From: Richard F. <fa...@gm...> - 2023-01-18 18:40:18
|
Apologies for taking space on this discussion list for something that has already consumed too many messages. Large expressions (several pages long) are difficult to comprehend whether in '2d" or linear form. Looking for the "/" to see where the denominator begins, or searching for the matching ")" may be important and yet difficult. What to do? There used to be (but I can't find it!) a program to abbreviate the display. Thus we would see something like E = sum of 140 terms (Which can be examined by E[1], ... E[140]...)) or E= product of... or quotient... etc. An alternative to E[140] might be something like ratcoef(E, x, 140) etc. This kind of abbreviation was used recursively, e.g. E = ( product of 50 terms) + (product of 40 terms) + .. Linear output (as used by grind, or display2d:false) uses far less memory and computing, and so is faster and doesn't break things so much. But really, other that to show we can do it, why display an expression explicitly that is many pages long? I recall that in the past I proposed a program that produced a spreadsheet instead of a conventional display. That is coefficient | x exponent | y exponent | .... 12345 3 4 ... which is perhaps of some use to people who are more familiar with spreadsheets than (say) multi-dimensional arrays.. Looking for the program, I came across Bruce Miller's paper on formatting expressions in Macsyma in various ways... https://nvlpubs.nist.gov/nistpubs/Legacy/IR/nistir5618.pdf Even lacking access to the source code it should not be too challenging to (re-)implement these designs, should someone take on this project. |