Re: HTML/SGML/charsets

Terry Allen (
Fri, 31 Mar 95 12:23:44 EST

>This is a very simple issue, but very hard to choose. We will either be
slightly incompatible with SGML, or we will be slightly incompatible with
MIME. I believe we made the decision to choose the former. Am I recalling
this incorrectly?

The issue was given only a short time for discussion, which ended with
Larry promising to supply some language. In the draft that ensued
(01) this stuff about charsets was qualified with the statement that
the SGML standard was definitive. That has changed.

In order to have a parseable SGML document I must have an SGML declaration.
We must define how to produce the desired SGML decl from the info
given in the MIME charset parameter. That in no way requires us not
to conform to ISO 8879, the SGML standard.

Furthermore, the issue of non-8859-1 charsets was put off until
2.1, as shown by the fact that we are supplying an HTML SGML decl,
mentioned in section 2.1 of the 02 draft (though the reference
to DTDs in the plural is dated).

There is simply no need to go and break forever numeric charset
entities in HTML. The need is to define how we want to convey
charset info, and that is explicitly evaded (as we agreed in San
Jose, to my memory) in this 2.0 draft.

As Murray pointed out, "slightly incompatible" is like "slightly
pregnant" so far as SGML is concerned. I don't see why any
incompatibility with MIME is involved *for 2.0 and 8859-1*.


Terry Allen  (   O'Reilly & Associates, Inc.
Editor, Digital Media Group    101 Morris St.
			       Sebastopol, Calif., 95472
occasional column at:

A Davenport Group sponsor. For information on the Davenport Group see or