- From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
- Date: Fri, 19 Nov 2010 11:47:46 +0100
- To: www-svg@w3.org
Hello www-svg, because there seems to be currently some work on the SVG1.1 test suite on http://dev.w3.org/SVG/profiles/1.1F2/test/svg/animate-elem-82-t.svg this might help to get this file in a somehow usable state: third test (paced translation): As already discussed for SVG tiny 1.2 the chosen values are not useful to test paced behaviour, because with these values for many distance functions the result is the same, what means, that the test is not sensitive on implementation bugs. Better use values like -40,40;40,-20;40,40 fifth test (paced scale): It is known, that the text in http://www.w3.org/Graphics/SVG/Test/20080912/htmlObjectHarness/animate-elem-82-t.html does not fit to the SVG tiny 1.2 definition, therefore this has to be fixed anyway. For SVG 1.1 (first edition) the intended method to determine a distance was wrong and was therefore removed in the second edition. What remains is the definition of calcMode paced and the description in SVG tiny 1.2 for paced scaling, that fits to the definition of calcMode paced, therefore this is the only meaningful method to calculate the distance. Therefore the distances are 2 and sqrt(5). This means scale(3,2) is reached at (2/(2+sqrt(5)))*3s = 1.4164...s and not at 1.5s as mentioned in the description. A longer active duration or other values might be better to see deviations more prominent - additionally it can be pretty useful, to provide red shapes below with animations using values-lists and keyTimes-lists or a group of animations to get the same timing for comparison - this applies for the other subtests as well of course. test seven (paced rotate): Because there is no meaningful distance function for rotation with a changed rotation center, there was a compromise for SVG tiny 1.2 to use only the angle for the distance function to get at least something meaningful, if the rotation center is not changed. However the description both in the SVG tiny 1.2 test suite and that in the SVG 1.1 second edition test suite is still wrong (does not fit to calcMode paced or the compromise formula for SVG tiny 1.2). Obviously rotate(45,-15,-20) is reached at 0.75s for the SVG tiny 1.2 compromise. I think, at 1.5s it is rotate(110,0,3.3333333333) (needs to be checked however). test eight (linear rotate): descriptions seems to be forgotten. (If I have it in mind correctly, I provided already an advanced fixed version of this test, while I was a member of the SVG-group - well, seems to be lost ...) Good luck with fixing this test ;o) Olaf
Received on Friday, 19 November 2010 10:48:25 UTC