Re: Math and MathML [input syntax]

Patrick Ion said:
> In another place, you castigate MathML for
> 2) lack of an adequate input syntax (see original CanonMath program)
>
> A universal input syntax  is something that early math markup
> developments
> thought would be a very good idea, and many imagined would be rather
> easy to specify.  They were quite wrong, of course.
>
> The W3C Math WG  did consider input syntax questions, but decided it did
> not have enough resources to specify such a thing when there was such
> clear disagreement amongst WG participants as to what the syntax
> should be.

If the MathML WG generates a markup is so verbose cannot be typped by
hand, and introduces on the markup capabilities are _not_ present in TeX
and similar system, I think that conclusion that any _a priori_ advantage
of MathML over TeX is lost when later is used as input syntax is really
easy to obtain without need for masterminds ;-).

It is very funy reading one of MathML prograndistic article in AMS
bulletin, where are explained the advantages of MathML over TeX but after
the same MathML WG endorsing like “nice” the IteX approach is unaware of
report in another AMS bulleting that IteX input

{}_a^b X

will generate the completely incorrect, inaccesible, not searchable, and
structurally invalid

<msubsup><mrow></mrow> <mi>a</mi> <mi>b</mi></msubsup><mi>X</mi>

whereas the specific MathML markup for prescripts will be not used. As
trivial consequence most of code is being spread over the internet is just
an ultraverbose version of the previous markups models that the WG
critized and rejected as inadequate.

> I think the new Math WG should revisit the question of input syntaxes.
> In my own present opinion, there should probably be a recognition that
> locally preferred input syntaxes are useful to their special
> communities.

This was addressed here sometime ago. Fortunately, I see in the 2006
Charter that WG plans reconsideration of its lon-time error and will begin
to investigate an input syntax. The original CanonMath program addresing
an input syntax for MathML was eventually abandoned but maybe some idea
could be useful and reused.

>
> Patrick

Juan R.

Center for CANONICAL |SCIENCE)

Received on Monday, 17 July 2006 11:51:33 UTC