On 01/13/2010 07:27 PM, José Matos wrote:
On Wednesday 13 January 2010 18:51:19 Andre Poenitz wrote:
- what to do with math?
Try to use the 'layout oriented' bits of MathML and invent our own for
the few cases where there is no clean match.
Richard you have been fiddling with MathML, what is your opinion about the
roundabout trip?
How difficult would it be?
I don't think this is feasible, really. MathML is *very* limited, and I
have often found myself using the same MathML construct for very
different LaTeX constructs. The content-oriented form might actually be
better, but I haven't done anything much with it, because it's not
relevant to display. So we'd end up "rolling our own" for a very wide
range of cases, in which case we might as well just "roll our own"
completely.
It seems to me that the easiest thing to do would be to do with math
pretty much what we do with other insets: Write some structured
representation that tracks the internal representation in LyX. Then
parsing becomes easy, as compared with what we have to do now. I see the
advantages of the current method: we already need to be able (a) to
write math as LaTeX and (b) to read LaTeX into math, so why not reuse?
Maybe those advantage are even enough to keep things as they are, more
or less.
Anyway, for now, it seems to me, we should leave math alone and just do
the boring thing I suggested earlier:
<math display="false">a+b=2^x</math>
That is, unless you think that there is a pressing need to change the
way math is represented, a need that is part of what is driving the
change to XML.
And perhaps another thing to say here is that any XML-like
representation of math will definitely make sed-like script processing
of LyX files harder.
Richard
