Programming challenges: processing vectors and matrices

[pier4r]

(02-14-2019 12:10 PM)grsbanks Wrote:  Given what a vector actually represents I'm not sure what practical applications these things could have, but on the 50g at least you could always use AXL to convert to a list, do whatever you need on the resulting list and then use AXL to convert back to a vector.

Well exactly that is to be avoided (that is, actually ending up working with lists).
With large vectors it may not pay off.

All this is born from the observation that manipulating large vectors may be faster (orders of magnitude faster) than manipulating large lists. If you do such operations over end over, the difference in time is marked.

For small number of elements (hundreds?), the list data structure is still the best - I think - due to the large amount of functions available.

And in general filling the gap for missing functions on vectors and matrices is a nice challenge.

edit: also the "practical applications" appears once one has the tools. If the same operation makes sense on a list, then it makes sense on the same sequence of number in general, being it represented as a list, a vector/matrix or a string (are there any other built-in data structures in the RPL/RPN world?).

edit2: I also remember to have said "I don't mind if a process is slow, if it is faster than my pace in processing the solutions". It is still the case. If I know I won't be able to process the result of a procedure in the next weeks, the procedure can as well last a week (I have many 50g exactly for this). Nonetheless in the long run striving to optimize this or that bit of code is not bad (in terms of speed, features, reliability, ease of maintenance, ease to read, accuracy, searchability, whateverbility).