Camelia, the Perl 6 bug

IRC log for #bioperl, 2010-03-10

| Channels | #bioperl index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:08 Forrest joined #bioperl
01:24 brunov joined #bioperl
03:12 kraut_ joined #bioperl
03:13 brunov joined #bioperl
03:22 neL joined #bioperl
03:27 flu joined #bioperl
05:10 shellfish_gene joined #bioperl
05:28 dnewkirk left #bioperl
06:58 Furiiku joined #bioperl
07:14 bag_ joined #bioperl
07:33 balin joined #bioperl
09:29 neL left #bioperl
15:48 deafferret rbuels: I tend to just create stupidly small .tt files.  :)   https://clabsvn.ist.unomaha.edu​/anonsvn/CLAB/RT884/run.pbs.tt
15:49 rbuels deafferret: you are a busy man lately
15:49 deafferret rbuels: what? sorry, too busy right now
15:50 rbuels heh
15:50 deafferret rbuels: what? sorry, too busy right now
16:24 driley joined #bioperl
16:43 brunov joined #bioperl
17:23 brunov joined #bioperl
17:32 deafferret rbuels: !
17:32 deafferret 11:32 -!- rbuels [~rob@turkeyberry.sgn.cornell.edu] has left #tt [Leaving]
17:32 deafferret 11:32 <@jhannah> no? first they made the giant rabbit. that got launched back at them with the catapult. then the throw-away line at the
17:32 deafferret end... dammit
17:33 deafferret is "if we construct this giant badger..."
17:33 deafferret then they cut away from the whole storming the castle thing
17:33 deafferret rbuels: what? sorry, too busy right now
17:33 rbuels lol
17:33 * rbuels slaps his forehead
17:37 deafferret what the...
17:37 deafferret http://www.ncbi.nlm.nih.gov/BLAST/developer.shtml
17:37 deafferret "The original implementation of BLAST is part of the NCBI C Toolkit. Source code (.tar.gz, .zip) is available. This codebase is deprecated. "
17:37 deafferret where's the tarball of blastall??
17:37 deafferret deprecated? huh?
17:39 deafferret who got peanut butter in my chocolate???
17:40 rbuels blastplus i guess ....
17:40 deafferret the C++ Toolkit was a f'n mess when I was in there 2 years ago
17:45 rbuels biologist: "This software stuff isn't so hard!  See, look, this program runs!"
17:46 rbuels biologist: "We don't need to hire any software guys!"
17:46 rbuels and now, they have an entire codebase that they are rewriting from scratch.
17:50 deafferret what codebase is this then?
17:50 deafferret programmers are like car mechanics
17:50 deafferret 'auto
17:50 rbuels 'This codebase is deprecated'
17:51 rbuels whatever that one was you were pointing at
17:51 deafferret NCBI BLAST?
17:51 rbuels yeah
17:51 deafferret I don't know why NCBI doesn't run their software opensource style
17:51 rbuels cause they don't know how
17:51 deafferret probably cause so much of it is shite
17:52 deafferret embarrasing dirty laundry
17:53 rbuels well also, how many contributors would they get for it
17:53 rbuels there are so few software devs in bio with the competence to participate in an open-source project
17:55 rbuels but SGN is going to start running that way
17:55 deafferret so what? make it public anyway
17:56 deafferret it's tax funded, it should be an open repo
17:56 deafferret if nothing else, it'll make devs think about committing crap
17:56 rbuels yeah
17:56 rbuels heh
17:56 rbuels wonder how many devs they have
17:56 deafferret "boy, this would be embarrassing if someone blogged this patch"
17:57 deafferret maybe I should be proud of my patches! epiphany!
17:58 deafferret too many egos
17:58 deafferret too much reputation protectionism
17:58 rbuels it's probably simpler than that, it's probably just not in their culture
17:59 rbuels my guess is that most of the people there are older than 40
17:59 deafferret and evil
17:59 deafferret EVIL
18:00 deafferret they're hiding drown_the_kittens.c
18:00 deafferret webblast is powered by clubbing baby seals
18:01 deafferret is ensembl in git or something?
18:04 deafferret err drown_the_kittens.cpp. pardon me.
18:05 deafferret .c was deprecated
18:09 deafferret "over 40"? crap, I'm 34. how old are you?
18:33 neL joined #bioperl
18:42 dnewkirk joined #bioperl
19:08 rbuels deafferret: i'm 19
19:09 rbuels JUST KIDDING!  ho ho ho
19:24 brunov I'm doing a immunoinformatics course. It's done in C. I have to learn C now.
19:25 brunov prof. said "perl is slow; a Gibbs classifier written in Perl takes 100 days. It takes 2 says in C. har har"
19:25 brunov and now I have to learn C
19:25 * brunov sobs
19:26 brunov I'll pay 5 bucks to someone that makes perl run as fast as C so I don't have to learn it. Anyone?
19:27 brunov no? dammit.
19:28 dnewkirk tell him you'd rather code it in Java and see what he says :)
19:30 * brunov hurls
19:30 brunov ...please dnewkirk, I'm trying to eat here
19:33 rbuels brunov: i've heard that python has nice c bindings....
19:33 rbuels brunov: you might be able to do at least part of it sanely
19:33 * rbuels does not have any first-hand experience with that
19:34 brunov rbuels, that's I good idea. I've thought of doing the final project in perl with Inline::C for the cpu-heavy parts
19:34 brunov I think it'll be saner than having to parse input files with C
19:34 rbuels oh god
19:34 brunov Inline::C no good?
19:34 rbuels no, was the "parse input files with C" part
19:34 rbuels i've never used Inline::C either
19:34 brunov yeah, I just peeked a future 10-slide ppt about command line parsing in C
19:46 deafferret brunov: if your prof has the C and Perl source code that benchmarks that way I'd love to see it
19:46 brunov deafferret, you think he was overstating that?
19:46 deafferret the last time anyone actually showed me source that was "Perl is slow", it was their fault, not Perls, and I sped it up 30X
19:46 deafferret it completely depends on what you're doing
19:47 deafferret tell him development takes 1000 times longer in C than Perl
19:47 deafferret both those statements are equally unjustified, unless he has the code
19:47 deafferret that benchmarks that way
19:47 deafferret people rarely actually pony up any proof for their use case
19:48 * deafferret har hars
19:49 deafferret C is the right answer for maybe 5% of the code I run
19:49 deafferret lines-of-code wise
19:49 brunov nod. I'll see what these cpu-intensive algorithms are, and try to port them to perl and see
19:49 deafferret sure, it all depends
19:49 brunov exactly. I don't see why I'd have to deal with parsing a fasta file in C. Really, getchar()? shoot me
19:49 deafferret modern GPUs are 1000 times faster than modern CPUs on certain operations... blah blah blah
19:50 deafferret the only reason to parse fasta in C is (1) your teacher is forcing you to (2) the aliens will destroy the Earth unless you shave 5% off the runtime
19:50 deafferret whatever... my tirade is just as silly as his
19:51 * deafferret stops
19:51 brunov for a 2-day-long running process, make that 0.0001% of the running time
19:51 deafferret "Show me the code and benchmarks or shut up"
19:51 deafferret ...which you spent 2 months writing
19:51 deafferret ...and run 4 times, ever, then shelf
19:51 neL left #bioperl
19:52 deafferret neL: I'm sorry! don't leave!
19:52 * deafferret clocks back in
19:52 brunov exactly. I'm actually looking forward to being forced to learn C, but there's no way I'm splitting text files with it
20:58 kyanardag_ joined #bioperl
21:08 kyanardag joined #bioperl
21:16 Biome joined #bioperl
21:16 Biome biome: latest_moose Chris Fields * a55970d (1 files in 1 dirs): work in latest Moose fixes (trivial) and add a few classes - http://bit.ly/bDNFMU
21:16 Biome left #bioperl
21:17 Biome joined #bioperl
21:17 Biome biome: master Chris Fields * a55970d (1 files in 1 dirs): work in latest Moose fixes (trivial) and add a few classes -
21:17 Biome left #bioperl
21:17 dnewkirk it's not all bad :P
21:18 deafferret Biome: what repo are you
21:18 deafferret ?
21:19 Biome joined #bioperl
21:19 Biome biome: master Chris Fields * 67d74a3 (1 files in 1 dirs): cluster-related roles - http://bit.ly/98ZEzM
21:19 Biome left #bioperl
21:19 rbuels deafferret: cjfields on github
21:20 deafferret Biome: thanks
21:20 rbuels deafferret: you could just click the link, those work  :-)
21:21 deafferret :)   I now have Pavlovian training not to trust links for 15m
21:30 rbuels oh, is that what causes you to lool?
21:34 * deafferret lools
21:44 bag joined #bioperl
23:16 brunov joined #bioperl

| Channels | #bioperl index | Today | | Search | Google Search | Plain-Text | summary