Camelia, the Perl 6 bug

IRC log for #cdk, 2007-11-30

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

All times shown according to UTC.

Time Nick Message
00:04 CIA-16 jmol: hansonr * r8695 /trunk/Jmol/src/org/jmol/ (3 files in 2 dirs):
00:04 CIA-16 jmol: version=11.3.50
00:04 CIA-16 jmol: # new feature: MOL isotope column read; interpreted for D, T, 11C, 13C, 15N (for now)
00:04 CIA-16 jmol: hansonr * r8696 /trunk/Jmol/src/org/jmol/adapter​/readers/molxyz/MolReader.java: version=11.3.50 # new feature: MOL isotope column read; interpreted for D, T, 11C, 13C, 15N (for now)
00:13 CIA-16 jmol: hansonr * r8697 /trunk/Jmol/src/org/jmol/adapter​/readers/molxyz/MolReader.java: version=11.3.50 # new feature: MOL isotope column read; interpreted for D, T, 11C, 13C, 15N (for now)
05:52 CIA-16 jmol: nicove * r8698 /trunk/Jmol/src/org/jmol/viewer/Jmol.properties: Version
06:35 jbrefort joined #cdk
06:51 rojas1 joined #cdk
08:19 steinbeck joined #cdk
09:03 egonw moin
09:33 jonalv joined #cdk
09:35 masak joined #cdk
09:51 egonw jonalv: you're the Spring dude here...
09:51 egonw can you say anything about using Spring for building websites...
09:51 egonw as compared to servlets or JSPs ?
09:52 * masak imagines it would help SoC
09:53 * jonalv has not done any websites but feels that Spring is very good in any project what grows beyond some point of complexity
09:53 egonw why?
09:53 egonw looked at the website...
09:53 egonw but ...
09:54 egonw puzzled at what problems Springs addresses
09:54 masak egonw: if it's any consolation, almost everyone is at first
09:54 masak Spring is very general
09:54 jonalv It took me a week and 150 pages of a Spring book to realise what it was all good for
09:54 egonw give me the executive summary :)
09:55 masak yes, let's hear the elevator pitch :)
09:55 jonalv Everyone just said "use Spring it's great!" and I was like: "Yea sure but for what?"
09:56 egonw yeah, so stop doing the same thing! :)
09:56 jonalv well the big benefits as I see it can be summoned up in the following things: IoC, AOP, great integration with ORMs
09:56 masak maybe you cannot be told what Spring is. you have to see for yourself
09:57 jonalv but I have also heard that it's got great integration with Struts and other web frameworks aswell
09:57 egonw worms ?
09:57 jonalv masak, Yes it's a little bit like that
09:57 masak egonw: ORM = Object-Relational Mapper
09:57 egonw ok, what are: IoC, AOP and ORMS?
09:57 jonalv egonw, yea like Hibernate, iBatis,
09:57 masak egonw: IoC, inversion of control
09:57 jonalv IoC = inversion of control
09:57 masak egonw: AOP, aspect-oriented programming
09:57 jonalv AOP = aspect oriented programming
09:57 masak :)
09:58 egonw mmm... more buzz words
09:58 masak yes, Spring brings all the buzz to Java
09:58 egonw so, what problems do these solve: IoC, AOP, ORMS
09:58 masak in a word, complexity
09:58 masak but from different directions
09:58 masak maybe I should let jonalv++ explain
09:59 jonalv It's not easy to just explain but I will give it a try
09:59 jonalv Really the best would be to read the first 100 pages of a Spring book...
10:00 jonalv IoC in the Spring flavour is a way to instantiate your service object in a container
10:00 jonalv How this different objects are put toghether are defined in XML
10:01 jonalv So it takes the complexity of the structure of your programs and let's you deal with that in XML (which is a great way to deal with structure)
10:01 jonalv AOP is a way to eject dependencys horisontally out of the class hierarchy
10:02 jonalv For example when you are working with a databse you first need to start a transaction and the either rollback or commit
10:02 egonw eject dependencys horisontally ???
10:02 jonalv You don't want to write this code everywhere in your methods but at one place
10:03 masak ...but it doesn't fit in the normal class hierarchy
10:03 jonalv So you eject the transaction code as an aspect and let Spring insert it where it is needed
10:04 jonalv Spring does this by wrapping your code in a proxy and catching the call to the method that needs to perform transactionally. It will run the code starting the transaction and listen for exceptions
10:05 jonalv If an esception is thrown it will rollback the transaction and if everything works it will commit it
10:06 jonalv The good thing with this is that when writing the method that needs to perform in a transaction you can just ignore the transactional stuff because Spring will take care of it
10:07 jonalv I am sure that we will want to do stuff in our service objects that can benefit from this way of seperation of convernes
10:07 masak from a birds-eye perspective, OOP is good at inheritance, AOP is good at Separation of Concerns
10:07 masak oh oh, OOP is good at encapsulation too
10:07 jonalv Yes and it is also important to realise that AOP is not a replacement for anything but a complement
10:08 masak AOP sort of breaks encapsulation in a sane, ordered way to enable even more SoC
10:09 masak the wikipedia article talks about code that's scattered or tangled
10:09 masak it's scattered when one thing (like logging) is all over the code
10:09 masak it's tangled when several aspects (logging, processing, security) are in the same module
10:09 masak AOP fixes both of these issues
10:10 masak ...and Spring enables AOP
10:10 egonw OK, /me thinks he has to buy the book :(
10:10 jonalv hehe
10:10 masak :)
10:11 masak jonalv: maybe our explanations were too scattered and tangled
10:11 egonw but, guys, nice try :)
10:11 jonalv yea maybee... ;(
10:11 jonalv I will try that again:
10:12 jonalv yea maybe... ;)
10:12 masak :)
11:14 egonw rojasm_: around?
11:36 egonw bbl
12:08 steinbeck joined #cdk
12:12 masak joined #cdk
12:16 rojasm_ (01:14:34 PM) rojasm_: Hi egonw: thanks for the email. Thanks that your are having a look.
12:16 rojasm_ (01:15:10 PM) rojasm_: Other proposal. For the restrictions of MolecularFormula I want to constract some like QSAR descriptors
12:16 rojasm_ (01:15:33 PM) rojasm_: so it is possible to implemente different rules to restrict valid MolecularFormula
12:16 rojasm_ (01:15:40 PM) rojasm_: what do you think
12:30 CIA-16 cdk: miguelrojasch * r9533 /branches/miguelrojasch/mf​/src/org/openscience/cdk/ (4 files in 2 dirs): New methods in MolecularFormulaManipulator to handle IMolecularFormulas
12:45 cpyang joined #cdk
12:45 egonw joined #cdk
12:45 steinbeck joined #cdk
12:46 jpansanel moin
12:47 CIA-16 blueobelisk: pansanel * r214 /structures/trunk/src/ (9 files in 5 dirs): Fix some spelling / file name
12:48 CIA-16 blueobelisk: pansanel * r215 /structures/trunk/tools/ (5 files):
12:48 CIA-16 blueobelisk: Link name has been added to formula index
12:48 CIA-16 blueobelisk: Some documentation has been added to the handler modules
12:49 CIA-16 blueobelisk: pansanel * r216 /structures/trunk/TODO: Update TODO with last modification and ideas
12:55 CIA-16 blueobelisk: pansanel * r217 /structures/trunk/ChangeLog: ChangeLog contains the last modification
12:57 CIA-16 bioclipse: biocoder * r4367 /trunk/plugins/net.bioclipse.statistics/src/​net/bioclipse/editors/MatrixGridEditor.java:
12:57 CIA-16 bioclipse: Added a selectionChangedListener to listen for clicks in the plot
12:57 CIA-16 bioclipse: Fixed a bug that caused the wrong data to be sent to the charting plugin
13:12 carsten joined #cdk
13:12 cpyang joined #cdk
13:13 jpansane1 joined #cdk
13:15 steinbeck joined #cdk
13:22 CIA-16 jmol: hansonr * r8699 /trunk/Jmol/src/org/jmol/ (7 files in 6 dirs): version=11.3.51_dev # code: minor refactoring in org/jmol/adapter
13:30 CIA-16 jmol: hansonr * r8700 /trunk/Jmol/src/org/jmol/ (6 files in 4 dirs): version=11.3.51_dev # code: minor refactoring in org/jmol/util
13:55 dleidert joined #cdk
14:15 CIA-16 jmol: hansonr * r8701 /trunk/Jmol/src/org/jmol/ (9 files in 4 dirs): version=11.3.51_dev # code: minor refactoring in org/jmol/adapter
14:25 CIA-16 jmol: hansonr * r8702 /trunk/Jmol/src/org/jmol/jvxl/ (calc/MarchingSquares.java readers/VoxelReader.java): version=11.3.51_dev # code: minor refactoring in org/jmol/jvxl
15:29 CIA-16 jmol: rkanters * r8703 /trunk/Jmol/src/org/jmol/adapter​/readers/molxyz/XyzReader.java:
15:29 CIA-16 jmol: Changed the setting of the titles so that the comment line is always the title.
15:29 CIA-16 jmol: The previous implementation got this somewhat messed up for a multi-structure xyz file.
15:29 CIA-16 jmol: I'll upload a 4struct.xyz file to the Jmol-datafiles so you can see what I mean (if you want to).
15:32 dleidert joined #cdk
15:34 CIA-16 jmol: rkanters * r8704 /trunk/Jmol-datafiles/xyz/4structs.xyz: Sample file xyz with three structures in it, to show that the older version messed up the titles
15:43 CIA-16 jmol: rkanters * r8705 /trunk/Jmol/src/org/jmol/adapter/​readers/more/GaussianReader.java:
15:43 CIA-16 jmol: I thought I had just committed this...
15:43 CIA-16 jmol: See comments in the code.
15:50 CIA-16 jmol: rkanters * r8706 /trunk/Jmol-datafiles/xyz/ (3structs.xyz 4structs.xyz): I can almost count. The file had three structures in it, so I renamed it...
16:02 masak joined #cdk
17:40 CIA-16 jmol: hansonr * r8707 /trunk/Jmol/src/org/jmol/ (6 files in 3 dirs):
17:40 CIA-16 jmol: version=11.3.51_dev
17:40 CIA-16 jmol: # bug fix: load =xxxx broken in 11.3.50
17:40 CIA-16 jmol: # bug fix: amino not a subset of protein
17:40 CIA-16 jmol: # bug fix: proteins with just C CA N not recognized as such
17:43 CIA-16 jmol: hansonr * r8708 /trunk/Jmol/src/org/jmol/viewer/Jmol.properties:
18:17 egonw_ joined #cdk
18:17 egonw_ moin
18:21 CIA-16 cdk: miguelrojasch * r9534 /branches/miguelrojasch/mf​/src/org/openscience/cdk/ (15 files in 5 dirs): New package implementation for MolecularFormula. It gives sense to validate Molecular formulas. Still in progress and in corrections.
18:22 rojasm_ Hi egonw
18:22 rojasm_ I go now at home. We speak later if you will be there still
18:23 rojasm_ see you. May be you take a look and make some commentars in chat
18:24 rojasm_ a10
18:24 rojasm_ left #cdk
19:18 dleidert_away joined #cdk
19:29 egonw_ hi dleidert_away
19:52 rojasm joined #cdk
19:54 egonw_ hi rojasm
19:54 rojasm hi egonw
19:55 * dleidert_away says hi to egonw_ and is leaving ;)
20:18 egonw__ joined #cdk
20:43 CIA-16 jmol: hansonr * r8709 /trunk/Jmol/src/org/jmol/ (3 files in 3 dirs): version=11.3.51_dev # bug fix: PDB CONECT records not creating bonds for ALL models
21:01 CIA-16 jmol: hansonr * r8710 /trunk/Jmol/src/org/jmol/adapter/ (readers/cifpdb/PdbReader.java smarter/AtomSetCollection.java): version=11.3.51_dev # bug fix: PDB CONECT records not creating bonds for ALL models
21:19 egonw_ rojasm: shall we still chat?
21:19 egonw_ I got some 30 minutes left
21:19 rojasm perfect
21:20 rojasm I wanted to ask you what you think about to make some think similar as QSAR descriptors for Rules to validate the MolecularFormula
21:21 rojasm So could be possible to implement easier more restrictions
21:21 egonw_ I think it is conceptually different
21:21 rojasm ok
21:21 egonw_ true
21:21 egonw_ you could still use a plugin-like system...
21:21 egonw_ using a common interface...
21:21 egonw_ the REaderFactory uses taht too
21:21 egonw_ so, the idea is good...
21:21 egonw_ but would not use the IDescriptor interface
21:22 rojasm I don't say to use the IDescriptor interface but some think similar in his structure
21:23 rojasm But you say more similar to ReaderFactory then
21:23 egonw_ well...
21:23 egonw_ what is the functionality...
21:23 egonw_ what does it need to do
21:23 egonw_ that defines the interface
21:23 egonw_ so, something like:
21:24 egonw_ IMFRule.isValid()
21:24 rojasm did you check the MMElementRule class
21:24 egonw_ IMolecularFormulaValidator.ad​dRule(IMolecularFormulaRule)
21:24 egonw_ only briefly...
21:24 egonw_ hang on...
21:24 egonw_ will check now
21:25 rojasm I thought that the good thing about the structure of the  IDescriptor is that is good implemented and linked with name, results and descriptions
21:26 rojasm And I thought may be It should be also so implemented
21:26 egonw_ not sure if that is really need in this case...
21:26 egonw_ there are not that many rules
21:27 egonw_ and unlike QSAR descriptors... there will not be many instances outside the CDK
21:27 rojasm yes only one
21:27 rojasm because I wanted to know you opinion first
21:27 rojasm ups
21:27 rojasm yes
21:27 rojasm that is true
21:27 rojasm I think maximal 10
21:28 rojasm or as maximum 20
21:28 egonw_ there are 38 citations for the first CDK article now :)
21:28 rojasm that ist very good
21:35 rojasm So you mean not to do too complicate. And simplifacate the package
21:36 egonw_ yes, always as simple as possible
21:36 egonw_ the more complicated, the more easy it is to break it
21:37 rojasm ok
21:52 egonw_ about IRule...
21:52 egonw_ the rules don't have params, do they?
21:52 egonw_ do you need getResultType()?
21:52 egonw_ isn't it just true/false?
21:53 egonw_ and leaving out the getSpec() gives you a nice and clean interface
22:02 rojasm not exactly
22:02 rojasm because some times you can obtain a score function
22:02 rojasm comparing isotope pattern
22:03 egonw_ but comparing spectra is something else...
22:03 CIA-16 cdk: djiao * r9535 /trunk/cdk/src/org/openscience/cdk/isomorphi​sm/matchers/smarts/LogicalOperatorBond.java: Fixed a bug reported on Operator Bonds
22:03 rojasm well ist a restriction for validate the MolecularFormula that you are interested
22:04 rojasm I thought you can have differents nivel to validate the MolecularFormula according your necessities
22:04 egonw_ yes, I see where you're going...
22:12 egonw_ simply return a float [0,1] then ?
22:17 rojasm I thought to return a double in this case a DoubleResult
22:20 egonw_ yeah, double is fine instead of float
22:20 egonw_ DoubleResult is an overkill, I think
22:20 rojasm you mean that already exists
22:20 rojasm yes
22:20 egonw_ no, that it is not needed
22:21 rojasm achso
22:22 rojasm But then how you can return for differents rules a boolean and a double(or float) and pack with the same thing
22:22 egonw_ no, only double
22:22 egonw_ for boolean-like rules... return either 0.0 or 1.0
22:23 rojasm ok
22:24 rojasm makes easier
22:24 rojasm :)
22:27 CIA-16 jmol: hansonr * r8711 /trunk/Jmol/src/org/ (11 files in 5 dirs): version=11.3.51_dev # bug fix: write FILE not handling binary file formats properly
22:44 rojasm I am removing all methods which are not necessaries and only leaving setParameters() getParameters() and isValid()
22:46 egonw_ why get/setParam() ?
22:47 egonw_ for spectra comparing etc, I guess?
22:51 CIA-16 cdk: miguelrojasch * r9536 /branches/miguelrojasch/mf​/src/org/openscience/cdk/ (6 files in 4 dirs): removed unnecessary methods
22:54 rojasm For example to specify what elements you want to analyze and what maximal and minimal occurrence they should have
22:55 rojasm I thought setParameters is a godd way
22:56 CIA-16 cdk: miguelrojasch * r9537 /branches/miguelrojasch/mf​/src/org/openscience/cdk/ (2 files in 2 dirs): added new rule to validate MolecularFormula. Occurrence of the elements.
22:56 rojasm What do you think. We should change it?
22:56 egonw_ not sure...
22:56 egonw_ can't think anymore...
22:57 rojasm ups
22:58 rojasm I hope I don't give you too much work
23:01 CIA-16 jmol: hansonr * r8712 /trunk/Jmol/src/org/jmol/viewer/Jmol.properties: 11.3.51 VERSION
23:31 egonw_ rojasm: cu later
23:52 CIA-16 jmol: nicove * r8713 /trunk/Jmol/examples/basic/Export.java: Warning

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