Re: Is this more or less correct, about the XHTML faq?

In article <G9-dnU9dMsk4B8PUnZ2dnUVZ_v_inZ2d (AT) bresnan (DOT) com>,
salmobytes <salmobytes (AT) closenuf (DOT) org> wrote:

I'm asking here, not asserting. So I'm not looking for a fight.
Is the following more or less correct?

If you like the logical consistency of XML and XHTML (lower case
attributes, all attributes must be quoted, all tags closed, etc) you can
write XHTML-like documents THAT DO NOT throw IE6 into quirksmode,
if and only if you serve the document as ContentType: text/html
....

Others will say "why bother, this accomplished nothing. Just use 4.01
strict!"

You can use lower case and closing tags and quoting in 4.01 Strict. You
can use an editor that warns of 'missing' closing tags, you can even
make up your own doctype to make it an offence. As for small case,
surely this is so simple a habit to get into? I also want to say "why
bother" with xhtml for now? But if it suits you to use xhtml, it is not
a definite mistake.

--
dorayme

dorayme wrote:

In article <G9-dnU9dMsk4B8PUnZ2dnUVZ_v_inZ2d (AT) bresnan (DOT) com>,
salmobytes <salmobytes (AT) closenuf (DOT) org> wrote:

I'm asking here, not asserting. So I'm not looking for a fight.
Is the following more or less correct?

If you like the logical consistency of XML and XHTML (lower case
attributes, all attributes must be quoted, all tags closed, etc) you can
write XHTML-like documents THAT DO NOT throw IE6 into quirksmode,
if and only if you serve the document as ContentType: text/html
...
Others will say "why bother, this accomplished nothing. Just use 4.01
strict!"


You can use lower case and closing tags and quoting in 4.01 Strict. You
can use an editor that warns of 'missing' closing tags, you can even
make up your own doctype to make it an offence.

Of the practices the OP mentioned, the one that you can't do in 4.01
Strict is closing all elements, because of the elements that by the
specification don't have an end tag. But still, if his intent is to be
able to validate his own documents to his own specification, he can
create his own DTD in which those tags *are* closed, while delivering
the pages to clients with the 4.01 Strict doctype and letting them
ignore the trailing slash in the <input>, etc., tags as the
specification requires them to do.

Harlan Messinger wrote:

Of the practices the OP mentioned, the one that you can't do in 4.01
Strict is closing all elements,

You cannot explicitly close all elements with separate end tags, but by all
HTML specifications, all elements are always closed, either explicitly or by
well-defined implicit closing rules.

But still, if his intent is to be
able to validate his own documents to his own specification, he can
create his own DTD in which those tags *are* closed, while delivering
the pages to clients with the 4.01 Strict doctype

There's no point in doing that. It's neither practical nor theoretically
clean. Besides, how would you write an SGML DTD that allows <input />?

and letting them
ignore the trailing slash in the <input>, etc., tags as the
specification requires them to do.

No, HTML specifications up to and including HTML 4.01 do not require any
such thing. Au contraire, they specify HTML as an SGML application, which
implies quite different things as regards to processing <input />. People
use to say that old web browsers ignore the slash, as they generally do, but
that's a bug (a violation of specifications).

ObCSS: None of this topic has anything to do with stylesheets. The original
message had some CSS implications, since it discussed doctype sniffing. No
matter how you write an <input> or <img> tag, it remains an odd bird from
the CSS perspective, treated as "replaced element", which is largely just a
manner of saying they're odd and follow their own rules.

--
Yucca, http://www.cs.tut.fi/~jkorpela/

"Jukka K. Korpela" <jkorpela (AT) cs (DOT) tut.fi> writes:

Besides, how would you write an SGML DTD that
allows <input />?

Not in the document type declaration subset of course, but generally in
the same (Annex K) way that the SGML declaration of XML does it
(i.e. nothing else but ‘<input //’ in disguise).

That is not to say that this is useful in the given context; on the
other hand, SGML validation is never useful for HTML in the first place,
besides you there are maybe 3 people alive who understand it (and
interested in both HTML and SGML, that is); if your authoring tools
understand all that and you really know what you are doing, I can see
the benefit instead of the problem, though.

--

|| hexadecimal EBB
o-o decimal 3771

--oOo--( )--oOo-- octal 7273
205 goodbye binary 111010111011

Eric B. Bednarz <bednarz (AT) fahr-zur-hoelle (DOT) org> writes:

(pardon me, fullquote below & followup-to set)

"Jukka K. Korpela" <jkorpela (AT) cs (DOT) tut.fi> writes:

Besides, how would you write an SGML DTD that
allows <input />?

Not in the document type declaration subset of course, but generally in
the same (Annex K) way that the SGML declaration of XML does it
(i.e. nothing else but ‘<input //’ in disguise).

That is not to say that this is useful in the given context; on the
other hand, SGML validation is never useful for HTML in the first place,
besides you there are maybe 3 people alive who understand it (and
interested in both HTML and SGML, that is); if your authoring tools
understand all that and you really know what you are doing, I can see
the benefit instead of the problem, though.