Camelia, the Perl 6 bug

IRC log for #cdk, 2010-10-01

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

All times shown according to UTC.

Time Nick Message
03:59 azeem left #cdk
04:00 azeem joined #cdk
04:54 egonw joined #cdk
05:29 sneumann joined #cdk
06:04 bag_ joined #cdk
06:24 bag_ left #cdk
06:41 bag_ joined #cdk
06:45 s_wolf joined #cdk
06:50 jbrefort joined #cdk
06:59 Gpox joined #cdk
07:02 bag_ left #cdk
07:06 maclean joined #cdk
07:07 maclean morning
07:09 maclean Ah bug reports which are actually (possibly) feature requests....
07:09 maclean "am creating a  database so i would like to incorporate Jmol into the database but for that i have to write a script which would should each molecule when clicked upon it."
07:21 egonw hej maclean
07:21 maclean hej
07:21 egonw yeah, actually sounds like a support request to me :)
07:22 maclean It also has a "accidentally the whole thing" error (IE : a missing word).
07:25 * egonw is install the eclipse maven2 plugin (again)
07:26 egonw but now in a separate eclipse install
07:26 egonw as I hate that extension...
07:26 egonw (which could change...)
07:26 maclean why, what does it do that's so bad?
07:26 egonw it starts at random time downloading huge amounts of data...
07:26 maclean Ohhh. Yes, this is a maven feature.
07:26 egonw and that is not useful when I am hacking on Bioclipse or CDK that do not really use maven
07:27 egonw yes
07:27 maclean Any use of maven is likely to trigger huge downloads.
07:27 maclean I think that there is a way to turn this off...
07:27 maclean Like in global settings.
07:27 egonw possibly...
07:27 egonw I might look that up...
07:28 egonw OK, there it goes...
07:30 maclean " I agree, it is not pretty. Ideally we could allow to to have "update
07:30 maclean on startup" flag (or "update at such intervals") per repository, but I
07:30 maclean was hesitating to add setting like this in an attempt to keep
07:30 maclean configuration UI simple."
07:30 egonw :)
07:31 maclean Yes, 'ideally' you could add that obvious and necessary feature. But basically you don't want to...
07:31 egonw yeah, it's always good software engeering practice to think what you like best is what the customer wants ...
07:31 * egonw is guilty as charged (wrt to the CDK)
07:31 * maclean too
07:31 egonw who wants scientific accuracy... we want those damn aromatic SMILES
07:31 egonw who wants scientific accuracy, or specification conformance... we want those damn aromatic SMILES
07:32 * egonw wonders of _ilbot2 filters out cursing ...
07:32 egonw zarah: do you do RTs ?
07:32 zarah egonw: are you talking to *me*?
07:33 egonw zarah: would be cool if you could have a command @rt
07:33 zarah egonw: eh... wha'?
07:33 egonw @tell masak can I haz @rt ?
07:33 zarah Consider it noted.
07:33 maclean retweet?
07:33 maclean oh nonono.
07:33 egonw :)
07:33 egonw yeah, it might scare away people
07:34 egonw with these logs... no one knows that anything they say here is public anyway
07:34 egonw who cares about reading channel topics...
07:34 maclean ah. channel topic. that actually answers a question I had.
07:42 slyrus_ can i haz decent 2d rendering? :)
07:42 maclean >:|
07:43 maclean Writing unit tests and javadocs is hard...
07:43 maclean But yes, I think many of us (including me) want decent 2d rendering.
07:43 egonw slyrus_: we are working on it...
07:43 slyrus_ yay!
07:43 egonw but, say, 65k euro would help a lot
07:43 egonw allowing us to put someone on it full time
07:45 maclean that's like 56K in real mo...er in pounds !
07:45 maclean Quite a bit.
07:47 maclean Also, I like the range of google queries that end "65,000 ..." such a "65,000 a year amount per month" :)
07:49 egonw that's with overhead etc...
07:49 egonw not what a person would earn...
07:54 slyrus_ is there an academic paper to be had from a more fully-functioning 2d renderer or is that "solved"?
07:54 egonw yeah, there is some related literature
07:54 egonw the JChemPaint paper from 2000
07:54 egonw this year a paper by Ertl
07:55 egonw 2-3 years ago a paper on MCDL
07:55 egonw but not so much really...
07:55 slyrus_ so, no, in other words? :)
07:55 egonw it is believed a solved problem
07:55 slyrus_ oh, great, since it's solved I'll just wait for the answer then :)
07:55 egonw but for many it's also core business, and thus closed source
07:55 slyrus_ ah, right...
07:55 egonw so, it's both unimportant and important at the same time
07:56 slyrus_ right. the 2 things I've been wanting are good 2d rendering and molecule -> IUPAC names
07:56 slyrus_ both of which exist in commercial libraries but not (well enough for 2d drawing anyway) in CDK
07:58 egonw indeed
07:58 egonw actually, I don't think many companies to IUPAC names either
07:59 slyrus_ chemaxon does, I think
08:02 maclean I think naming is one of those problems that looks easy at the start, then rapidly gets difficult.
08:02 slyrus_ indeed
08:02 slyrus_ sort of like 2-d drawing :)
08:03 maclean Like "Oh well there's alkanes, and alkenes" but then "oh and bridged cycles, and spira-fused rings and..."
08:03 maclean yes, much like 2d drawing.
08:03 egonw slyrus_: anyway... CDK has a framework for it...
08:03 egonw and does IUPAC naming...
08:04 egonw but only for a small set of molecules...
08:04 egonw curtosy to Manchester University
08:04 egonw slyrus_: patches welcome
08:04 maclean egonw: unfortunately, that's pretty much the same as "CDK doesn't do it".
08:04 slyrus_ oh?
08:04 maclean It's like pregnancy - there's no "almost works" for some things.
08:05 egonw maclean: indeed... the stress should be on 'framework'
08:05 maclean true.
08:05 slyrus_ where is that?
08:05 egonw slyrus_: looking up details, hang on...
08:06 slyrus_ there's the IUPAC name -> molecule framework, but I haven't found the opposite direction
08:06 slyrus_ not that it's not there...
08:06 egonw oh, bugger...
08:06 egonw indeed...
08:07 egonw Manchester did the *parser*...
08:07 egonw ah, then that generator framework was mine...
08:07 egonw and even less functional...
08:07 slyrus_ oh, cool
08:07 egonw and hence pulled...
08:07 slyrus_ not cool :)
08:07 egonw should be somewhere in the archives...
08:07 egonw would be a nice student project
08:07 egonw but it is so complex...
08:07 egonw the rules contradict
08:08 egonw so, prioritization is important
08:08 egonw the framework did all that...
08:08 egonw but I never had time to persue it further
08:08 * egonw install egit
08:08 slyrus_ anyway, time for some sleep. talk to y'all another night.
08:09 maclean night
08:09 * maclean should get to work
08:10 egonw slyrus_: ack...
08:11 egonw slyrus_: please do ask on blueobelisk.shapado.com about an OS IUPAC name generator...
08:11 egonw slyrus_: perhaps there is critical mass to bootstrap some project
08:16 maclean later.
08:16 maclean left #cdk
09:12 caligula joined #cdk
09:13 caligula need help with standard dev
09:13 caligula left #cdk
09:17 egonw huh?
09:42 s_wolf left #cdk
10:04 maclean joined #cdk
10:15 s_wolf joined #cdk
10:18 s9asad joined #cdk
11:02 sneumann left #cdk
11:23 s9asad left #cdk
11:24 maclean left #cdk
11:46 sneumann joined #cdk
12:59 CIA-98 cdk: Egon Willighagen master * r714c51e / pom.xml : Updated version and confirmed inception year - http://bit.ly/agWAhl
13:02 maclean joined #cdk
13:04 maclean egonw : I just wiped out the molecule layout code :(((
13:05 maclean Oh no! Wait.
13:06 egonw deleted the branch
13:06 egonw ?
13:07 egonw or the whole git folder?
13:07 maclean Actually, git did the right thing for once.
13:08 maclean I cloned a project into the same directory that was already there.
13:08 maclean And...it's preserved the branch I had.
13:09 maclean But git (or eclipse?) has put the other checkout I had into a sub-directory.
13:09 maclean Weird.
13:10 maclean Anyway : false alarm. The moral of the story is... don't try to work in coffee shops.
13:12 egonw no, joints are not good for concentration :)
13:14 maclean Oh. Not a Dutch coffee shop, no :)
13:14 egonw :)
13:15 maclean Oh, and with regard to Taverna-CDK, I think the problem was that the CDK plugin doesn't work with the latest Taverna, not a CDK problem.
13:15 maclean As in, the Taverna API changed, or something, and the CDKplugin wasn't updated.
13:15 egonw but he was using 1.7.2, right?
13:16 egonw the code is being ported to 2.2 ...
13:16 egonw Andreas is working on that, and I believe they will have a workshop soon
13:16 egonw devel workshop, that is
13:17 egonw and it worked on other platforms...
13:17 egonw just not on that vista system
13:17 egonw I hope he saw the question about him using 32 or 64 vista...
13:19 maclean Ok, right yes. Am I not awake today, or something?
13:19 egonw no worries
13:19 egonw I had that earlier this week
13:20 maclean What are you working on today?
13:21 egonw stuff with maven
13:21 egonw e.g. the CDK pom.xml
13:21 egonw in master
13:21 maclean ah cool.
13:21 egonw upon popural request...
13:22 egonw unlike some like to believe, I *can* bend backwards...
13:22 maclean Probably ignore maven when it says it prefers src/main/java and src/test/java.
13:22 maclean As the CDK has just src/java and src/test. I don't think it matters.
13:23 egonw yeah, we don't have that extra java/
13:23 egonw not sure where the pom.xml is supposed to teach maven that...
13:24 maclean hmmm. not sure.
13:25 maclean there's a child element of _build_ called _sourceDirectory_ and _testSourceDirectory_
13:25 maclean sorry, there are two children of...
13:26 maclean I have them just before the plugins element
13:26 egonw ah, indeed!
13:26 egonw found it...tack!
13:26 maclean n/p
13:29 maclean oh god. It's like a neverending nightmare. The zombie Data/data directory.
13:33 maclean nothing works. git reset --hard, git clean -f, git checkout -f -b newbranch
13:34 egonw try on #git
13:36 maclean It's a known problem - the case-sensitivity thing on macs.
13:36 maclean The workaround is to set up a filesystem on a partition or some such crazyness.
13:40 maclean Or, alternatively. Just branch, delete the dir, and forever work on that branch. Sigh....
13:40 egonw (OS/X)--
13:40 egonw I always thought that BSD was case sensitive...
13:42 maclean Well it's not quite BSD. Or some parts are, and some are not.
13:42 maclean I guess the filesystem is not the same.
13:43 egonw indeed...
13:43 egonw that's probably more significant
14:01 sneumann left #cdk
14:05 s9asad joined #cdk
15:32 s_wolf left #cdk
15:33 maclean haha! the day is looking up, as I fix a bug.
15:33 egonw maclean++
15:33 * egonw is going offline for now
15:33 maclean ok, cu
15:44 egonw left #cdk
15:50 slyrus_ maclean: you can repartition your hard drive with case-sensitive HFS+ or whatever they call it
15:51 maclean I probably /can/ (with much cursing, I expect) - but do I really want to? :)
15:51 slyrus_ heh, well, no
15:52 maclean It's one of those issues that occurs infrequently enough that by the next time it comes up, I will have forgotten a) why it happens and b) how to fix it.
15:55 * maclean sometimes wonders if people troll software mailing list with bizarre questions.
15:58 maclean I mean "I am working on a database where there are several molecules in the drop down menu..." ?
16:16 egonw joined #cdk
16:22 maclean egonw : I assigned that signature/aromatic patch to you - do feel free to re-assign it :)
16:23 egonw ok
16:23 egonw yeah, will try to review some patches this weekend
16:24 egonw no promises though
16:25 maclean anytime. It's only necessary that it exist in the patch list, not that it be reviewed quickly.
16:31 egonw left #cdk
16:46 maclean layout code :
16:46 maclean http://github.com/gilleain/cdk/commit/5​7a3d5432faa56783c1353a67898f0cb74681517
16:46 zarah maclean's link is also http://tinyurl.com/2wg4tth
16:49 s9asad left #cdk
17:07 maclean left #cdk
17:15 Gpox left #cdk
18:12 mgerlich left #cdk
21:18 jbrefort left #cdk
23:36 bag__ joined #cdk
23:48 bag__ left #cdk

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