Not to criticize someone's hard work (OK, I guess I am...) but aren't we
barking up the wrong tree here? Shouldn't it be possible to discover
everything we need to know about HTML by parsing a DTD? That way as the
language changes, software would automatically recognize new tags. :)
(Of course, *which* DTD to parse could be problematic... we might need a
"sloppy" DTD covering all the unofficial and obsolete tags that exist in the
real world.)
The stuff with in_head in_body etc. could be generalized -- as mchase notes,
td is "in_tr" not just "in_body".
Also re <TR> <TD> etc. there are a few (million) ill formed documents out
there, so you need to know that a </TR> also ends a pending <TD> even if some
bozo forgot to say </TD>
From: mchase%ix.netcom.com@Internet on 99-12-28 08:21 AM
To: libwww%perl.org@Internet, sburke%netadventure.net@Internet
cc: (bcc: Marvin Simkin)
Subject: Re: HTML::Known, containing HTML parameters
Would it be worthwhile to mention required parents? For example, <tr>
doesn't mean much without an enclosing <td>.
Will this module also include information about boolean attributes?
I think I like HTML::Tags best so far.
--
Mac :})
----- Original Message -----
From: Sean M. Burke <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: 27 December, 1999 15:09
Subject: HTML::Known, containing HTML parameters
> A few days back we kicked around the idea of having one module
> containing all the facts about all the HTML tags. Here's my first
> stab at such a thing, made by just combining things from Element,
> TreeBuilder, and LinkExtor. Thoughts, anyone?
>
> (Suggestions for the name are especially welcome. So far I've
> considered HTML::Tags, but that's bad because it's too close to
> HTML::Element, and HTML::DocType, which is not so good because it
> implicates this is a DTD, or the kind of information you'd find in a
> DTD.)
