Re: Conformance ratings (was: Extra! Microsoft beats Netscape in the race for...

Jonsm@aol.com
Tue, 13 Feb 1996 03:29:50 -0500


From: Jonsm@aol.com
Date: Tue, 13 Feb 1996 03:29:50 -0500
Message-Id: <960213032949_320615123@emout09.mail.aol.com>
To: murray@spyglass.com, www-html@www10.w3.org
Cc: eric@gauthier.centre.edu
Subject: Re: Conformance ratings (was: Extra! Microsoft beats Netscape in the race for...

In a message dated 96-02-12 22:49:55 EST, murray@spyglass.com (Murray
Altheim) writes:

>Jon Smirl <Jonsm@aol.com> writes:
>>1) Errors - pages containing severe errors like overlapping tags,
>
The basic idea here is to ensure lexical correctness. A SGML parser can do
this by suppressing all tag/attribute related error messages. It might be
simpler to build this check from a lex/yacc grammar. Special work will need
to be done to try to detect unclosed comments. NS allows some crazy things in
comments, like <! comment > and <!-- comment > (this one searches to EOF for
-- and if not found declares the > as terminating the comment). These
problems need to be detected and flagged.

DOCTYPE is not really an issue. I would just try to parse the page using all
of the tests/DTDs and report the error free ones and an error count for
failures.

>>4) Extended HTML2  -  tables and extensions common to several browsers.
A...

This rule is amended to require four non-related browsers to support the
feature, four Spyglass or Netscape variations don't count.

>>5) Vendor specific - pages that will only work on specific browsers
(frames,
>
>If you have some magic formulae for *requiring* vendors to post a DTD for
>their products, you are pretty welcome to post that here....

If a vendor doesn't provide a DTD their pages will be flaged as unknown
vendor extended. This is the catch all category for pages that are lexically
correct but fail on all of the DTD's. The penalty is that no one will know
which browser the pages have been extended for. Skull and cross bones might
be a good icon for this.

Jon Smirl, jonsm@aol.com