Camelia, the Perl 6 bug

IRC log for #bioclipse, 2007-09-13

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

All times shown according to UTC.

Time Nick Message
05:35 Nout joined #bioclipse
06:23 rojasm_ joined #bioclipse
07:00 edrin joined #bioclipse
07:29 jonalv joined #bioclipse
07:45 egonw joined #bioclipse
07:48 edrin hi egonw
07:48 egonw hi edrin
07:56 * jonalv summons olas
07:59 edrin where is he?
08:00 jonalv at home?
08:04 edrin holidays?
08:05 jonalv he was home sick yesterday
08:05 edrin oh
08:05 jonalv But he was hoping to be better today...
08:09 thomas_ku joined #bioclipse
08:15 edrin ok, cu later
08:46 EskilA joined #bioclipse
08:56 olas joined #bioclipse
08:58 masak joined #bioclipse
08:59 olas hi all
08:59 masak hi olas
09:02 olas hi masak
10:15 EskilA_ joined #bioclipse
10:54 edrin joined #bioclipse
10:54 edrin hi again
10:54 edrin hi olas
10:57 olas hi edrin
11:00 edrin olas: what are the current deficiencies of Bioclipse 1 so we need a full rewrite of the core API?
11:04 olas edrin: lots of stuff
11:04 olas stability mostly
11:04 olas no junit tests
11:04 olas and resource management
11:05 olas and we should do it the Elcipse way with IResource instead of BioResource
11:05 olas so that we can hook in external plugins like SVN
11:05 olas Maybe also let plugins register support for an extension
11:05 olas we shall investigate all this
11:06 olas also a more tighter relying on scripting language
11:06 jonalv Is IResource big and complex?
11:06 olas so that we can record user activity and build scripts interactively
11:06 olas jonalv: not really
11:06 olas but very customizable what I think
11:07 olas not really expert in it (yet!)
11:07 jonalv ah but you are looking in to it?
11:07 olas I will
11:07 olas http://wiki.eclipse.org/The_Official_Ec​lipse_FAQs#Workspace_and_Resources_API
11:08 olas We are moving towards more Eclipse-based design, instead of rewriting our own
11:08 olas I think the idea is good, let's see where this ends
11:09 olas but we will NOT ship the IDE with bioclipse
11:09 olas 99% sure about that
11:09 jonalv What is in the ide anyway?
11:09 olas This somewhat limits the scripting capabilities, but we are building a user-centric application, not a developer-centric application
11:10 olas IDE is for development
11:10 jonalv In what way is it limiting scripting capabilities?
11:10 olas yes
11:10 olas It is something to investigate
11:11 olas but generally the IDE clutters the workbench with developer options
11:11 olas not focused on life science
11:11 olas we MUST make bioclipse simple!
11:11 jonalv ah but are they really related to scripting?
11:11 olas what do you mean?
11:11 olas IDE contains developer tools
11:11 olas for programming mostly
11:12 jonalv Hm you said that it limits scripting capabilities and I wonder in what way...
11:12 olas automatic builds..
11:12 olas launch configurations..
11:12 olas ANT support
11:12 olas programming langugages editor support
11:12 olas and so on
11:12 masak if someone really wants that, we could perhaps provide easy links to it
11:13 masak for the 1% of the people who want it
11:13 olas yup
11:13 olas we should
11:13 masak but I agree that it should not be the default
11:13 jonalv Well an editor could be a good thing for scripting but not everyone will need it
11:13 olas but Bioclipse is for the masses
11:13 edrin olas: I think one good point for a rewrite is in fact the aim to make a product stable release. That's very important
11:13 olas yes
11:13 olas edrin: that's the key point
11:13 edrin but it will be a lot of work
11:14 olas yes, but we are some people who will share the burden now
11:14 edrin i think so, too
11:14 olas mostly here in Uppsala
11:14 olas jonalv and masak and EskilA will assist me here in Uppsala
11:14 olas with design and implementation
11:14 edrin good efficient
11:15 edrin olas: maybe have a scripting workspace?
11:15 olas yes, that is one idea
11:15 jonalv Well I need to get going. I have a meat space meeting planned in with someone named Ola... :)
11:15 olas scripting should be hidden from standard user
11:15 olas sure
11:15 olas I'll be in my office
11:16 edrin i had several ideas how to design a scripting based workflow thing when i traveled back from Uppsala
11:16 olas cool
11:16 olas I'm looking into EMF now
11:16 masak edrin: tell us more :)
11:16 masak and have a look at the dedicated page for it
11:17 * masak gets the link
11:17 masak http://wiki.bioclipse.net/index.php?ti​tle=Talk:A_Meta_Language_for_Bioclipse
11:17 olas If you define your model in EMF it's easy to do a graphical editor for it using GEF, with objects. DnD, connectors etc
11:17 masak hm, and http://wiki.bioclipse.net/index.php?​title=A_Meta_Language_for_Bioclipse of course
11:17 olas We should definitely design in EMF
11:18 olas and then consider building a GUI on top of it (like a workflow)
11:18 edrin there are two important key points to consider in the design: 1. have scripts be called recusrively in a synchronous or asynchronous way. 2. how to push data from one script to another...
11:19 masak edrin: yes, I've also started to think about script interaction
11:19 masak asynchronous calling basically requires threads, or something equivalent
11:19 edrin threads, that's right
11:20 masak by pushing data, what do you mean besides files and method params?
11:20 edrin on the other hand there should be a synchronous way, too. for tasks that complete quickly. in example web service calls will take long, local database calls not so long...
11:21 masak edrin: sure
11:21 masak I'd make the synchronous call the default, since threads can be hairy
11:21 masak and do unexpected things
11:22 edrin masak: for example: if you screen a protein for certain properties, you might use webservices... several webservices. they should run in parallel... on the other hand in some cases it does not make sense
11:22 masak maybe an asynch { ... } block or something similar
11:23 edrin so you have script1 that loads proteinsequence. then it calls script2, script3, script4 that are actually web service calls for analyzing the protein sequence. on the one hand they should run in parallel. on the other hand they need a mechanism
11:24 edrin to get the sequence loaded into script1 and give the result back to script1 when complete
11:25 edrin masak: for this script to script data transfer i would suggest a special mechanism...
11:26 edrin anyway, have to continue here now
11:26 masak edrin: I'll write what you suggested on the wiki page
11:26 edrin after this is done, i will continue with the XMPP webservice library and then..
11:26 edrin with the scripting
11:26 edrin k, tks
11:26 edrin thanks
11:26 masak np
11:28 edrin ah, and creating a workflow in my opinion should be done with a wizard...
11:28 masak that's one possibility
11:40 jonalv joined #bioclipse
11:40 jonalv_ joined #bioclipse
13:22 EskilA joined #bioclipse
13:59 jonalv joined #bioclipse
16:36 olas left #bioclipse
17:38 lambdabot joined #bioclipse

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