Re: Flagging & in URL in HTML 4.01 transitional type.

On 08.06.01 at 23:57, Mike Heins <mheins@redhat.com> wrote:

>[...] Yet why do we have an HTML 4.01 transitional spec, and an HTML 4.01
>strict spec?

For the same reason that a C compiler has a "-ansi" or "-posix" switch. The
"-pedantic" or "-Wall" switch is a bit different. The "Strict" in "HTML 4.0
Strict" talks about a different kind of strictness then what you are
addressing here.


>And why does a program like lint(1) have differing levels?

Because lint(1) is just a fluff-checker. The W3C HTML Validator is a formal
validation tool; it validates the supplied input, using a SGML Parser,
against the DTD for the Formal Public Identifier that you've chosen and the
SGML Specification. Lints just look at your code and makes suggestions
based on arbitrary, subjective, notions of "good practice" or "style".


>Maybe I am arguing that the HTML 4.01 transitional spec is wrong
>and should be changed.

No, you are arguing that the HTML 4.0 Reccommendation as a whole is wrong
and should be changed. This is an entirely valid argument and I'm half-way
inclined to agree with you. However...

>I think I just ended up on the wrong list.

...the Validator can do nought but enforce the specification as it exists.
The proper way to deal with this is to get in touch with the HTML Working
Group at the W3C and raise the issue there. The www-html@w3.org mailinglist
may be a good place to discuss this. Here on www-validator@w3.org there is
very little we can do about it. So, yes, you ended up on the wrong list.


>All I wanted to do was find out solid reasons why the validation flagged
>that, and I haven't found that out.

Did this perhaps answer your question a bit better? Not satisfactorly, of
course, but at least give you the explanation for why the Validator flags
these things?


>No real reason for this has been shown other than the case of &copy=, and
>this is defended because the semicolon is optional in an entity, as
>defined by the spec. Why the heck would the semicolon be optional? What
>good reason could there be for that? No one seems to know or care. It is
>the spec, after all, and it must be validated. If that is the totality of
>the mission statement, congratulations to the authors.

HTML is a big body of work, and SGML even more so. It may very well be that
this is a simple oversight and based on existing practice. Again, you'd
need to talk to the HTML WG; they make the rules, we just enforce them. :-)


>If you look at the dictionary definition of pedant, it has a word in
>the definition -- "needlessly".  And pedants are eventually ignored by
>most people, as I feel HTML 4 compliance is being ignored. I think
>have found out why.
>
>This list appears to be silently moderated and without charter, not
>allowing me to post or subscribe after my first post. I find that rude
>in the extreme, particularly so in a quasi-public forum.
>
>Thanks for your response, and please bid the folks there a kind adieu,
>as I cannot.

I think perhaps you'll find that you will get much further if you adopted a
slightly less confrontational attitude. Your first message to this thread
-- and, incidentally, your first message to this list at all -- gave the
impression of a preconceived opinion; you'd already decided that the
Validator was in error and managed to imply that it must be a completely
useless piece of junk for making such a stupid error. The sorts of replies
you got were somewhat colored by that. Though, admittedly, some may also
have been colored by the sort of questions we often get on this list
regarding unencoded entities and such.

To quote spaf:

      Lighten up! It's not real life, it's just ones and zeros. :-)

Received on Thursday, 21 June 2001 21:46:38 UTC