- From: Tim Bray <tbray@textuality.com>
- Date: Fri, 27 Sep 1996 14:44:31 -0700
- To: w3c-sgml-wg@w3.org
At 04:03 PM 9/27/96 -0500, R A Milowski wrote: >It seems to me that we are taking a big step backwards by not using and >requiring DTDs. I would rather see us push forward with the SGML standards >process... >It would seem that it is a requirement that DTDs be used in the >authoring process. If DTDs are used in the authoring process, they >they should be available for interchange. If they are available for >interchange, they should be used in some way be the receiving application. I share Alex's position that DTD's should be used for authoring. Otherwise, as Charles has pointed out, you just get HTML with extra tags. However, there are many other applications of structured text where the DTD is not particularly helpful. These include display, where you need a stylesheet but not a DTD, and full-text-search, where you need some metadata but no DTD. There is also the problem that it is often desirable to deliver small fragments of SGML, and given the bloated size of many DTD's, it is likely to be the case that the DTD would be *much* bigger than the fragment you're delivering. As has been said more than once: things that require knowledge of the grammar should use the DTD. Things that *don't* require such knowledge shouldn't have to. With SGML, because of minimization and SHORTREF & so on, you can't do *anything* without a DTD. This is a primary motivation for XML's design. Cheers, Tim Bray tbray@textuality.com http://www.textuality.com/ +1-604-488-1167
Received on Friday, 27 September 1996 17:41:12 UTC