[tei-council] SGML or not to SGML, that is the question

Syd Bauman Syd_Bauman at Brown.edu
Thu Sep 14 19:38:47 EDT 2006


> > > what would be involved? 

Supporting SGML means a lot, actually. Including, but not limited to,
the following from the top of my head. (Many of you may have answers
or responses to some of the following issues -- please refrain from
interjecting them until we've decided to support SGML, if we do.)

* Providing an SGML declaration.

* Discussing the ramifications of using a different SGML declaration.

* Creating rules for conformance and recommended practice that
  include SGML as well as XML.

* Deciding whether the Gentle Introduction needs to include SGML as
  well as XML.

* Deciding what to do about non-Unicode characters in an SGML
  environment (where SDATA entities mean gaiji is not needed).

* Deciding whether or not to discuss use of SGML with non-Unicode
  character sets.

* Deciding whether the chapter on multiple hierarchies should address
  CONCUR or not.

* Deciding what to do with <egXML>, which is in a different namespace
  (namespaces don't really exist in the SGML world).

* Deciding what to do about pointers. Our pointers rely on the target
  being XML.

* Deciding what to do about attributes that have "xml" in their name
  -- xml:id is a bit of a misnomer, isn't it?

Now that I think about it, I'd say that supporting SGML would push
back release of P5 1.0 by months. Thus I'm no longer just mildly
against it because I'm lazy -- now I'm seriously against it because
P5 needs to get finished, and I see very little use for SGML.





More information about the tei-council mailing list