Camelia, the Perl 6 bug

IRC log for #bioclipse, 2007-11-22

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

All times shown according to UTC.

Time Nick Message
04:36 ilogger2 joined #bioclipse
08:45 olas joined #bioclipse
08:47 jonalv joined #bioclipse
09:05 olas egonw: around?
09:05 olas http://www.ncbi.nlm.nih.gov/sites/entr​ez?cmd=Retrieve&db=PubMed&list​_uids=17962754&dopt=AbstractPlus
10:14 miguelrojasc1 joined #bioclipse
10:34 ilogger2 joined #bioclipse
10:38 jonalv btw olas: I checked out jfeature a bit more. I liked thier idea but the program is completely broken. It is not working with Eclipse 3.3
10:39 jonalv miguelrojasc1: did you ask about hasmaps in #CDK a while ago?
10:40 miguelrojasc1 yes! But I find other way for no to use it
10:40 miguelrojasc1 but it is possible?
10:41 jonalv not with a hasmap no but there is something called linkedhashmap which stores the order of the items
10:41 miguelrojasc1 ok thanks I will have a look
10:41 jonalv so you could use a linkedhashmap for it
10:45 miguelrojasc1 Thank jonalv
10:48 jonalv np
10:50 jonalv soon lunch, olas did you heat up the sauna? ;-)
11:13 egonw yohoo!!!! the manuscript committee approved my thesis! I'm done !
12:48 steinbeck joined #bioclipse
12:48 steinbeck cheers
13:08 thomas_ku joined #bioclipse
13:21 egonw hi steinbeck
13:22 steinbeck Hi egonw
13:22 steinbeck got my reply?
13:22 egonw yes
13:22 egonw so either you or Ola will go...
13:22 egonw there is funding for one
13:22 steinbeck ack
13:23 steinbeck Any idea when Ola will decide?
13:23 egonw hopefully when he gets back from lunch/sauna
13:23 egonw :)
13:23 steinbeck hehe
13:23 steinbeck I'll open a private chat
13:24 egonw steinbeck: you've met steffen neumann (sneumann) in person, not?
13:24 steinbeck yes
13:26 olas left #bioclipse
13:26 egonw argh...
13:26 egonw Uppsala... any there to bring Ola back online?
13:27 egonw any*one*
13:28 EskilA joined #bioclipse
13:35 jonalv joined #bioclipse
13:35 egonw jonalv, EskilA: is ola back from lunch?
13:35 EskilA yeah
13:35 egonw can you ask him to go online, if you happen to be near his room
13:36 EskilA I can give him a call
13:36 EskilA he seems to be on the phone
13:36 egonw OK,
13:36 egonw I see he's online with skype now...
13:36 egonw thanxx
13:36 EskilA :)
13:41 egonw joined #bioclipse
13:44 EskilA Is it possible to force an AWT frame to redraw?
13:44 egonw repaint() I think
13:45 EskilA I've tried, however the frame doesn't draw unless i resize the window containing the frame
13:55 olas joined #bioclipse
13:56 olas steinbeck: around?
13:56 steinbeck Hi
13:57 jonalv reagarding Spring OSGI <http://www.infoq.com/interviews/osgi-adrian-colyer> very interesting. (30 minutes)
14:20 masak joined #bioclipse
14:29 edrin joined #bioclipse
14:29 edrin left #bioclipse
15:14 jonalv masak: <http://blog.interface21.com/main/2007/04​/05/so-whats-the-deal-with-spring-osgi/>
15:16 masak jonalv: thx
15:27 masak http://bioclipse.blogspot.com/2007​/11/scripting-in-bioclipse-2.html
15:41 egonw olas: while you're still here?
15:42 egonw how do I get a .pdb file in my BC2 environment?
15:44 masak egonw: I'm no expert, but let me see if I understand your question
15:44 masak you want to open a .pdb file, like we do today in Bioclipse 1's BioresourceView?
15:44 egonw indeed
15:44 masak I imagine it will eventually be done the same way
15:45 masak i.e. double click
15:45 masak or "Open" in context menu
15:45 masak only BioResourceView is now Navigator
15:45 CIA-18 bioclipse: biocoder * r4320 /trunk/plugins/net.bioclips​e.chart/src/net/bioclipse/ (chart/ChartUtils.java plugins/views/ChartView.java):
15:45 CIA-18 bioclipse: Reworking chart plugin to integrate Swing components properly
15:45 CIA-18 bioclipse: Currently only Scatter plot has been integrated.
15:46 egonw masak: sure...
15:46 egonw but the problem really is:
15:46 egonw how do I get something shown in the new view...
15:46 egonw it's empty!
15:46 masak :)
15:46 masak the operative word here is "eventually", methinks
15:47 masak bc2 is more of a shell today, with not much ghost in it
15:50 olas egonw: Im back
15:50 olas did you solve it?
15:50 egonw no
15:50 olas in your hoem dir, there's a runtime-bioclispe dir
15:50 egonw no idea what you're talking about...
15:51 egonw but have only been reading...
15:51 egonw oh...
15:51 egonw mom, let me check
15:51 olas you have BC2 and would like to have a PDB in it, right?
15:51 olas make a project
15:51 jonalv egonw: regarding your comment to the scripting blog post. I don't really think I understand what you mean...
15:51 egonw what's the beasty called?
15:51 jonalv Could you elaborate?
15:51 olas then find runtime-bioclipse
15:51 egonw got no $HOME/runtime-bioclipse
15:51 olas hmm
15:51 olas look in eclipse
15:52 olas run dialog
15:52 olas there it says where your runtime-workspace is
15:52 egonw jonalv: what EP do I use to have my custom plugin define a new data object to beused in the scripting langauges...
15:55 jonalv egonw: That is one of those things that I would like to manage by Spring but righl now there is an EP which I don't know the name of right now...
15:55 masak egonw: it's called INamespaceProvider in bc2
15:55 egonw OK, might you blog a bit about that EP?
15:55 masak egonw: sure
15:56 masak but as jonalv says, it's likely to change soon
15:56 masak maybe.
15:56 egonw ok, then wait :)
15:56 jonalv perhaps
15:57 masak it's a very interesting to be living in a changing data model, but it feels a bit strange too
15:59 masak egonw: FWIW, olas++ already blogged about how he set up the EP
15:59 masak egonw: http://bioclipse.blogspot.com/2007/11/add​ing-scripting-commands-to-bioclipse.html
16:00 masak (note that INamespaceProvider was called IScriptProvider at this time, but everything else is the same)
16:01 egonw ah... but that's really on method level
16:01 masak yes...
16:01 jonalv egonw: I still can't see what you think is a problem wiith all this...
16:01 egonw so, does that work for data types too?
16:01 egonw did not mean the idea is not correct...
16:01 masak egonw: hm, no
16:01 egonw just that I don't understand how I can define a BibTex object to be used in the scripting
16:02 masak egonw: I imagine many of the central data types will be interfaces defined by bc2
16:02 egonw that would be a problem...
16:02 egonw you should not hard code that...
16:02 masak egonw: I would be very happy to discuss such interfaces with you at some point
16:02 masak egonw: hard code?
16:02 masak we figure that interfaces will be minimal
16:03 * jonalv feels that maybe we should use emf for these domain objects...
16:03 * egonw forsees many maintainance hours for the Uppsala team
16:03 masak :)
16:03 egonw and people screeeming... he's where's that method to do .... <BLA> ... ?!?!?
16:03 jonalv with emf you can even create new "classes" dynamicaly
16:03 egonw he's -> He!
16:04 masak egonw: so you think that minimal interfaces only will make people unhappy?
16:04 olas egonw: we can generate an EMF model based on an OWL
16:04 masak we can? cool!
16:04 olas so every time the OWL changes, we can just regenerate the EMF model
16:05 olas well, in theory :-)
16:05 masak olas: and all old scripts break :-P
16:06 olas I know
16:07 olas http://www.eclipse.org/mode​ling/mdt/?project=eodm#eodm
16:08 olas EODM allows users to load a RDF/OWL ontology into EMF-based Java objects, to manipulate and infer the EMF Java objects which represents the RDF/OWL ontology in memory, and to serialize the Java objects to an ontology in RDF/XML syntax. EODM also provides model transformation capabilities between RDF/OWL and other modeling languages, such as ECore, UML and ER etc.
16:08 olas sounds really cool
16:09 masak ...maybe bioclipse 3 material? please? ;)
16:09 olas maybe
16:10 olas you could be right
16:10 jonalv I'll give it 9.5 out of a maximum of 10 on the many-three-letter-names-in-one-sinlge-text scale
16:10 olas we should start with something simpler
16:10 masak I vote for a process where everything gets accepted that people can get to run with what we already have
16:10 masak :)
16:11 * jonalv asks himself: "is it beginning to be late in the afternoon?"
16:11 masak it's not the end, but the beginning of the end
16:13 jonalv seriously though: I think emf might be the way to go with these domain objects. My feeling is though that maybe we should stick with the easiest possible solution using emf and the fancy autogenerating birds-awake-at-night auto updating solution...
16:14 jonalv ...could be left for bioclipse 3... :)
16:14 masak *phew*
16:14 masak was worried there for a while
16:14 * masak is uneasy with any software solution involving birds awake at night
16:16 jonalv koay so to sum things up the ideal solution in my world would be a domain model made up with emf objects and a service layer constructed using Spring but hey that's just what I would prefer right now...
16:17 masak I'm with you on Spring, and theoretically on EMF too
16:18 jonalv do I hear a but?
16:18 masak though I can't claim to grok EMF, or even the need for it
16:18 masak to be honest, I hate it
16:18 masak but that's just me
16:18 masak maybe we would indeed be wise to use it
16:18 jonalv could you elaborate on why you hate it?
16:18 masak no, it's very irrational
16:19 masak it looks so... enterprisey
16:19 masak a way of writing "Hello World" using only 5 components
16:19 jonalv and that is a bad thing?
16:19 * masak cries
16:20 jonalv I mean is enterprisey really a bad thing?
16:20 masak technically, it's just a bad name :)
16:20 jonalv what emf?
16:20 masak no, 'enterprisey'
16:20 jonalv it was you who used it first...
16:21 masak it means doing easy things the rigid, complicated way
16:30 olas left #bioclipse
16:44 masak @tell edrin I just understood what your dynamic classpath loader does :)
16:47 edrin joined #bioclipse
16:47 edrin hi
16:54 masak I just understood what your dynamic classpath loader does :)
16:56 edrin me?
16:56 jonalv yes he meant you
16:57 jonalv :)
16:57 edrin ah
16:57 jonalv so it was just one line of code in the manifest or what?
16:57 edrin i think so
16:57 edrin or what does he mean?
16:57 jonalv sounds cool
16:58 jonalv So it actually let's you use classes in all loaded plugins?
16:58 edrin was there a longer discussion regading this subject?
16:58 jonalv Yes but in meat space here in Uppsala... :)
16:59 jonalv What happens if there is a verioning confligt?
16:59 edrin it enables access to all classes/objects in the runtime/RCP that are at least exported by other plugins...
16:59 jonalv ah yes of course they have to be exported otherwise it would be to magical... :)
17:00 edrin jonalv: there is no versioning conflict. well, the reason for this line is just that it is very cool to have this for a very dynamic scripting engine ...
17:00 jonalv but what happens if there is a name collision or many versions of something
17:01 * jonalv is thinking about using it for a somewhat less fancy Spring solution not involving the spring osgi framework
17:01 edrin jonalv: that's a good point. indeed this can be a problem. therefore it is highly recommended to use very strict bioclipse or eclipse naming conventions
17:01 jonalv I am not sure that it will work but it should...
17:02 jonalv SO I am thinking about trying it
17:02 edrin but you might get into trouble if two plugins come with the same .jar of a lib (that might have different versions)
17:02 jonalv It will totally break (or rather ignore) the plugin structure however...
17:02 edrin in example one plugin has axis1.2 and another has axis1.4
17:02 jonalv what happens?
17:03 edrin in this case the dynamic classpath loader will be not predictable
17:03 edrin that means it can return either the one or the other but nobody knows
17:03 jonalv but it doesn't crach and burn anyway. That's good... :)
17:04 edrin yes, well, this wont happen
17:04 edrin but indeed this could be a problem
17:06 jonalv yes it could be a problem and it is something of a hack but if we only use it for strictly named bioclipse classes I hope we should be okay...
17:06 jonalv or?
17:06 edrin yes
17:07 jonalv that's what my Spring stuff sould do anyway...
17:07 jonalv But I can't shake the feeling that this might not be the best way to do it...
17:07 edrin indeed you need this only if you want to define/do stuff at runtime.
17:07 edrin if your plugin's actions are all well defined you wont need this.
17:07 edrin i can only recommend this if you are wrting a scripting plugin (or something comparable)
17:08 edrin what should your Spring stuff do?
17:08 jonalv It should instantiate objects defined in other plugins and wrap them in proxies for aop stuff...
17:09 jonalv so it needs an xml file from another plugin and the classes in order to instantiate them...
17:10 jonalv the xml file shouldn't be anu problem that will simple be a file path right? but the classes need to be in the plugins class path I guess...
17:11 edrin i dont know. i still did not understand what "spring" is supposed to be
17:11 jonalv This was my first idea about how to use Spring in bc2 but then I found out about spring-osgi which seems alot more complex and more tightly integreated with the plugin structure
17:11 jonalv What I want out of it is the aop feature
17:11 edrin is it something like "MFC for java"?
17:11 jonalv Do you know about aop?
17:11 jonalv not sure what is mfc?
17:11 edrin what is aop?
17:11 jonalv aspect oriented programming
17:12 jonalv but what does mfc stand for?
17:14 jonalv edrin? ping
17:15 edrin it's a warper around win32 c++ api that is supposed to make c++ development in windows more easy. a developer does not have to care about everything but just use existing c++ classes that already contain the "usual time consuming and always the same code to create a window/define a push button/..."
17:23 jonalv aha no it's not that
17:23 jonalv :)
17:23 edrin yes , i am reading about it
17:24 jonalv I am sorry it's almost likte int the matrix: "no one can be told that the Spring is..."
17:24 jonalv But only almost
17:24 jonalv :)
17:24 jonalv But it's far easier for you to read what someone else sais than it is for me to explain it...
17:25 jonalv It took me or less 200 pages of a book to realise what I could use it for but that was a year ago hopefully things are a bit clearer nowadays...
17:26 edrin where do we want to use it?
17:26 edrin or better, where do you want to use it?
17:27 jonalv I want to use it for loggin and exception handling but also to record stuff done in the gui and making a script out of it
17:27 jonalv It aslo great when working with databases through object relational mappers like Hibernate and iBatis
17:28 jonalv but in those cases it could be encapsulated into just one plugin
17:28 jonalv in the logging and exception hadnling it would need to be spread out over many plugins and that's what troubling me...
17:29 edrin is spring a "new" java runtime with "extended" features?
17:29 jonalv no not really
17:29 edrin can you give me a line of code as an example?
17:30 jonalv nope but I can give you an example from the structuredb plugin
17:30 jonalv say we want to edit and store the name of a bunch of structure stored in a database
17:31 jonalv normally we would have to first create a connection, start a transaction
17:31 jonalv and if something went wrong rollback the transaction or if nothing went wrong commit the transaction
17:31 jonalv right?
17:32 jonalv With Spring AOP the transaction dealing is encapsulated into another layer
17:32 jonalv so When the method on a manager object get's called Spring starts the transaction for us
17:33 jonalv and if the method throws an exception Spring will catch it and rollback the transactin
17:33 jonalv this is done by proxiyng the manager object
17:34 jonalv So it's actually a wrapper which gets the method call and performs something (anything) before the method get's runned and then also can do thiungs afterwards
17:34 edrin it's something like a debugger
17:34 jonalv So you won't have to worry about transaction management
17:34 edrin you can just add event listeners to "your code"
17:34 jonalv well in debugger you can step forward and such things
17:34 jonalv yes kind of
17:35 edrin that means it is a java extension
17:35 jonalv In Spring you define all of this wiring in an xml file
17:35 jonalv Well you could view it like that but it's nothing magical about it
17:36 edrin you define it and the compatible code/compiler will just add it to the code
17:36 edrin hehe
17:36 edrin the matrix, that's a good example
17:36 jonalv the manager get's instantiated in a factory and you ask the factory for your manager object. What you get is a proxy which encapsulate the real object
17:36 jonalv There are other aop frameworks that do precompiling and stuff like that
17:37 jonalv What page are you reading?
17:37 edrin the code executes/lives it's "life", and the spring sits in the back/behind taking care of everything
17:37 jonalv There are many ways to do AOP and Spring is doing it mostly though proxies
17:38 edrin i wonder if java will get this as default
17:38 jonalv yep you let's spring take care of those strange things that you otherwise would have spread out all over your code...
17:38 jonalv I don't ehink so...
17:39 edrin this thing is really great for hackers
17:39 jonalv why?
17:39 * jonalv needs to go soon...
17:40 edrin because hacker can hook everthing.
17:40 jonalv hehe yea..
17:41 jonalv well I am off
17:41 jonalv bye
17:42 edrin cyu
18:30 edrin bye egonw
18:30 edrin or cu later
18:30 edrin ;)
18:30 edrin left #bioclipse
20:42 edrin joined #bioclipse
20:43 edrin hi
20:49 edrin sneumann, yangming; who are you?

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