freenode/#lisp - IRC Chatlog
Search
4:12:59
shrdlu68
ACTION hasn't slept all night and has made the mistake of taking coffee in the morning.
4:40:06
jasom
my dad used to say that he didn't use coffee to stay awake for work because he'd have to rewrite all the code after he got some sleep
4:41:07
jasom
instead he added a line to his makefile that would echo a bunch of ^G characters to the terminal after it was done building and nap until the beeps woke him up
7:11:22
Marcy
How can I get slime to load something as soon as it starts? I tried editing my .emacs file and it's not working.
7:30:15
jackdaniel
then you can put anything in ~/.sbclrc (in case of sbcl), or ~/.eclrc (in case of ecl) etc
14:25:59
White_Flame
There's a pretty simple document listing all the prime numbers, but I ran out of harddrive space trying to download it
14:28:05
rtf34
sorry but i just have a problem with my lisp program, it's have to processing about totient phi of euler project
15:00:33
flip214
White_Flame: you should choose some other encoding, the Kolmogorov complexity of prime numbers is quite small.
15:03:27
flip214
White_Flame: no, the gzip format is too limited for that... but there are compression programs that try to build algorithms, perhaps one of these would succeed.
15:04:33
White_Flame
flip214: that would be amazing, because afair there's still no non-search algorithm to generate all the prime numbers, right?
15:05:17
flip214
White_Flame: well, sieve of eratosthenes wouldn't be called searching, would it? they just come out, one after another.
15:06:41
flip214
White_Flame: if you start from 2, no guessing is involved. if you have enough time and storage, that is.
15:07:27
White_Flame
or put differently, a purely mathematical function to compute the next prime given a prior, as opposed to an algorithm
15:09:01
Bike
whether an algorithm to compute the function proceeds by "testing candidates" doesn't matter
15:09:47
White_Flame
as mentioned, we can describe how to find all prime numbers, just not feasible mean
15:12:05
Bike
and i don't know why the time complexity even matters given that you're talking about computing all primes, which would obviously take forever
15:13:09
White_Flame
well, give me an algorithm to compute the Nth prime in constant time, and I'll be happy that primes are "solved" :)
15:13:22
flip214
Bike: well, if computers keep speeding up by some constant factor every year, it won't.
15:16:12
flip214
we wondered why the machines didn't accept the PCI devices, a reboot ago they were still present?
15:20:14
Bike
anyway, AKS is polynomial in number of digits, which is obviously pretty tractable asymptotically. though the useful algorithms are higher O but quicker in practice. but most of the "biggest primes" we fine are mersenne, which we describe by an exponential power, so naturally it's going to get harder and harder quickly
15:21:13
flip214
well, IMO the randomly-generated primes for cryptography are more important than some "size matters" findings
15:22:23
flip214
still, I guess you'd be generating "my" random primes much earlier than your "big" ones, at least when enumerating them ;)
15:26:19
flip214
_death: if we'd only skip the randomness part, and actually start generating them one-by-one, this wouldn't happen!
15:26:50
_death
flip214: sure.. here's a program to generate all primes: (loop for i from 2 do (output i)) ;)