Re: Cloture on presentation vs. structure debate
Fri, 28 Oct 1994 06:44:30 +0100


In article <>, (Brandon Plewe ) writes:

|>Perhaps we should vote--the debate is going in circles, and the longer we talk
|>about it, the more bad solutions will arise--something needs to be _done_. I
|>see several possible futures (not mutually exclusive) that have been brought up
|>in this discussion, and I'd like to know what people see as being the best one.
|>Does someone want to throw up a quick voting form?
|>1. HTML+++
|> HTML is continually expanded to include more direct-presentation-control
|> tags, some accepted by the HTML WG, some done ad hoc (ala MCOM).

Not acceptable. HTML is an abstract presentation language. Its development is
formally within the IETF standards process. This is an open process which began
quite a long time ago and has debated these issues at length. Preempting the
process cannot be allowed to set a precedent.

|>2. HTML 3.1
|> HTML as it has been, plus stylesheets to control presentation.

This is the route that was agreed upon by the participants. I don't thnik that
anyone voiced significant objections.

|>3. SGML
|> WWW Browsers become SGML browsers, allowing providers to use the
|> DTD best suited to their documents: PDL (presentation display lang)?,
|> HTML, MDL (math display language)?, any other DTD, which can all be
|> mixed in a single document.

HTML 3.0 has maths.

The SGML issue is a red herring. SGML only defines a structure, not a
presentation. For that try <? Process Directives>.

|>4. Multi-format browser
|> Mosaic, etc. become larger, able to read and display HTML, Web-PDF,
|> Web-TeX, CGM, SVF, VRML, etc.

It would start to get unmanageable eventually. But this is the way that
most see the Web developing.

|>5. Chimera-style multi-format browser
|> Single browser which uses separate rendering plug-in programs to
|> interpret different formats

Plug in dynamic link routines is a more appropriate solution.

|>6. Multi-browser (CCI, CORBA)
|> Several smaller browsers, one for each format, which can communicate
|> with each other.

This is an implementation issue that I don't see as a proper subject
for a vote. If someone wants to do this then they can go ahead with or
without a vote. It is simply one of several valid technical solutions.

I would be very interested in seeing such a browser.