Why do you think the "factor" command exists?

dggoldst · 09-01-2006, 07:21 AM

I've noticed there's a command line call "factor", which prints the prime factors of a number.

Does anybody have an idea why this is included as a standard Unix/Linux routine? Is it really that important to do prime factorizations on the command line? Most programming languages don't have a factor command, so why does a very popular OS have one?

There must be some interesting history to this I'm missing. The only use I've gotten out of factor is getting a rough comparion of processor speeds on my various boxes by doing
factor 18446744073709551557

...which takes from 52 seconds to over 2 minutes, depending on the machine

Yours in curiosity,
Dan

jlliagre · 09-01-2006, 07:48 AM

Well, I think it exists because it has always been there.

The venerable Unix Version 7 manual (Jan 1979) has it and states:

Maximum time to factor is proportional to √n and occurs when n is prime or the square of a prime. It takes 1 minute to factor a prime near 10 14 on a PDP11.

It improves since that time (x240), but not as much as I was expecting:

Code:

time factor 99999999999971
99999999999971
     99999999999971


real    0m0.25s
user    0m0.17s
sys     0m0.00s

dggoldst · 09-01-2006, 08:23 AM

Quote:

Originally Posted by jlliagre

Well, I think it exists because it has always been there.

Oui, exactement. But why did someone put it there in the first place?

jlliagre · 09-01-2006, 08:38 AM

The benchmark theory may be correct, I have never seen this command actually used in a script.

Note that OpenSolaris is providing a source code which is certainly very close to the original one. There are no comments about the intended use of it though.

http://cvs.opensolaris.org/source/xr...actor/factor.c

Gnu/Linux is having a different implementation without the original 10^14 limit, so these guys took their job seriously ...

dggoldst · 09-15-2006, 12:31 PM

Any other thoughts on this?

J.W. · 09-15-2006, 04:31 PM

I would guess that it's there because at the time Unix was originally developed around 1970, the only real market for computers was large, well-funded organizations, and at that time, number crunching was one of a computer's primary uses. Factoring is a useful function for certain areas of study (eg, in scientific labs, universities, etc) and by including a simple-to-use utility to compute factors in Unix, it may have been a minor but useful selling point. Sort of like cup holders in a car - they aren't really necessary, but to a potential customer, having them is better than not having them. Same with a factor command. Just speculation, of course