FYI and, I hope, some action: Last Call for MathML 3.0 spec

[Chris Rowley]

This is now available and all comments need to be made by sometime in
November. I hope you have some!
It contains lots of stuff about Unioode maths and how to use it; and
lots of other stuff.

So go !!NOW!! to

http://www.w3.org/Math/

for further information.

Thanks, chris

[Will Robertson]

On 27/09/2009, at 4:18 AM, Chris Rowley wrote:

> This is now available and all comments need to be made by sometime in
> November. I hope you have some!
> It contains lots of stuff about Unioode maths and how to use it; and
> lots of other stuff.

I've skimmed over a fairly large chunk of it, skipping the section on
content maths because that's something I'm not yet ready to think about.

Basically, in my opinion it all looks good. MathML is a simple idea
underneath but the effort it must have been covering the breadth
necessary to describe basically every possible way to notate (?)
mathematics is commendable.

I don't know if that sentence made sense.

Anyway, what we have in unicode-math so far is approaching the problem
from the LaTeX direction; obviously MathML could never be handwritten.
Whether, in the future, we stick to a [unicode+syntax -> TeX-variant]
typesetting approach or move to something that involves converting to
MathML before rendering the maths.

But although we're both dealing with a syntax to eventually result in
a visual representation of some maths, I don't see a whole lot of
overlap. (Any "user-level features" of unicode-math to aid convenience
would surely be implemented at the application level *above* MathML.)

Particularly since symbols don't have a "built-in" meaning and require
explicit markup, there's not much to take away for suggestions on how
the unicode-math "linear input" should be modelled. (Although the
recommendation discouraging the shorthand that <mo>x'''</mo> stand for
"x" with the unicode triple-prime symbol goes against a recent feature
of unicode-math. But, again, that's on a different level of processing
to what unicode-math is for, so I don't feel bad about doing here what
they're not.)

Interesting read, though; thanks for passing it around.

Will