Camelia, the Perl 6 bug

IRC log for #cdk, 2010-05-16

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

All times shown according to UTC.

Time Nick Message
03:40 s9asad joined #cdk
07:37 jbrefort joined #cdk
09:08 egonw joined #cdk
09:50 egonw wtfhhtmsfp
09:52 egonw stupid dependencies
10:42 egonw_ joined #cdk
10:42 maclean joined #cdk
10:43 maclean hi egonw how's it going?
10:44 egonw_ hi maclean
10:45 egonw_ too few hours per day :)
10:45 maclean :) So it goes.
10:47 maclean I'm trying to decide whether to try and implement a Schlegel graph layout algorithm for fullerenes or finally do some work I should have been doing for several years.
10:47 egonw :)
10:48 egonw I'd say, pick the one that leads to a publication
10:48 egonw that is, I try to force myself to do that too...
10:49 maclean Hah. Yes. That is a good meta-algorithm.
10:50 maclean do { Step 1: Pick an algorithm that can be published. Step 2 : Implement it. Step 3: Publish.} until(Famous).
10:50 maclean anyway, off to get some food first...
10:52 egonw :)
11:03 maclean_ joined #cdk
12:17 egonw_ joined #cdk
12:27 maclean joined #cdk
13:28 egonw_ maclean: ping
13:28 egonw_ maclean: have you met Rajarshi yet?
13:28 egonw_ he seems to be looking for a pub dinner in Cambridge
14:00 carsten joined #cdk
14:07 maclean joined #cdk
14:07 maclean egonw_ - Right country, wrong city :) I'm in London...
14:07 egonw_ ack
14:08 egonw_ yeah, going down to London from Hinxton is a bit trickier :)
14:08 maclean A bit, but quite near, really. It's a fast rail connection.
14:10 egonw_ yeah, but you'd have to walk down to G-Chesterford or Wittlesford
14:10 egonw_ Hinxton needs a station itself :)
14:10 maclean That's true.
14:10 egonw_ would be so good for EBI/WT/Sanger
14:10 egonw_ but bad for cab companies :)
14:11 maclean :)
14:12 * egonw_ is pretty happy with tripit
14:15 maclean egonw_ : I realised today that it would be quite easy to do random structure generation with signatures as bond constraints.
14:15 maclean Combined with using signatures intead of HOSE codes, for NMR prediction, it would be interesting, I think.
14:15 egonw_ this requires two things, I guess (I have not read the paper yet)
14:15 maclean Unfortunately, I really should be getting on with other things :(
14:15 egonw_ 1. enumarate valid and unique signatures
14:16 egonw_ 2. convert a signature in to a chemical graph
14:16 egonw_ the latter is already working, right?
14:16 maclean Yes.
14:16 maclean But 1) is not necessary.
14:16 maclean You just randomly choose atoms to bond, and then check that they are signature-compatible.
14:16 egonw_ will that garantee generating all structures given a MF?
14:17 maclean Eventually.
14:17 maclean But you would still need to examine and remove duplicates.
14:18 maclean No, the generation process would still be constrained by the match of predicted spectrum to experimental.
14:18 egonw_ ah
14:18 egonw_ Faulon wrote a paper on structure generation based in signatures
14:18 egonw_ I printed it, but not read it...
14:19 egonw_ what's your opinion on that solution?
14:19 maclean I've read it. Many, many times :)
14:19 egonw_ yeah, I'd guess so
14:19 maclean That I can't do it - I'm missing a vital piece. :(
14:19 maclean I have a long, detailed blog post about it, if you are interested...
14:20 maclean :)
14:21 maclean Anyway, that is published work, so it fails the meta-algorithm.
14:21 egonw_ unless you practically apply it of course
14:22 egonw_ sometimes refered to as: "data set - PCA"
14:22 egonw_ PCA has been published a long time ago...
14:22 maclean Principal Components Analysis?
14:22 egonw_ but just apply it on a new data set, and you get new scientific knowledge == publishable
14:22 egonw_ yes, that one
14:22 maclean Oh, right, you mean papers that just take PCA and apply it to some random data set.
14:22 egonw_ so, structgen + new data set == publishable
14:23 egonw_ indeef
14:23 egonw_ indeed
14:23 maclean I guess that happens with many techniques - ANNs, SOMs, ...
14:23 egonw_ quite so
14:41 CIA-30 cdk: Egon Willighagen master * rbe1b70a / (21 files in 10 dirs): (log message trimmed)
14:41 CIA-30 cdk: Patch to generalize the stereo chemistry handling.
14:41 CIA-30 cdk: * introducing IStereoElement, a super class indicating some form of molecular stereo chemistry
14:41 CIA-30 cdk: * IAtomParity extends IStereoElement
14:41 CIA-30 cdk: * IAtomContainer no longer has get/set of IAtomParity
14:41 CIA-30 cdk: * IAtomContainer now takes a list of IStereoElement's
14:41 CIA-30 cdk: * Added the helper method AtomContainerManipulator.ge​tAtomParity(IAtomContainer, IAtom) to
14:50 maclean Hmm. Hope that was right.
14:50 egonw_ that commit?
14:50 egonw_ it's part of the CIP R,S assignment stack
14:51 maclean I tried to pull those changes from upstream.
14:51 egonw_ oh...
14:51 maclean but I may have re-commited them in my name.
14:51 maclean Or something.
14:51 egonw_ :)
14:51 egonw_ yeah, I can recommend to not push things to your master branch
14:52 egonw_ (who came up with those terms, master, pull, ... ?!?!? )
14:52 maclean Hmmm. So should I revert it?
14:52 egonw_ I try to keep my master exactly the same as the upstream master
14:52 egonw_ and put all my work in branches
14:53 egonw_ which I can then always easily rebase on master
14:56 egonw_ Ich bin freu das mein Deutch noch nicht ganz verloren ist
14:56 egonw_ es ist nicht sehr gut, aber doch funktioniert :)
14:57 maclean "freu"?
14:57 egonw_ yeah, that's the one word I am worried about :)
14:58 egonw_ happy
14:58 maclean freülich?
14:58 CIA-30 cdk: Egon Willighagen master * r9abead5 / (4 files in 4 dirs):
14:58 CIA-30 cdk: Implemented stereo chemistry for atoms with four ligands
14:58 CIA-30 cdk: Signed-off-by: Rajarshi Guha <rajarshi.guha@gmail.com> - http://bit.ly/auZf1v
14:59 maclean No, "glücklich", according to google translate.
15:00 egonw_ http://dict.tu-chemnitz.de/dings.cgi?o=302​;service=deen;iservice=de-en-ex;query=freu
15:00 zarah egonw_'s link is also http://tinyurl.com/2w9sras
15:00 egonw_ seems a bit old
15:01 maclean "Don't halloo till you're out of the wood!" :D
15:02 maclean 'halloo' is archaic.
15:02 egonw_ :)
15:02 maclean Or at least, obscure.
15:02 egonw_ but glücklich is not the right word either
15:02 egonw_ it's the correct meaning...
15:02 egonw_ but you would not use that
15:03 egonw_ ah... it's "froh"
15:03 egonw_ http://www.dict.cc/deutsch-englisch/​Ich+bin+froh+das+zu+h%C3%B6ren.html
15:03 zarah egonw_'s link is also http://tinyurl.com/36u8z25
15:03 egonw_ so, it was only phonetically correct :)
15:04 maclean I got that too. Google translate got it from "I am glad"
15:04 egonw_ right
15:05 maclean Er..so what sort of steps do I take to un-screw my github cdk repository (fork?).
15:06 maclean Rebase to remove recent commits, I guess, then make branches in future.
15:08 egonw_ git fetch upstream
15:08 egonw_ git checkout master
15:08 egonw_ git reset upstream/master
15:09 egonw_ git push -f github master
15:09 egonw_ where:
15:09 egonw_ upstream = cdk/cdk
15:09 egonw_ github = gilleain/cdk
15:09 maclean Ah, ok. I'll give that a go (after reading up what they should all do :)
15:09 egonw_ :)
15:11 maclean Vielen dank.
15:12 egonw_ kein Problem :)
15:24 CIA-30 cdk: Egon Willighagen master * r4c2e9b7 / src/test/org/openscience/cdk/​smiles/SmilesParserTest.java :
15:24 CIA-30 cdk: Added unit tests with various chiral SMILES situations.
15:24 CIA-30 cdk: * ring closures
15:24 CIA-30 cdk: * disconnections
15:24 CIA-30 cdk: * implicit/explicit hydrogens
15:24 CIA-30 cdk: Signed-off-by: Rajarshi Guha <rajarshi.guha@gmail.com> - http://bit.ly/cyndFm
15:24 CIA-30 cdk: Egon Willighagen master * r2dd575e / src/main/org/openscience/cd​k/smiles/SmilesParser.java :
15:24 CIA-30 cdk: SMILES @ and @@ chiralities are now fully read.
15:24 CIA-30 cdk: * takes into account branching
15:24 CIA-30 cdk: * implicit hydrogens on chiral atoms are made explicit
15:24 CIA-30 cdk: Signed-off-by: Rajarshi Guha <rajarshi.guha@gmail.com> - http://bit.ly/dcSlcd
15:24 CIA-30 cdk: Egon Willighagen master * r0f4649c / src/test/org/openscience/cdk/​smiles/SmilesParserTest.java :
15:24 CIA-30 cdk: Added two more unit tests, related to ring closing
15:24 CIA-30 cdk: Signed-off-by: Rajarshi Guha <rajarshi.guha@gmail.com> - http://bit.ly/cM9yhu
15:24 CIA-30 cdk: Egon Willighagen master * r9cbc242 / src/test/org/openscience/cdk/​smiles/SmilesParserTest.java :
15:24 CIA-30 cdk: Added two test cases by Daniel from my blog: http://chem-bla-ics.blogspot.com/2010/05/​cip-rules-2-parsing-and-from-smiles.html
15:24 zarah CIA-30's link is also http://tinyurl.com/22pxhag
17:14 sneumann joined #cdk
17:37 egonw_ joined #cdk
18:10 jbrefort joined #cdk
19:11 sneumann joined #cdk
20:02 s9asad joined #cdk
21:54 sneumann joined #cdk

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