MMA 3.0 The new is wearing off
- To: mathgroup at smc.vnet.net
- Subject: [mg7084] MMA 3.0 The new is wearing off
- From: sch at mitre.org (Stu Schaffner)
- Date: Wed, 7 May 1997 01:58:10 -0400 (EDT)
- Organization: The MITRE Corp.
- Sender: owner-wri-mathgroup at wolfram.com
I eagerly awaited MMA 3, and was very pleased with its increase in
speed and the snazzy new formatting. However, when I converted some
of my old programs none of them worked. The primary reason seemed to
be that I had format statements that displayed complex data structures
as text, and the converter substituted that text in all of my Modules,
even in the argument lists. The converter also threw away the comments
I had carefully added to each module specification and made an
incomprehensible hash of the indentation. I put MMA back on the shelf
for awhile until I could find the time to pick up the pieces.
I am now more-or-less resigned to losing much of my prior work, but I
am still having trouble using MMA 3.0 at all to write well-structured
libraries of function definitions. I'm not talking about using Do loops,
but simply writing, commenting, and altering ordinary functional programming
code. I'm not experienced enough in MMA 3.0 yet to even be able to
fully characterize the source of my difficulties, but here are a few
thoughts:
1. The cell conversion algorithms seem to lose a lot of information.
I can usually reduce a cell to garbage with a few conversions
between StandardForm and InputForm and back again.
2. Almost any attempts to use indenting to show module structure
produce a blizzard of invisible Cells and other structural
gobbledegook. The only way to see these little "helpers" is to
Show Expression, but the full expression is so complex that
it doesn't help much.
3. The invisible Cells make cutting and pasting a harrowing
experience. It is often difficult to even select the proper
sequence of code because the selection often overlaps some
counterintuitive Cell structure. Pasting such a collection
of cells often has strange effects on the hidden structure
MMA has placed the rest of the cell in.
4. I thought "Why not use the formatting to advantage? Let's produce
a formatting function for function definitions." This didn't
work because the contents of a Cell apparently have to be a valid
expression. Thus, you can't (or I don't know how to) put
"foo[ a_ ] := Module[" into a single cell.
I sure would appreciate some suggestions.
Stu Schaffner,
The MITRE Corp.
sch at mitre.org