"Réflexions" - langages scientifiques

[Robert Keeble <RKeeble@xxxxxxxxx>]

Salut,

It's Friday, so I'm taking time out to get back to "Réflexions"

We have already talked about some of the ideas in this section, but here are
some general comments and questions, and to some degree, a summary. Sorry
about the dull, specification-like language, but it  is more compact.

(In general, by "formula" I mean Math, and by "expression", any assemblage
of symbols in some notational system)

The base application should support scientific languages of all sorts and
allow arbitrary adjustment of the symbols in any particular scientific
layout. It should allow relative positioning of the symbols in a formula or
expression. Allow coarse positioning, but also positioning at the finest
increment possible. The base application need not provide the most efficient
ways of manipulating the formulas or expressions, but should not prevent any
adjustment or manipulation, and the base architecture should allow
specialized, optional modules tailored to particular fields such as
chemistry or biology. These modules would provide an efficient way to
compose expressions for a particular scientific field. The base application
should provide a macro or markup language that allows direct (non-GUI)
control over positions, and a simple interface to it. It should also provide
an equally capable graphical interface for controlling the layout.

I think a reasonable way to approach the problem would be having a generic
markup language as the internal format, but allow the possibility of parsing
other markup languages like MathML -- with the addition of an optional
module. This will work well for users importing expressions created with
another tool for instance, and make the manual, non-GUI control of
expressions straightforward. It would be much harder to accomodate the GUI
user who does not own an optional Math module, for instance, the person who
creates formulas by adding the right symbols and moving them around until
the expression looks right. I doubt you could easily generate meaningful
MathML from arbitrary groupings of symbols, so it might be difficult to make
these two forms compatible.

Lines of text perhaps aren't meaningful for scientific layout, although
orientation to a global baseline or specific symbol may be useful.

Question: Would it be useful to designate an arbitrary group of characters
(1..n) as the "anchor" (like sigma in a summation) and adjust positions of
other symbols relative to the anchor?

Symbols like the curly brace ("accolades") and the integration sign must
scale correctly.

Graphic designers and generalists don't want to learn some silly, arcane
layout language for scientific expressions.

Mathematicians want to use a formal language they already know, and they
don't want to fiddle with the silly mouse. Especially that hockey puck that
comes with the new Macs ;^)


Rob Keeble
Quark, Inc.