- From: Gérard Talbot <www-style@gtalbot.org>
- Date: Tue, 7 Jun 2011 14:49:40 -0700
- To: "W3C www-style mailing list" <www-style@w3.org>
- Cc: "Daniel Glazman" <daniel.glazman@disruptive-innovations.com>, "Bert Bos" <bert@w3.org>, "Ian Jacobs" <ij@w3.org>
Cascading Style Sheets Standard Boasts Unprecedented Interoperability CSS Test Suite Key to Stable Standard that is Foundation for New Features http://www.w3.org/2011/05/css-pr.html.en " More than 9000 CSS tests have made it easier for designers to create style sheets that work across browsers, and across devices. " Hello all, I want to express a different and much more nuanced opinion regarding CSS 2.1 and its test suite. Maybe we all now have good reasons to celebrate at this point. But maybe we should not celebrate so much ... or not as much as if everything is perfect, complete, impeccable and "done for good". ================== Many if not all of the issues I brought up wrt CSS 2.1 have not been incorporated into the last version (June 7th 2011) of the Proposed Recommendation. I do not know why; maybe it was too late. Maybe it will be added into the errata. 2 examples here: 1st example Section 17.6.1 http://www.w3.org/TR/2011/REC-CSS2-20110607/tables.html#separated-borders still uses an incorrect diagram. I reported the problem, explanations in http://lists.w3.org/Archives/Public/www-style/2011Feb/0652.html along with corrections http://www.gtalbot.org/BrowserBugsSection/css21testsuite/width-table-BC-separated.html 2nd example [CSS 2.1] WD 07 Dec. 2010: section 9.5.2 Introductory subsections and sentences to orient reader http://lists.w3.org/Archives/Public/www-style/2010Dec/0194.html and no significant change happened. Some other people (eg Peter Moulder in http://lists.w3.org/Archives/Public/www-style/2011May/0506.html ) also reported issues and problems with CSS 2.1 which do not appear to have been incorporated in the CSS2-20110607 version. As outrageous as it may seem, I wish CSSWG would consider to release one day a CSS 2.2 spec that would fix all known remaining issues. ================== The weakest point of the CSS 2.1 spec and the weakest point of the CSS 2.1 test suite is that the perspective of web authors are still not met. The CSS 2.1 spec and the CSS 2.1 test suite still look like it was designed by & from & with very technical people "geeks" and for very technical people "geeks". Unfortunately, I wish I had time, energy to develop this opinion here and now but I do not right now. ================== " More than 9000 CSS tests have made it easier for designers to create style sheets that work across browsers, and across devices. " There are at the very least 200 CSS testcases in the latest (RC6 and/or nightly-unstable) version of the test suite which must be tuned, tweaked, corrected, improved here; you can assume that 50 testcases are invalid. You can safely approximate that at least 100 new, additional testcases need to be submitted for various reasons: coverageability, more stringent testcases, more demanding/requiring testcases. CSS 2.1 test suite: Known issues http://wiki.csswg.org/test/css2.1/issues Overall, many testcases (let's say, 400 testcases) among those 9000 CSS tests are not that good, not so good, in particular if/when they are passed by "rusty" and buggy browsers (IE5, IE6) which are known to not support related CSS properties. Several dozens of tests should be redone just because of that. One example among many: http://test.csswg.org/suites/css2.1/nightly-unstable/html4/max-width-058.htm will be passed by browsers which entirely ignore max-width ! http://lists.w3.org/Archives/Public/public-css-testsuite/2011Mar/0040.html There are *many* CSS 2.1 test suite testcases whose expected results are lenient. Out of 9000 testcases, there are many which are basic testcases. We may end up finding more/many browser implementation issues if/when doing more complex testcases. ================== " More than 9000 CSS tests have made it easier for designers to create style sheets that work across browsers, and across devices. " The fact that 9000 CSS tests have been submitted and that the fact the CSS 2.1 spec is official do not, by itself, make CSS coding entirely reliable, perfectly predictable in all browsers. Known compliance scores on the test suite among browsers (IE9 beta1, Firefox 4 beta 6, Chrome 6, Opera 10.70, Safari 5.0.2) were gathered, achieved 9 months ago and ranged from 85% to 97% .. and a bunch of testcases have changed since then, new versions of browsers have been released and new testcases have been submitted, etc. ================== W3C CSS validator has a few bugs. This should and must be stated. Such bugs do not contribute to make the CSS 2.1 spec and test suite perfectly reliable. regards, Gérard +CC: Daniel Glazman, Bert Bos, Ian Jacobs P.S.: Unfortunately, I will be taking a necessary 3 week vacation time starting tomorrow. -- CSS 2.1 Test suite RC6, March 23rd 2011 http://test.csswg.org/suites/css2.1/20110323/html4/toc.html Contributions to CSS 2.1 test suite http://www.gtalbot.org/BrowserBugsSection/css21testsuite/ Web authors' contributions to CSS 2.1 test suite http://www.gtalbot.org/BrowserBugsSection/css21testsuite/web-authors-contributions-css21-testsuite.html
Received on Tuesday, 7 June 2011 21:50:12 UTC