Presentation and Semantics debate

Nathan Torkington (Nathan.Torkington@vuw.ac.nz)
Tue, 3 Aug 1993 12:53:53 +1200


(I accidentally sent this only to Bill J --- my apologies to him for
his getting it twice)

Bill Janssen writes:

> I completely agree. The point is that there are many many formats which
> do provide this capability. Simply use one of them (MS-Word,
> Postscript, FrameMaker MIF, TeX, troff, etc.), instead of trying to use
> a format which does not provide this kind of thing.

The trouble is that there is a need for hypertext markup that has
procedural elements. I want to write hypertext, *and* specify some of
the output (starting a list at (e) for instance).

You would have me write in Postscript or TeX or some other format that
doesn't have hypertext ability, or write in HTML which doesn't have
presentation capabilities. You are effectively preventing me from
doing something, and this isn't on.

SGML's non-procedural emphasis comes from the need to use texts for
multiple purposes --- it's easier to know how to create an index if
the headings are marked as ``headings'' rather than as ``some text at
14 pt Palatino''. HTML will be used for a variety of purposes, but
none of those purposes (as far as I know) precludes a limited set of
procedural elements. Please enlighten me if I am labouring under a
misapprehension.

HTML[+] is not, and doesn't seem likely ever to be, a language for
presentation-independent documents. It's designed for interactive
browsing by humans and possible indexing by a computer, and as such
should be tailored for these purposes. Anything else is subverting
the goals of HTML[+].

Nat