DESIGN -- Was: What are your plans for MathML macros?

Dear Friends --

I find that my reaction to this discussion bifurcates on the issue
of whether we are thinking about (1) whole document MathML -- consistent
with what I perceive still to be the W3C theme of whole document XML,
as demonstrated in the W3C testbed item "Amaya" or (2) encapsulated
MathML, i.e., MathML embedded in larger documents.

Overall design is very important.  One could have both things possible,
but everything has a cost.

My understanding of the history of the W3C HTML-Math WG, born in the
early time after the math in the draft for HTML 3.0 was scrapped, is
that the encapsulation idea was going to be "temporary".  It is, of
course, very reminiscent of the use of bitmap images for math symbols
by early "latex2html".

The strength of encapsulation is that it enables you to go off on
your own.  (By 1993 with the early 1.0 level of HTTP, we could, in
fact, do that, and we would have been happier than we are now
if by 1995 a descendent of "xdvi" or even "ghostview" (with no tax
on authoring) had been distributed with "netscape", "mozilla", and
"MSIE".

The weakness of encapsulation is, as David Carlisle said, the
availability of processing for the encapsulated object.

If math is to be "special" in our HTML or XML, has it been agreed that
all processing of the encapsulated objects will be governed by DSSSL
instructions?

If math is not to be encapsulated for "MathML", then we need an
XML document type definition (DTD) for the whole thing.

And, either way, it might be good to have a few public "entrance"
DTD's (probably looser than XML) that could be targets to which
authoring systems could render as a preliminary authoring-platform
step toward our ultimate whole document format.

I would not perceive such public "entrance" DTD's as good DTD's
for authoring.  Rather I would want to flow from an authoring
DTD toward an entrance DTD.  (Composition of pseudo-morphisms in the
pseduo-category of markup languages.)

For that matter the large preprint archive at Los Alamos National
Laboratory [http://xxx.lanl.gov/] could benefit from a few
"archival" DTD's as a step toward "smartness".

                                   -- Bill Hammond

Received on Friday, 30 October 1998 09:10:44 UTC