Next message: Christian Smith: "Re: Bug in W3C HTML validator"
Date: Mon, 19 Mar 2001 07:32:58 -0500 (EST)
From: Brian Gilkison <gilkison@one.net>
To: Frank van Wensveen <frankvw@euronet.nl>
cc: www-validator@w3.org
Message-ID: <Pine.LNX.4.21.0103190723310.4871-100000@shell.one.net>
Subject: Re: Bug in W3C HTML validator
On Sun, 18 Mar 2001, Frank van Wensveen wrote:
> I just lost an argument with someone who claimed that Doctor HTML was
> a better validator than the W3C validator. Rather painful - he turned
> out to be right. :-(
You had best start arguing again; it turns out he was wrong... ;)
> Doctor HTML (http://www2.imagiware.com/RxHTML/) found that one of my
> pages was missing both a </TD> and a </TR> tag. I claimed that it was
> wrong, since the W3C validator had OK'ed it. Unfortunately it turned
> out that my page DID miss the aforementioned tags.
It sounds like "Doctor HTML" is not so great a validator; TR and TD tags
are optional, per the HTML
spec: http://www.w3.org/TR/html4/struct/tables.html#edef-TD
A TD opening tag is implicitly closed when the next TD or TR opening tag
appears (or a closing TABLE tag), while TR is implicitly closed when
another TR opening tag or TABLE closing tag appears. This is not to say
that matching closing tags are not A GOOD THING (tm); stylesheets work
much better when any opening tag with an optional end tag HAS that end
tag...
> As I said the first page is OK'ed by the W3C validator and marked
> incorrect by Dr. HTML (and it does indeed miss tags) while the second
> is OK by both validators. "Manual" syntax checks on both versions
> confirm this.
Congratulations, they're both valid HTML 4!
> Now I won't hold this loss of face I suffered agains you... :-)) But
> I'd appreciate if somebody could look into this, since now I feel that
> my rocksolid trust in the W3C validator was not completely warranted
> after all. It *did* miss a few syntax errors, and nasty ones at that
> (missing table tags can wreak havoc in Netscape Navigator especially).
Your trust in the validator is still warranted; however, your trust in the
accuracy of Navigator's interpretation of valid HTML is another :)
Brian