Re: [Bug-gnupedia] math content

[Thomas E. Vaughan]

On Wed, Jan 17, 2001 at 06:11:04PM -0000, Duncan Lock wrote:
>
> I am not a mathematician but I looked into this a while ago for another
> project. I would say MathML is the way to go here. Unfortunatly it isn't
> very well supported in browsers at the moment. IE4/5 half arsedly
> supports it and Mozilla aslo has some support (although I can't remember
> how much) As for creating MathML content you've got even less choice
> really. On the free front you could look at amaya from the wc3 no less
> (www.w3.org) which is CLUNKY but will produce MathML (amongst other
> things) Maybe MathML on the backend but translated to something else (an
> image?) for the client end?

It seems to me that there are three client ends, or one client end with
three heads:

(1) what the user's browser gets when the user clicks on a link to an
    article in the encyclopedia,

(2) what the user submits in order to search the encyclopedia, and

(3) what the author submits in order to add content to the encyclopedia.

Even if MathML doesn't (initially) play a role in any of these, MathML may
still be useful as an internal format if it provides us a standard way to
store mathematical content so that it is searchable.

Surely, though, almost all substantial mathematical content will be typed
up by the author in (La)TeX, and so what we'll need is some software that
will convert at least the formulas in a TeX document into MathML.  For
example, if latex2html could produce not only a PNG image file for each
formula but also MathML for each formula, then we would have the
flexibility to serve in-line either a nicely antialiased image or MathML
code to the user's browser.  Even if it is not served out directly, the
MathML could, unlike graphic images, be queried for certain mathematical
expressions in an as-yet undetermined way.

-- 
Thomas E. Vaughan <address@hidden>
CIMMS/NSSL, Norman, OK, USA