14:59 <msiism> I've implemented command-line parsing using lists. But now I wonder whether vectors might be better suited for the job.

14:59 <msiism> One potential problem I see is the speed of going through a list vs. going through a vector.

15:01 <msiism> I guess that, if you have a lot of operands, storing them in a list and then going through that list will tend to be slower than doing all that vector-based.

15:02 <msiism> There's probably a good reason, `current-command-line-arguments` returns a vector.

15:03 <msiism> Also, obtaining the length of a list probably causes much more overhead than the same operation on a vector.

15:05 <ncf> going through a list exactly once should take as much time as going through a vector exactly once, in theory

15:06 <msiism> I see. But when you need the length, my understanding is that that means going through the whole thing for lists, but not for vectors because they have a fixed length.

15:07 <ncf> right, getting the length of a list will take O(n) operations, while the length of a vector would take O(1)

15:08 <msiism> Okay.

15:09 <msiism> I was thinking that going through a list would take longer than going through a vector because a list might be spread out over a comparably big area of memory, while a vector is stored in consecutive cells.

15:13 <ncf> yes, in practice there are reasons it might not be quite as fast (although the asymptotic complexity is the same); i don't think this should matter much for command line processing, though

15:14 <ncf> but, well, if the built-in functions already return vectors, might as well go with that

15:15 <msiism> Yeah, that's what I thought. I probably shouldn't be converting to a list in the first place.