Camelia, the Perl 6 bug

IRC log for #bioperl, 2009-12-03

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

All times shown according to UTC.

Time Nick Message
00:48 brunov joined #bioperl
01:47 ptl joined #bioperl
03:08 ptl_ joined #bioperl
03:39 ptl joined #bioperl
03:55 ptl_ joined #bioperl
04:37 ptl joined #bioperl
08:08 bag joined #bioperl
09:04 faceface google do mix ads with 'hits' these days, which is why they were popular in the old days (because they didn't)
09:11 faceface Don't be ghettoized
10:55 faceface hi
10:55 faceface can someone test this script: http://www.bioperl.org/wiki/Merging_sepa​rate_sequence_and_quality_files_to_FASTQ
10:55 faceface with zero length sequence / quality file?
10:55 faceface outputs an extra newline in the qual
10:56 faceface I mean in the fastq
11:02 xdg joined #bioperl
11:03 xdg joined #bioperl
11:04 xdg Hi, are any of the 1.6.1 alpha developers around?
11:05 faceface xdg: may have to wait, but they should turn up
11:06 xdg faceface, thanks.  I tried early in case any were in Europe/Asia and will check later when more of the US is awake
11:10 faceface right
11:11 faceface rbuels: will be able to identify the appropriate devs (when he wakes up!) ;-)
11:11 faceface timzone?
11:11 xdg I'm in NYC, so US/Eastern
11:14 faceface cool
11:14 faceface europe here
11:16 xdg I'm the current release manager for Module::Build and some recent dev changes look like they break the custom Module::Build subclass in BioPerl.
11:17 faceface ic
11:17 faceface I'm just a user
11:17 xdg I don't want to break BioPerl, so I want to discuss a plan with them.
11:18 faceface at one point I tried to debug the "t/LocalDB/BioDBGFF.......​....................1/279" bug, but I couldn't track it down
11:18 faceface (it looks for localhost even if you tell it to use some other MySQL server)
11:18 xdg that's annoying
11:19 faceface xdg: prolly worth hiting the mailing list, this channel isn't well represented relative to the number of ppl there
11:19 xdg OK.  I'll try that instead.  Thank you!
11:19 xdg left #bioperl
11:20 faceface but don't leave...
11:51 deafferret joined #bioperl
11:51 faceface joined #bioperl
11:51 rbuels joined #bioperl
11:54 faceface joined #bioperl
13:17 driveby_bot joined #bioperl
13:17 driveby_bot /home/svn-repositories/bioperl: r16435 (maj) : reverse feat order on revcom_with_features: Thanks for the patch Roy--
13:25 brunov joined #bioperl
13:37 brandi joined #bioperl
13:51 brandi1 joined #bioperl
14:58 brunov I feel like screaming in anger now
14:59 brunov so you write an algorithm. It does it job. But you don't completely specify it in the paper you write. The only implementation then it's the one you give, and it sucks
14:59 brunov Because it's poorly packaged and only outputs to a file, and reads a whole database for each input sequence
15:00 brunov so it's unscalable (is that a word?)
15:03 brunov Ideally, I'd like it to be a thin wrapper over a library that I can use programatically
15:03 flu joined #bioperl
15:04 brunov but I am willing to settle with something that 1) accepts multiple sequences as input, 2) writes to standard output in some standard-ish format
15:05 brandi1 left #bioperl
15:06 brunov </rant>
15:23 driveby_bot joined #bioperl
15:23 driveby_bot /home/svn-repositories/bioperl: r16436 (maj) : _collate_subcmd_args() really belongs here...
15:27 driveby_bot joined #bioperl
15:27 driveby_bot /home/svn-repositories/bioperl: r16437 (maj) : tweak
15:58 driveby_bot joined #bioperl
15:58 driveby_bot /home/svn-repositories/bioperl: r16438 (maj) : fixed mixed dash issue
16:19 rbuels brunov: welcome to the world of non-programmers trying to write software
16:20 brunov thank you, it sucks
16:21 brunov maybe they have some arrangement with hard drive manufacturers
16:21 brunov "let's make that bitch spin!"
17:25 faceface hey rbuels
17:25 faceface time to talk tomato?
17:26 * faceface migrates himself dynamically to #gmod
17:34 * deafferret feeds the phant
17:36 deafferret brunov: what source code are you referring to?
18:38 * deafferret strokes his @ lovingly
18:39 rbuels :-X
18:46 brunov deafferret, some programs that predict solubility and secondary structure
18:49 deafferret brunov: so the paper make the software appear useful, then the software itself is immature?
18:49 deafferret is it in an open repository so you can patch it?
18:49 rbuels oh these naive questions
18:49 * deafferret deals mostly with abandonware from grad students that have left
18:50 brunov yes. No. There's no VCS, poor packaging ("copy the library files to this directory, change your home directory in configure.pl"), inflexible interface
18:50 deafferret "but, but, the paper says the software does THIS. and THAT is not happening. fix it!"
18:51 deafferret me: Oh, sure. I'll take 4 years of grad student spaghetti and make it actually cure cancer this time, in 4 hours"
18:51 brunov LOL
18:52 faceface DSSP
18:52 deafferret code which, by the way, NO ONE HAS EVER ACTUALLY LOOKED AT OR USED except the grad student him/herself
18:52 brunov this is what I'm talking about: http://pastie.org/726065
18:53 brunov it makes me want to stab myself in the eye
18:53 brunov eyes, even
18:54 faceface Don't modify the code below
18:54 deafferret COMMENTS?! YOU LUCKY BASTARD!
18:54 deafferret lol
18:54 brunov It's still unusable
18:54 faceface I don't know why people write such shoddy perl.. I'm betting when the world has gone over to python, it'll be just as shoddy
18:54 deafferret rbuels told me I'm not allowed to open(SOLPRO_SH any more  :(
18:56 deafferret that code looks a lot better than what I usually get thrown into
18:56 deafferret Perhaps UC Irvine > U Neb Omaha?  :)
18:56 brunov It's still horrible "generate solpro.sh script"
19:20 * deafferret is guilty of generating .sh files on the file for the purposes of submitting jobs to our clusters  :)
19:20 deafferret one the fly
19:30 brunov isn't it better to have a script that takes arguments?
19:30 brunov or a config file?
19:31 deafferret yes, but that's not how @admins want us to use qsub
19:32 deafferret http://en.wikipedia.org/wiki/Portable_Batch_System
19:33 deafferret Firefly, our 2000 node cluster, has 3 different schedulers, but I've only used PBS so far
19:33 brunov oh, so they have to be runnable without args then?
19:34 rbuels yeah, qsub-like things are kind of....
19:34 rbuels eh.
19:34 rbuels Torque, PBS, Sun Grid Engine, etc
19:34 rbuels not my favorite.
19:34 deafferret um, yes? the idea is that job1.sh is NFS-mounted on all the cluster nodes, and the scheduler picks a node and runs it
19:34 rbuels we use torque at SGN too
19:34 rbuels and i have a perl module that generates a shell script that contains a perl script for submitting jobs to our cluster
19:35 brunov ok, then maybe that's why this does it too
19:35 rbuels and what's more, the perl script inside the shell script uses the module that did the generating.
19:35 brunov I take back 10% of my rant
19:35 deafferret "Firefly" has 3 different scheduler head nodes: Sun Grid Engine, PBS Pro, Moab
19:36 brunov rbuels, aren't you worried that it'll generate a singularity in the universe and doom us all?
19:36 deafferret mmm.... Perl -> shell -> Perl -> C (blastall)
19:37 deafferret kinda like my Perl -> TT -> Javascript -> dynamic HTML
19:37 deafferret brunov: fear not. I'm keeping an eye on him.
19:37 brunov tame the beast
19:38 deafferret the fastest way to his heart is through his rib cage
19:38 brunov but he's protected by his iPhone!
19:38 deafferret I've got a counter-App for that
19:38 brunov http://www.mobilecrunch.com/2009/0​9/16/dont-worry-i-have-an-iphone/
19:39 deafferret lol
19:40 brunov I like the part where he heals him with his phone
19:40 deafferret hmm... missed that part. /me looks again
19:40 * deafferret was never good at research
20:56 brunov thinking is hard
20:56 rbuels it's significantly easier if you reverse the two:  then you can just hardly think
23:08 ptl_ joined #bioperl
23:33 ptl joined #bioperl

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