Re: A proposal for a computer readable mathematical object notation

I think this discussion is very valuable since the preferred way today
seems to be to develop an abstract data model and then talk about
serializations; cf. the work in nthe Annotation and POE WGs. From a
semantic point of view, I also consider the problem of expressing
layout+semantics simultaneously as critical, especially in a fashion that
is independent of the rendition.

Neil wrote:
> No one wants to type in an editor and say "this 'x' is a real valued
number" just so that some evaluation down the line knows how it interacts
with some other variable. Indeed, people don't event want to say whether
h(x+y) is function application or multiplication. Computational systems
solve that problem because their syntax defines what is meant. Neither TeX
nor most WYSIWYG editors provide that information in general.

I see two problems with this statement.

On the one hand, where's the data? My experience is that authors very much
do want to express this "x is a real valued number".

On the other hand, why don't they? In my experience, the dominant problem
here is print. Presentation MathML was designed to capture print layout
traditions of equational content; it does this very well but is not suited
to do much beyond that. From what I've gathered from its history, MathML
was also heavily influenced by the needs of the publishing industry (since
its success rests there) and authors working with publishing entities
generally think only of print (since publishers alas still do).
Accordingly, authoring tools generally lack the means to express semantic
information (efficiently).

Most importantly, there is little incentive for expressing more than purely
visual information while authoring for print (mainly, convenience such as
macros or templates). The web changes this on a fundamental level; writing
a component that expresses a thought both visually and rich in semantics is
perfectly feasible and the means of expression for both is vastly expanded.
I'm very interested in anything that helps move towards this.

Peter.