- From: Arjun Ray <aray@q2.net>
- Date: Tue, 5 Oct 1999 00:02:59 -0400 (EDT)
- To: W3C HTML <www-html@w3.org>
On Mon, 4 Oct 1999, Russell Steven Shawn O'Connor wrote: > On Mon, 4 Oct 1999, John Whelan wrote: > > > XML, unlike SGML, can be parsed without a DTD, As Russell points out, SGML isn't subject to that restriction any more. But there's a further point, in that his statement is exactly correct: > Now SGML documents don't require declarations. That is, the issue isn't some entity called a 'DTD', but the need for declarative information. > > which seems like the major selling point of XHTML; Yes, in terms of buying mindshare among people already predisposed by popular myths to consider SGML as incurably ugly. There's nothing mysterious in the utility of declarative information, or that it can be useful to have a *distinct* syntax for such stuff (the only alternative is to use "reserved words" by which a whole mess of tags can still be partitioned into "meta-information" and "instance data".) Mainly, it's a concession to the "reality" that the web is what Netploder supports, and so since Netploder has only a crippled notion of tag syntax, anything that gravitates to that crippled form hasta to be kinda sorta good. Virtue out of necessity. > > XML browsers can be forward-compatible without fetching a separate > > file to tell them whether they've just passed a start tag or a > > stand-alone. This is a red herring (it's a myth that the DTD has to be an external subset, and also a myth that what might be pointed to by a public or system identifier "is" the DTD.) The DTD is *incorporated* by the doctype declaration. That is, the doctype declaration is where one finds the DTD. How the stuff got there is an irrelevant detail. In particular, it is not *necessary* that an external subset exist to be incorporated by reference. Everything is allowed to be inline. Arjun
Received on Monday, 4 October 1999 23:21:28 UTC