Camelia, the Perl 6 bug

IRC log for #cdk, 2007-08-12

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

All times shown according to UTC.

Time Nick Message
05:43 jbrefort joined #cdk
06:25 egonw joined #cdk
06:25 egonw moin
06:25 egonw jbrefort: let me know when you're around
06:26 jbrefort egonw, I'm there
06:26 egonw but it write access would be handy to maintain the space group list, or not?
06:26 egonw in principe, there is not much work to be done anyway...
06:27 jbrefort yes, if it needs any change
06:27 egonw you got a SF account?
06:27 jbrefort and I'll probably add data for point groups (character tables)
06:27 jbrefort yes, jbrefort
06:28 egonw ok, you're added
06:28 jbrefort thanks
06:39 jbrefort egonw, install path for the file? $(pkgdatadir)/groups or just $(pkgdatadir)?
06:39 egonw mom... brb
06:48 egonw ${pkgdatadir} would be fine
06:48 egonw btw, please add alsoa bibxml file...
06:48 egonw citing the sources of the information...
06:49 egonw put things in trunk/bodr/crystal
06:49 egonw (elements/crystalstructures.* should move there too...)
06:50 jbrefort ok
06:59 jbrefort egonw, it's the first timeI play with bibxml, does the following seems correct?
06:59 jbrefort <bx:entry>
06:59 jbrefort  <bx:misc>
06:59 jbrefort    <bx:url>http://cci.lbl.gov/sginfo/hall_symbols.html</bx:url>
06:59 jbrefort    <bx:title>Concise Space-Group Symbols</bx:title>
06:59 jbrefort  </bx:misc>
06:59 jbrefort  <bx:misc>
06:59 jbrefort    <bx:url>http://www.mkmc.dk/crystal/group.html</bx:url>
06:59 jbrefort  </bx:misc>
06:59 jbrefort  <bx:misc>
06:59 jbrefort    <bx:url>http://cci.lbl.gov/cctbx/explore_symmetry.html</bx:url>
06:59 jbrefort    <bx:title>cctbx - Explore symmetry</bx:title>
06:59 jbrefort  </bx:misc>
06:59 jbrefort </bx:entry>
07:00 egonw yes, but please add month/year too... so that we have an idea when the data was extracted
07:00 egonw oh...
07:00 egonw and each misc should be in a separate entry
07:02 jbrefort ok
07:03 jbrefort the month/year entry is missing in most bibxml files in bodr
07:03 jbrefort the tags are bx:year and bx:month?
07:04 egonw yes
07:11 jbrefort <bx:entry>
07:11 jbrefort  <bx:misc>
07:11 jbrefort    <bx:url>http://cci.lbl.gov/sginfo/hall_symbols.html</bx:url>
07:11 jbrefort    <bx:title>Concise Space-Group Symbols</bx:title>
07:11 jbrefort <bx:author>Sydney R. Hall<bx:author>
07:11 jbrefort <bx:author>Ralf W. Grosse-Kunstleve<bx:author>
07:11 jbrefort    <bx:year>2007</bx:year>
07:11 jbrefort    <bx:month>August</bx:month>
07:11 jbrefort  </bx:misc>
07:11 jbrefort </bx:entry>
07:11 jbrefort <bx:entry>
07:11 jbrefort  <bx:misc>
07:11 jbrefort    <bx:url>http://www.mkmc.dk/crystal/group.html</bx:url>
07:11 jbrefort    <bx:year>2007</bx:year>
07:11 jbrefort    <bx:month>August</bx:month>
07:11 jbrefort  </bx:misc>
07:11 egonw looks good
07:11 jbrefort </bx:entry>
07:11 jbrefort <bx:entry>
07:11 jbrefort  <bx:misc>
07:11 jbrefort    <bx:url>http://cci.lbl.gov/cctbx/explore_symmetry.html</bx:url>
07:11 jbrefort    <bx:title>cctbx - Explore symmetry</bx:title>
07:11 jbrefort <bx:author>Per Stoltze<bx:author>
07:12 jbrefort    <bx:year>2007</bx:year>
07:12 jbrefort    <bx:month>August</bx:month>
07:12 jbrefort  </bx:misc>
07:12 jbrefort </bx:entry>
07:12 jbrefort correct?
07:19 CIA-1 cdk: jbrefort * r3 /trunk/bodr/ (7 files in 2 dirs):
07:19 CIA-1 cdk: 2007-08-12 Jean Brefort <jean.brefort@normalesup.org>
07:19 CIA-1 cdk:  * Makefile.am: add sace groups descriptions.
07:19 CIA-1 cdk:  * configure.ac: ditto.
07:19 CIA-1 cdk:  * crystal/Makefile.am: ditto.
07:19 CIA-1 cdk:  * crystal/space-groups.bibxml: ditto.
07:19 CIA-1 cdk:  * crystal/space-groups.xml: ditto.
07:41 cpyang joined #cdk
07:44 cpyang joined #cdk
07:59 neksa_ joined #cdk
08:00 neksa_ changed the nick to neksa
08:03 neksa hi egonw
08:03 egonw hi neksa
08:03 neksa fwd an email from OSRA author
08:05 egonw looks good...
08:06 egonw he's eager to get things done... and very open to discussion...
08:06 egonw btw, about the lack of recall...
08:06 egonw even if 40%... that means quite a lot of useful links between molecules and images/PDFs...
08:11 neksa what about false positives?
08:11 egonw that's more worrying... but those you found will make a good test suite :)
08:11 egonw seriously... this type is data mining will always be error prone...
08:11 neksa anyway, substructure search could perform well ... probably...
08:12 egonw and that's no problem, people should really use InChI, CML files in supplementary info, etc...
08:12 egonw yes, indeed...
08:12 egonw and any molecule we can recover with these tools from literature is good...
08:12 egonw in that sense...
08:12 egonw recall is less important than precision
08:14 neksa I just took the appendix with 64 molecules from my CUBIC thesis ;)   I can freely use them in our test repository
08:15 * neksa went for coffee
08:18 egonw nice
08:18 egonw send Igor a copy too then
08:29 neksa thanks for the call in BO list
08:30 egonw yeah, it's really outside the scope of your project to get that done, but is important...
08:30 egonw there should be a few qualified people with packaging knowledge on the list...
08:30 egonw so I hope it is picked up
08:32 neksa can you help to bring OSRA to the SF?
08:33 egonw let's await the response of the NIH legal department...
08:33 carsten joined #cdk
08:33 egonw it's more important to keep development in one place, than to have the source on SF...
08:34 egonw hi carsten: nice email..
08:34 egonw carsten: as you know, I would be rather happy with dependencies on chemical-mime-data :)
08:36 egonw carsten: is your student updating that status page?
08:39 jbrefort it's a pain that OSRA does not export cml instead of SMILES
08:39 jbrefort atoms 2D ositions are lost
08:39 jbrefort +p
08:40 egonw not sure if OSRA picks that up anyway... it could...
08:40 egonw should not be too difficult, as it needs to analyze that anyway...
08:40 egonw BTW, patching the source to output InChI instead of SMILES would be simple...
08:40 egonw have not been able to get it compiled...
08:40 neksa jbrefort: it's not the problem of export. you can change 2 lines in source to export in any OpenBabel supported format
08:41 egonw well, did not really have much time for that anyway...
08:41 * jbrefort will have a closer look later
08:43 neksa egonw: Igor may be right here. poor quality of my sample may be explained by 150dpi res of internal renderer
08:43 egonw yes, could be
08:44 neksa that was: 8 molecules on 8 pages, each page rendered with 150dpi.
09:01 carsten egonw: yes, he is
09:01 carsten egonw: nice mail? I guess that was irony :)
09:01 egonw no, it wasn't ...
09:01 egonw sounds interesting those patches...
09:01 egonw but you're rigth when quoting Linus: show me the code
09:02 carsten egonw: ah, you mean the kpars mail
09:02 carsten egonw: I thought you mean the soc mail
09:02 carsten egonw: that mail to kalzium@ is not about soc. that guy is doing his work on his own :)
09:03 carsten egonw: my student is the author of Step is he dropped from earth
09:03 egonw neksa has had trouble with internet too...
09:03 egonw 20 days is not that bad yet... but good to report it anyway...
09:04 egonw I guess you got no home address, phone or whatever...
09:04 carsten lets hope that is true
09:04 carsten but on august 20th the students are supposed to upload their work
09:04 * carsten is afk for 10 min
09:05 egonw carsten: oh... really...?
09:05 * egonw feels it is time to check the GSoC time schedule :)
09:32 carsten back
09:32 carsten egonw: or not?
09:32 * carsten and Julia are backing a Marmor- and an apple pie
09:34 egonw in your new kitchen?
09:34 * carsten doewsn't have a new kitchen
09:34 carsten we'll move ~25 oct
09:34 carsten still in my old appartement
09:35 egonw ah...
09:35 * egonw is baking an article
09:54 jbrefort osra is a nightmare to build: ill-formed dependencies like tclap :(
09:58 carsten jbrefort: but it sounds really cool
09:58 jbrefort of course
09:58 carsten Templatized C++ Command Line Parser Library
09:58 carsten that sounds like a superflous thing....
09:59 jbrefort the issue with tclap is that the .h files have #include "config.h"
09:59 jbrefort which is quite stupid
10:02 * carsten things all coders should at least use GCC 4.2 with --strict and without gcc extensions
10:02 carsten and run Krazy from time to time
10:15 egonw carsten: yeah, that would be a good blog item: run Krazy on the OSRA source code
10:15 egonw out dump the output
10:15 egonw out -> and
10:18 neksa Krazy on OSRA is crazy! you've been warned.
10:18 neksa jbrefort: if you need help in building OSRA ping me, I passed this nightmare and got a binary
10:19 egonw not a .deb I guess? :)
10:19 jbrefort hings seem to compile now, but gocr is a strange thing too.
10:20 neksa egonw:  I can't build a decent OpenBabel  deb :-/
10:20 jbrefort that's strange
10:21 carsten neksa: I created a rpm with the buildservice
10:21 carsten why don't you use that?
10:21 neksa carsten: it is alien!
10:22 egonw right... apt-get install alien
10:22 carsten neksa: no, you can use the buildservice to create a native deb
10:23 jbrefort osra built, but crashes
10:23 neksa carsten: I know. The current .deb from debichem is not that bad. It has some minor problems. I like the way it splits the packages
10:24 carsten neksa: my openbabel has two rpms: one for debug, one for openbabel itself
10:25 neksa carsten: debichem rules create libinchi-dev, libinchi0, libopenbabel-dev, libopenbabel2, openbabel and python-openbabel packages
10:27 neksa the problem is when I build it on my Debian unstable, the dependencies are too strict to use the resulting .deb package on Ubuntu, for example.
10:27 jbrefort which dependencies?
10:30 neksa jbrefort: libopenbabel2_2.1.0 Depends: libc6 (>= 2.5-5), libgcc1 (>= 1:4.2-20070516), libinchi0, libstdc++6 (>= 4.2-20070516), libxml2 (>= 2.6.28), zlib1g (>= 1:1.2.1)
10:30 jbrefort osra always crashes there, and gdb is unable to provide a trace :(
10:30 jbrefort neksa, yes, these are quite common libs, one is missing in ubuntu?
10:30 neksa jbrefort: I just need to teach myself   ".pc"-magic
10:31 neksa no, the versions are problematic
10:31 jbrefort the pc magic is quite easy ;)
10:32 jbrefort pkg-config --cflags --libs libxml2
10:32 neksa e.g. to turn libgcc1 (>= 1:4.2-20070516)  into libgcc1 ( >= 4.2) or even into  libgcc1
10:33 neksa jbrefort: thanks, I'll try! haven't spent much time on that..
10:33 jbrefort hmm, works better with: pkg-config --cflags --libs xml2
10:34 * jbrefort is a pkgconfig expert (or almost)
10:38 neksa $ pkg-config --cflags --libs libxml-2.0
10:38 neksa -I/usr/include/libxml2  -lxml2
10:44 rojasm joined #cdk
10:53 egonw hi rojasm
10:53 rojasm hi Egon
11:05 carsten egonw: what is greasemonkey?
11:05 carsten there is no readme file it
11:05 carsten there is no readme file in it
11:06 egonw that's the paper I am writing up today
11:07 egonw see the BO wiki
11:07 carsten well, greasemonkey is a javascript thing
11:07 CIA-1 blueobelisk: egonw * r81 / (4 files in 3 dirs): Renamed root directory: other browsers support userscripts too, no longer restricted to FireFox/Greasemonkey
11:07 egonw yes, they are userscripts to change the HTML content in web pages... on the fly, on client side
11:08 egonw e.g. to remove pop-up advertisement
11:08 egonw http://blueobelisk.sourceforge.net/wiki/index.php​/Using_Javascript_and_Greasemonkey_for_Chemistry
11:08 lambdabot2 Title: Using Javascript and Greasemonkey for Chemistry - BlueObelisk, http://tinyurl.com/37ahq7
11:08 egonw lambdabot2: thanx
11:10 carsten ok, so greasemonkey is a script system not related to chemistry but some folks wrote scripts for it to to chemical things
11:11 egonw yes, and those folks include your truly
11:12 egonw +s x2
11:13 egonw jbrefort: please send that reply to Igor too...
11:13 egonw maybe even with suggestions on how to improve things...
11:14 egonw e.g. use a more accepted CLI library, or whatever...
11:19 * carsten also replied to that mail
11:21 egonw to the list?
11:21 egonw did not see your reply yet...
11:26 egonw ok, arrived now
11:49 jbrefort egonw, done
11:50 egonw ack, thanx
11:53 jbrefort carsten, I know your position about the autotools/cmake issue ;)
11:53 carsten hehe :
11:53 carsten )
11:54 carsten jbrefort: CMake is simply simple and elegant
11:54 carsten and both cannot be said about autohell
11:54 egonw anyone here who know a nice first person shooter/space game for (k)ubuntu?
11:54 carsten and I still don't know why OpenBabel includes generated files
11:54 egonw I feel the need to loose some aggressiveness
11:54 jbrefort hmm, but it was rejected by the gnome community because of needed features lack
11:54 carsten egonw: counterstrike and Quake 1,2,3
11:54 egonw apt-get-able?
11:55 egonw arg... "Firefox can't establish a connection to the server at packages.ubuntu.com."
11:55 jbrefort carsten, I agree wit you about the configure, Makefile.in and other generated stuf in OB, they should not be in svn
11:55 carsten jbrefort: which is a very very stupid argument. KitWare would add neeeded features. b) cmake is extendable with macros so those features can be added and c) CMake is able to build KDE so I seriously doubt there are a lot missing features
11:56 carsten jbrefort: my suggestion to remove them was shot dead
11:56 jbrefort egonw, I have the same issue with packages.gentoo.org, seems ff does not like packages
11:56 egonw nah. Konqueror complaints too...
11:56 jbrefort ;)
11:57 egonw oh no... not another flame war :)
11:57 egonw please don't bring in the nazis
11:58 egonw (sorry, was just reading comments on this Linux Foundation guy's statement to respect MS ...)
11:58 egonw yeah, we're really actually very bad open source developers...
11:58 egonw we never flame war over OB vs CDK or so...
11:58 egonw bad us
11:58 carsten I am not flaming, I just say that even if that is true and could be changed. KitWare is releaseing a CMake version every few weeks, 2.5.7 is out already with some new features. the upcomming 2.4.8 will include several macros KDE created so that other projects can use them without any hacks
12:00 egonw carsten: what package is counterstrike?
12:00 carsten egonw: you need the windows version
12:00 carsten egonw: I don't know any good 1st person shooter for free
12:01 carsten egonw: there is a wolfenstein3d clone for free, but I don't know the name
12:01 carsten I think Quake is free and everything
12:01 carsten (GPL)
12:01 carsten but I don't know if there is a .deb
12:01 carsten Quake is a very good 1st person shooter (was commercial, the company simple gpl'ed old version, the latest is always commercial)
12:28 egonw quake2 does not run on my machine...
12:28 egonw Error: Couldn't load pics/colormap.pcx
12:28 egonw and trouble detecting sound...
12:43 neksa egonw: http://sauerbraten.org/
12:43 lambdabot2 Title: Sauerbraten / Cube 2
12:45 egonw that last screenshot is beautiful! of that cave
12:51 egonw will try that...
12:52 egonw blah... gutsy has debs...
13:07 carsten neksa: how stupid is that name!?
13:08 neksa carsten: don't you like beef?
13:09 carsten I do, but I'd never call an engine like that
13:12 carsten egonw: in the wikipedia there are other shooters mentioned in the sauerbraten article
13:28 egonw two cool bug reports from Andreas in Frankfurt
13:28 egonw with JUnit tests, and fixes :)
14:14 rojasm left #cdk
14:47 jbrefort joined #cdk
17:46 carsten joined #cdk
18:29 shk3 joined #cdk
19:32 mashuuk joined #cdk
19:32 mashuuk Hi
19:32 mashuuk Anybody jere can explain me how to implement affine gap penalty ?
21:10 neksa mashuuk: gap of length l is then given a penalty:  opening_penalty + (l-1)extenson_penalty
21:10 mashuuk neksa, ok I found some implementation example
21:11 mashuuk I still dont understand why three arrays are used
21:11 mashuuk neksa, one its abvious as its for collecting a nodes scores
21:12 mashuuk and the way its initialized its fine..
21:13 * neksa looking for an implementation he did a year ago in python
21:13 mashuuk also biojava is buggy
21:13 mashuuk as it gives diffrent score than water from emboss
21:15 neksa my implementation is plain, no biopython, just pure algorithm. email?
21:19 neksa mashuuk: i'm going offline. ping me if you need my implementation

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