- From: Jukka K. Korpela <jkorpela@cs.tut.fi>
- Date: Fri, 31 Oct 2003 09:03:13 +0200 (EET)
- To: www-validator@w3.org
On Thu, 30 Oct 2003, Nick Kew wrote: > Why is it "better" to put in extra unproductive work, and use a > nonstandard DTD, when the object of the exercise is to conform with > a W3C DTD? I don't know the objects of everyone's exercises, but here the announced objective was to get more informative diagnostic messages. For that purpose, a suitable DTD will help. By the way, there is no such thing as a nonstandard DTD, except in the loose sense of something that is meant to be a DTD but does not conform to the applicable standard, namely the SGML standard. > And how exactly is someone who is confused about whether <basefont> > can go in <head> going to hack their own DTD? That requires just basic understanding of SGML. Why would people use an SGML validator if they lack a basic understanding of SGML _and_ are unwilling to learn? Yeah, right. Because the W3C claims that you get an interoperable Web page if you put your document thought a validator and get a "pass". Someone who is confused with a validator's error messages should check the description of the syntax of the markup system (such as HTML) he is using. For this purposes, prose descriptions can be used, such as the nice material at http://www.htmlhelp.com/ (though unfortunately this volunteer work has not been updated to HTML 4.01 level, but the differences are small). > IMO it is my business > as a tool developer to help users, not to put superfluous learning > curves in their way. > > Noone forces you to select fussy mode. The beta at http://validator.w3.org:8001/ _still_ has the fussy mode as the default. It has now the explanation "In this mode it will generate warnings about some things that are not strictly forbidden in the HTML Recommendation - -" First, that "are not strictly forbidden" is very odd. They are not forbidden, period. Actually, they are definitely allowed. But worst of all, the explanation says it will generate _warnings_. What the beta actually does is that issues _error_ messages and claims "This page is not Valid HTML 4.01 - -" for pages that are actually valid. As far as I have understood, this cannot be fixed, due to the way the "fussy" mode has been built into the beta. Does this ring a bell? -- Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/
Received on Friday, 31 October 2003 02:03:15 UTC