This is my perspective, and should be read as "as I understand it" and
"in my opinion". Other people might well disagree with parts or the whole.
Undoubtedly, like anything posted to Usenet. I have refrained from
answering the question, as I don't know the specifics of Dreamweaver,
and I don't know whether the OP is developing brand new pages or
maintaining/developing old pages. It makes the huge difference.
In any case, the question "which DTD should I be using?" is as such
rather vague. The doctype declaration has two functions: browser mode
selection, and validation. If you don't validate your pages, then the
short answer is: don't use any doctype declaration, if you want "quirks
mode"; and use <!doctype html>, if you want "standards mode". And if you
use validation, you should understand the doctype issue on your own;
otherwise the validator reports will just confuse you.
The question "which version of HTML should I be using" is something
completely different. But it is important to realize that browser don't
care about HTML versions. Well, they might care about HTML vs. XHTML
_if_ you serve XHTML as genuine XHTML in the HTTP headers, but nobody
should do that without understanding the consequences, and few people do
it anyway.
It really doesn't matter much which "version of HTML" you use. What
matters is what tags and attributes you use, and how. This is tagsoup
world. Even if your doctype declares your document as HTML 2.0, browsers
will happily implement your use of tags introduced in HTML5 drafts if
they support them at all. And even if you declare HTML5, which declares
many old tags as obsolete, cursed, anathema, and criminal (pardon my
slight exaggeration), browsers will still obey your pre- HTML 2.0 markup
as usual, to the extent that they know about it at all.
Historically, html was based on sgml.
No, it was just nominally retrofitted to SGML.
Html 4 was pretty much defined by
the turn of the century.
In 1997; later slightly edited in 1998 and 1999.
(depending who you ask, xml was or was not a development of html in the
first place)
XML is a simplification of SGML; seeing it as a development of HTML is a
gross category error.
XHTML 1.0 was basically a re-write of html 4.01 to match the rules of xml.
Yes.
It didn't add anything new to html 4.01, other than making it legal xml.
It did. The differences were never properly documented. Nobody was
interested enough.
The theory was that by making a few minimal changes, existing html 4.01
could be ported to the new shiny xml based xhtml.
Hardly. The idea of XHTML is that XHTML documents can be processed by
generic XML tools and combined with other XML documents. Using XHTML
just as a publishing format for web pages is absurd: it gains nothing.
is best mercifully forgotten.
Eventually, xhtml suffered stagnation and became a dead end. At the same
time, html 5 was becoming the next thing.
HTML5 has HTML serialization and XHTML serialization. So instead of
killing XHTML, it makes it an option.
html 4.01 - mature, well documented, predictable behaviour across all
browsers as long as markup is compliant
Well, nothing like that, except as compared with the alternatives.
I still prefer to work in html 4.01, basically because compliant 4.01
markup still has the best chance of appearing as intended across the
widest selection of browsers currently in use.
There's no magic in sticking to HTML 4.01. Most constructs there are
widely supported, though often with considerable rendering and
functional differences. But just being part of HTML 4.01 guarantees
nothing; browser developers have no interest in HTML 4.01 any more. And
some HTML5 features work fine in browsers, because they are not really
new, just standardization of common practice.
