Re: Review of WWW Mathematics Browsers, MathML and HTML.

> An authoring DTD is worth the effort of careful construction.

yes, of course.

> Regarding MathML, it might be helpful if one knew what kind of
> authoring DTDs people will actually be using.

Well who knows:-). They may of course not be authoring to a DTD at all.
They may be authoring in something else (like tex) and converting,
or using a dedicated tool like ezmath or a mathml aware mathtype or

If they are using an XML editor to author directly to some DTD
then one would hope that they will have the freedom to have many
different possible conventions, together with a transformation
mechanism to transform to MathML. So in that case it doesn't matter
too much if the authoring DTD differs from mathMl does it? (eg by having
elements if you prefer that).

> Has anybody in the MathML group ever tested non-default processing of
> "α"?  If so, with what?

Not as far as I know. In an XML/SGML based system (like for instance my
dsssl scripts for jade, an entity declaration in local part of the
doctype declaration has priority over one in the DTD, so within the
file with the MathML you could declare the entity alpha to be anything
you liked if you did not like the default definition (which is a
character entity pointing at some slot as determined by unicode).
`Dedicated' mathml renderers probably do not have any entity definition
mechanism built in, but there is nothing to stop you transforming your
document with dsssl or xsl etc from some extended or modified DTD to
standard mathml before passing it to a mathml renderer.

> Has it been agreed that the MathML group will produce a public dtd for
> voyager+mathml?

I'm not sure what has been formally agreed but the math and HTML groups
are committed to making sure that MathML can be easily integrated into
extended HTML systems. (including if necessary making changes to MathMl
to conform to whatever interfaces are finally decided, once they are
decided). The issue of including elements from some `external' XML DTD
into an HTML page is of course a general issue not specific to MathML.