and animation of CSS properties results in a complex sandwich of priorities. Maybe this could have been solved much simpler and more friendly for authors, when SMIL animation appeared, for example with some pseudo namespace for CSS, but this did not happen and I think, now it is a little bit late to change this (maybe unfortunately). Apart from a few errors the results with Opera for my group of related tests are indeed the best, what means, that some versions of Opera get more tests right than all other tested viewers. I cannot see a problem with this: > Animation of presentation attributes is equivalent to animating the > corresponding property. Thus, the same effect occurs from animating > the presentation attribute with attributeType="XML" as occurs with > animating the corresponding property with attributeType="CSS". Of course, there is itself no different effect, if it is interpreted correctly, what 'effect' means). Obviously the visible result for an animated attribute or property depends on many other things, on the underlying value and priorities of different animations, stylesheets etc. The animation function (See SMIL) is the same in the sense, that the computations are the same, not necessarily the final presentation result for a real graphical object at some specific time. Finally, directly before presentation on has to join together all these applicable presentation attributes, properties, stylesheets and animation functions in a proper order with the right priorities to be able to present the graphical object. In my tests and discussions it turned out, that this is often implemented wrong (for example for additive, cumulative behaviour and especially for to-animations) already without the additional problem of the attributeType, therefore it is no surprise, that one finds even more errors in current implementations, if the situation is complex enough, that attributeType becomes relevant ;o) If one manages to isolate a simple enough sample, the visible result can be the same. Of course, if there are other dependencies and priorities, this can be quite different, this happens with and without attributeType. And what applies is defined in the sandwich model in SMIL (note that there are minor clarifications in SMIL3 due to some discussion of this nasty problem). Olaf
Received on Friday, 22 January 2010 11:05:28 UTC