- From: Nick Kew <nick@webthing.com>
- Date: Sun, 30 Sep 2007 09:53:30 +0100
- To: www-validator@w3.org
Aaargh! Why doesn't this list give us a reply-to? Just accidentally sent this to Olivier instead of the list :-( On Sun, 30 Sep 2007 16:27:10 +0900 Olivier Thereaux <ot@w3.org> wrote: > Only few of the warnings are in the "this is strictly speaking OK but > very likely to cause problem" realm, mostly they are "I had to use > some guessing, and if the guessing was wrong, it may twist the > results, so please double-check". Um, shorttags warnings as implemented at valet and htmlhelp are very firmly in the "will cause problems" realm. Only if you select "fussy" mode at valet do you also get warnings of things that are likely not a problem. Happy to look up and repost details, but I'm sure they're on record at lists.w3.org from the days when I was more active in this field. > > e.g: > http://qa-dev.w3.org/wmvs/HEAD/check?uri=http://qa-dev.w3.org/wmvs/HEAD/dev/tests/sgml_customdtd.html The warning there has nothing to do with shorttags. And people who use custom DTDs are a different audience to those who go away in disgust when they discover their obviously-wrong typo that the validator ("ugh, how useless and buggy" had permitted. Bear in mind users who post here and get pointed to an explanation are sure to be a tiny fraction of those who abandon the validator as useless when they make the same discovery. > For these reasons, I think we should be extremely careful with the way > we make warnings prominent on otherwise valid/conforming documents. That's presupposing you make no distinction between different types of "warning". -- Nick Kew Application Development with Apache - the Apache Modules Book http://www.apachetutor.org/ -- Nick Kew Application Development with Apache - the Apache Modules Book http://www.apachetutor.org/
Received on Sunday, 30 September 2007 08:53:45 UTC