greptilian logo

IRC log for #sourcefu, 2015-03-05

http://sourcefu.com

| Channels | #sourcefu index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

All times shown according to UTC.

Time S Nick Message
02:18 searchbot` joined #sourcefu
02:32 pdurbin "Almost every non-trivial software system suffers from the accumulation of technical debt and structural erosion." http://www.nejug.org/events/show/170
02:32 pdurbin sounds uplifting
02:42 sivoais :-)
02:44 pdurbin "OpenSSL is written by scientific people" http://techblog.bozho.net/the-astonishingly-low-quality-of-scientific-code/
02:46 * sivoais knows "The Low Quality of Scientific Code" first hand...
02:46 pdurbin heh
02:46 pdurbin but is it really a problem, I wonder. that post seems pretty rant-y
02:47 sivoais #swcarpentry exists for a reason. :-P
02:47 sivoais Oh, and I work in a CS research lab...
02:48 pdurbin yeah
02:49 sivoais It's more on the science side of things, but I know that the engineering side of CS research isn't much better. I know a person that works in HPC and compilers. He's always venting about the state of the code.
02:50 pdurbin ok, so it's a problem if you're a scientist
02:50 sivoais writing good code as a scientist is thankless since few people read your code
02:51 sivoais they just care about the pretty graphs at the very end
02:52 pdurbin yeah. and it sounds like you just cite the software you used. it doesn't really matter
02:53 sivoais Yep. And not all of that software is openly available. Or has proprietary deps. And data... oh, I could go on and on...
02:53 pdurbin $1000 for a license for JMP, I hear
02:54 sivoais that's really cheap relative to some others
02:56 sivoais I don't remember the cost of ANSYS Fluent (CFD sim software), but the academic license was somewhere around hundred K $ for the engineering school for a version that was probably a decade old.
02:57 pdurbin yikes. must be good stuff
02:59 sivoais It's used heavily in industry. The version we had only ran on Linux, so everyone had to learn to use X11 forwarding. :-)
03:00 sivoais because the software was only on certain remote computers and there was a limit to the number of concurrent licenses
03:00 sivoais Oh, and the GUI for making meshes used the Motif toolkit, iirc
03:03 sivoais OK, if one looked at a scientist working alone, I would have no problem with them writing code however they want... if it works for them, great!
03:03 sivoais but scientists don't work alone ... it takes too much work to replicate every tiny detail
03:05 pdurbin that's what http://www.runmycode.org is for, I guess
03:08 sivoais hmm, I've seen that before, but for some reason I assumed they give you a VM to run your code on...
03:09 sivoais but that's not the case... it just lets you upload your code. Which is a start.
03:20 pdurbin you could ask skay about it: http://irclog.perlgeek.de/crimsonfu/2014-02-27#i_8357091
03:23 sivoais cool :-)
05:35 aditsu joined #sourcefu
06:08 aditsu joined #sourcefu
12:27 pdurbin sivoais: ah, nice, you pinged her and she replied: http://irclog.perlgeek.de/crimsonfu/2015-03-05
12:27 pdurbin so what's next for you? almost done with your CS masters?
12:39 sivoais almost. Just need to finish my thesis. Then hopefully I ace my med school interview in a few months... then on to med school :-)
12:44 pdurbin oh right, I think you mentioned this. good luck!
12:45 pdurbin I have a doctor friend who's getting into bioinformatics. learning linux, compiling software
12:48 sivoais thanks! hehe, awesome!
12:49 pdurbin last I checked he was doing it all in a VM (virtualbox) on his PC laptop
12:52 sivoais are they learning bioinformatics through Coursera? Or Rosalind?
12:54 pdurbin hmm. dunno. a lot of just hacking around, I'm sure
12:55 pdurbin maybe taking a course too
12:56 pdurbin didn't ask
12:57 pdurbin sivoais: should I get him to join this channel and ask you questions? :)
12:58 pdurbin sivoais: here's an example of data he's looking at: http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE34962
12:59 sivoais maybe... I'm not in bioinformatics though! I'm interested in learning it, but right now I'm doing biomedical imaging
12:59 sivoais that's a nice big tar file... :-)
13:00 pdurbin yeah
13:00 sivoais oooh! One thing I liked about #dataverse was that it has MD5s of the files! I don't want to download gigs of bad data (though it would be corrupted anyway)
13:01 pdurbin sivoais: yep! I even expose the md5 from the Search API
13:01 pdurbin we also have a parallel concept of UNF for tabular data
13:03 sivoais *something* normal form ?
13:03 pdurbin searchbot: lucky unf fingerprint
13:03 searchbot pdurbin: http://www.unf.edu/coehs/field-experiences/Fingerprinting.aspx
13:03 pdurbin d'oh!
13:03 pdurbin searchbot: lucky unf fingerprint dataverse
13:03 searchbot pdurbin: http://thedata.org/
13:03 pdurbin sigh
13:03 sivoais hahah
13:03 sivoais Universal Numerical Fingerprint
13:04 sivoais <http://dataverse.readthedocs.org/en/latest/Data_Citation/unf.html>
13:04 pdurbin Universal Numerical Fingerprint (UNF) — Dataverse.org - http://guides.dataverse.org/en/latest/developers/unf/
13:05 pdurbin huh. those readthedocs files are stale
13:07 sivoais isn't it supposed to build on push?
13:08 sivoais slightly related, I've wondered about using a webhook to build ikiwiki pages on a GitHub gh-pages branch
13:08 pdurbin no idea who set that up
13:09 pdurbin sivoais: I have ikiwiki doing a build every time I push
13:10 sivoais but does it do that on GitHub? :-) I'm trying to see if that would work as alternative to hosting it myself.
13:10 sivoais and allowing others to edit
13:11 pdurbin https://github.com/pdurbin/greptilian-vagrant/blob/master/modules/server1/files/usr/local/greptilian/bin/wiki-setup
13:11 sivoais And get the benefit of making it easier to contribute, more visibility
13:12 pdurbin my setup doesn't involve github
13:14 pdurbin there's a git hook in https://github.com/pdurbin/wiki/blob/master/wiki.setup
13:16 sivoais hehe, it seems you use your wiki like I use my vimoutliner files. Just put all your notes in it. :-)
13:21 pdurbin it's pretty... random
13:21 pdurbin but I like ikiwiki
13:21 pdurbin it lets me put stuff into folders
13:23 sivoais My workflow is: first, use vimoutliner for things that I haven't yet understood fully, then move those to a private ikiwiki to write out the ideas in longer form
13:23 pdurbin ah
13:24 pdurbin this channel is where I put things I haven't yet understood (obviously) :)
13:24 sivoais hahah, yeah, IRC is good for that too! :-D
13:25 pdurbin sivoais: see also http://irclog.greptilian.com/pdurbin/2012-11-28
13:28 sivoais oh, I saw this before... it inspired me to create my own, but turns out that I used it more for venting about bad code than actual notes. Kind of like Twitter...
13:28 sivoais :-P
13:30 sivoais ooh, O'Reilly has a "Learning iPython Notebook" course
13:34 sivoais I've been using IPython notebook to write up ideas now. ikiwiki integration for that would be wonderful...
13:56 pdurbin sivoais: I stopped using and logging #pdurbin a few months later. failed experiment
21:57 sivoais this is interesting: <https://notablemind.org/>, <https://app.notablemind.org/tutorial/>
21:57 sivoais via the IPython-dev mailing list
21:57 sivoais <http://mail.scipy.org/pipermail/ipython-dev/2015-March/016056.html>

| Channels | #sourcefu index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

http://sourcefu.com