Re: MathML vs HTML math, vs ???

On Mon, 26 Jan 1998, Richard J. Fateman wrote:

> I think that you have been misled into thinking there is only one
> version of the language in question. In fact, in response to various

Yes, this was indeed what I believed.  I guess the reason I got
that impression is because the page I was reading never said that
there was more than one MathML.

> To see that the short way may not always be enough, consider
> ambiguous notation like " a (x+y)"  which could depend on
> whether "a" is a function or a constant to determine if this
> is a multiplication or a function application.

This ambiguity is certainly a problem, and it is one that
those of us who publish equations (in whatever form) have to
worry about as we are putting together our documents.

> One might have to ask what  f_x meant, since the this might be a subscript
> with any number of meanings.  And for the notation to be useful we
> now believe we must distinguish these.

I disagree with this statement, but I think that is because my
desires and (what I perceive to be) the goals of the project are
orthogonal.

> 1. Well, it is a goal of the project, because otherwise we would just
> be discussing typesetting. There are fine typesetting languages
> already in existence.  TeX is quite popular in academic circles, and
> maybe you should consider that, if all you need is typesetting.

I use TeX (well, actually I use LaTeX, I'm not quite
sophisticated/masochistic enough to use raw TeX) for my typesetting 
needs.  Latex is fantastic for paper publication.  Unfortunately it is not 
suitable for publication on the web, because there are no browsers
available which can deal with it (at least none that I know of).

> 3. We are concerned with the fact that there is an abyss between conventional
> scientific publication and scientific computation which always requires a
> human reader to read the publication, interpret the statements, and
> "program the formulas" into a programming language such as Maple,
> Mathematica, Macsyma, Reduce, Axiom, Mathcad, Matlab, Fortran, C, Java, ...
> Of course not all scientific paper formulas, even in MathML would be
> instantly and completely understood by all (or any) of such systems,
> but let's at least give them half a chance!

I think this is a fantastic goal.  Really.  But I'm just a stupid
user, and the first thing I want to be able to do is present my
mathematics so that other *humans* can understand them.  

I like the idea of having some different specification for presentation
and content.  So if I'm interested in presenting something to other
humans, I can just deal with the presentation tags.  Unfortunately,
those are so dense and complicated that I would choose to not communicate
rather than to try and wade through that mess.

I'll say more about this in a later message.

> 	Since Waterloo and Wolfram are both involved in the Activity, you
> 	certainly know this.  If I want interactive math, I'm sure that I will
> 	be able to buy some kind of plug in from either of these companies.
> 
> Sorry, a plug-in that understands all of scientific notation, including
> ambiguous abuse-of-notation usages, that will solve this problem is
> highly unlikely.  Sure you can find half-baked solutions for (say)
> high-school level math. 

Yes, that's true.  But a plug in which understands Mathematica or
Maple notation would be "trivial", both of these packages already have
a perfectly useable notation for denoting the meaning of mathematics.  Not
only that, but it is a notation that a lot of people know already.

> 	Currently there is nothing that lets me display math, and it looks like
> 	what whatever is coming will not address these needs either.
> Try the TechExplorer plugin for a current solution.

I thought this was a great idea.  Then I went to the TechExplorer
page and discovered that it is only available for Windows systems.
So this is no help at all.

> 	So my question to you, the Working Group, is this:
> 	Have you actually talked to any scientists about what they want
> 	in a markup language for math?  
> 
> Personally I have considered what it would take to entirely encode a
> substantial table of integrals, series, and math identities, in such a
> system, so I think I have something of a sense of how this should
> differ from a typesetting language (like TeX). Considerably.

Okay, I made a mistake in my terminology.  I should not have
said "markup language for math".  What I should have said is
"something which allows the display of math on the web".  Clearly
I am setting my sights on a different target.  

I want something which allows me to communicate with people, what it
sounds like I am going to get is something which allows me to communicate
with machines (a much harder problem).

-greg


---------------------
Dr. Greg Landrum  (landrum@hal9000.ac.rwth-aachen.de)
Institute of Inorganic Chemistry
Aachen University of Technology
Prof.-Pirlet-Str. 1, D-52074 Aachen, Germany
Phone: 049-241-80-7004
Fax: 049-241-8888-288

Received on Tuesday, 27 January 1998 03:30:24 UTC