OE Webex 16Dec08

From GO Wiki
Jump to: navigation, search

12/16/2008 09:30:58 AM from Amina Abdulla: Hello folks
12/16/2008 09:32:57 AM from Amina Abdulla: Hi David
12/16/2008 09:33:05 AM from midori: what was it? I took a quick peek at her file, and there were all sorts of oddities
12/16/2008 09:33:25 AM from Amina Abdulla: did I miss something?
12/16/2008 09:33:30 AM from David OS: Hiya
12/16/2008 09:33:36 AM from Amina Abdulla: Midori are you talking about Jane's file
12/16/2008 09:33:46 AM from midori: yes
12/16/2008 09:34:06 AM from Amina Abdulla: yes I was going to ask you if you had any ideas on that one
12/16/2008 09:34:17 AM from Amina Abdulla: I just tried loading the file
12/16/2008 09:35:04 AM from midori: so did I, but I didn't get it to load (Nomi saw a dodgy xref that I missed; maybe that's it)
12/16/2008 09:35:34 AM from Amina Abdulla: Nomi I don't read you
12/16/2008 09:36:00 AM from midori: Nomi, switch to all participants (again ... sigh)
12/16/2008 09:36:29 AM from midori: anyway, Nomi's email just reached the oewg list
12/16/2008 09:37:34 AM from Amina Abdulla: ok
12/16/2008 09:38:59 AM from Nomi Harris: Sorry about that, Amina.
12/16/2008 09:39:07 AM from Nomi Harris: I just stepped away for a minute to get some tea.
12/16/2008 09:39:30 AM from Amina Abdulla: oh ok
12/16/2008 09:40:05 AM from Nomi Harris: So, as I said, I identified and fixed some oddities in Jane's file, but there could be more.
12/16/2008 09:40:35 AM from Amina Abdulla: ok does it still give an error on load?
12/16/2008 09:40:47 AM from Nomi Harris: Yes, I didn't fix all the errors.
12/16/2008 09:40:52 AM from Amina Abdulla: I haven't received the wg mail yet
12/16/2008 09:41:23 AM from Amina Abdulla: Ok, I'll look into it
12/16/2008 09:41:25 AM from Nomi Harris: When we encounter this sort of situation, we might want to put code in the OBO parser to deal more gracefully with these errors--at least give a useful error message.
12/16/2008 09:41:40 AM from Nomi Harris: But we can't possibly anticipate all possible malformations.
12/16/2008 09:41:46 AM from Amina Abdulla: Right I was just thinking about that
12/16/2008 09:41:53 AM from midori: that would be a big help for those of us who don't speak Java-errorese
12/16/2008 09:42:03 AM from Amina Abdulla: As and when we come accross some of these
12/16/2008 09:42:07 AM from Nomi Harris: Right
12/16/2008 09:42:36 AM from Amina Abdulla: Folks b52 has been delayed as I'm still fixing/ testing
12/16/2008 09:42:50 AM from midori: even just a better idea where in the file to look would help a lot, if it's feasible
12/16/2008 09:43:01 AM from Amina Abdulla: Testings been slow on b51 on my side
12/16/2008 09:43:08 AM from David OS: Yep - line numbers for errors are good.
12/16/2008 09:43:12 AM from midori: testing is good. I finally did a bit today!
12/16/2008 09:43:59 AM from Amina Abdulla: Yeah I'll put in a check that pinpoints line number
12/16/2008 09:44:25 AM from David OS: Actually - OE2 seems quite permissive of syntax errors. Arguably too permissive
12/16/2008 09:45:28 AM from David OS: Because of teh XP input bug, I do quite a lot of hand editing. Would be nice to catch more syntax errors early from theses
12/16/2008 09:45:43 AM from Amina Abdulla: It would be a good idead to report these to the wg David
12/16/2008 09:46:00 AM from Chris Mungall: David - yep, it's a fine line between sensitivity and extensibility. But more non-fatal warnings configurable warnings would help here.
12/16/2008 09:46:21 AM from Chris Mungall: configurable strictness
12/16/2008 09:46:51 AM from David OS: Yep - sorry - should get round to reporting. Difficult to find time to report everythiing
12/16/2008 09:48:21 AM from David OS: Hmmm - but where an OBO field has a required syntax, shouldn't that be fatal to loading?
12/16/2008 09:50:00 AM from Tanya: sorry I'm late, lots o bridge traffic
12/16/2008 09:50:30 AM from Amina Abdulla: David I'll have to check before I comment - it should report something like that though
12/16/2008 09:50:53 AM from Amina Abdulla: Hi Tanya
12/16/2008 09:51:04 AM from Chris Mungall: David - yes.
12/16/2008 09:51:10 AM from David OS: OK - just digging out an example.
12/16/2008 09:51:59 AM from Amina Abdulla: So if testing whole components seems challenging - would it help if I made the wiki more detailed?
12/16/2008 09:53:18 AM from midori: It probably won't hurt -- there are probably things that we ought to be testing, but don't even know are there (I know I found out about quite a few OE1 features just in the pre-release testing push).
12/16/2008 09:54:29 AM from Amina Abdulla: Yeah it would be easier to report errors back that way too
12/16/2008 09:54:52 AM from Jen Deegan: My testing tasks are okay as whole components.
12/16/2008 09:56:39 AM from Amina Abdulla: We haven't had many new bugs recently though - just some high priority ones I'm working on already
12/16/2008 09:56:41 AM from Amina Abdulla: ok Jen
12/16/2008 09:57:05 AM from Jen Deegan: We have quite a lot to be going on with don't we?
12/16/2008 09:57:44 AM from Amina Abdulla: Yes - I've fixed the TreViewer to display one-way disjoints
12/16/2008 09:57:59 AM from Jen Deegan: great!
12/16/2008 09:58:09 AM from Amina Abdulla: I have to look into that wrt to SO in the OTE
12/16/2008 09:58:36 AM from Jen Deegan: so that's as a stop gap so we don't need to think about whether to change how disjoints work for now.
12/16/2008 09:59:06 AM from Jen Deegan: I seem to have been volunteered to test the OTE
12/16/2008 09:59:22 AM from Jen Deegan: I think I have quite a lot already if anybody else wanted to do that.
12/16/2008 10:00:53 AM from Amina Abdulla: I guess I'll be looking into some sections of the OTE
12/16/2008 10:01:53 AM from Amina Abdulla: Jen how was your experience with protege?
12/16/2008 10:01:58 AM from Jen Deegan: it was good
12/16/2008 10:02:12 AM from David OS: I use it every day. I'm not aware of any unreported bugs with it. But perhaps I should try to thnk of some things to do with it I don't usually do.
12/16/2008 10:02:14 AM from Jen Deegan: I worked all day today with the Protege 4 developer
12/16/2008 10:02:27 AM from Jen Deegan: we were comparing functionality in P4 and OE2
12/16/2008 10:02:47 AM from Jen Deegan: I have some notes that I could tell you if you are interested.
12/16/2008 10:03:00 AM from David OS: I'm interested.
12/16/2008 10:03:00 AM from Amina Abdulla: Yeah go ahead..
12/16/2008 10:03:21 AM from Jen Deegan: We were looking to see if editing with incremental reasoning turned on might be quicker in P4
12/16/2008 10:03:41 AM from Jen Deegan: they don't currently have incremental reasoning, but they have lots of different reasoners for different situations.
12/16/2008 10:03:54 AM from Jen Deegan: We tried to load a set of files with cross products
12/16/2008 10:04:04 AM from Jen Deegan: but we trouble with obo 2 owl conversion.
12/16/2008 10:04:15 AM from Jen Deegan: however, we got one set loaded, and it reasoned in 8 seconds.
12/16/2008 10:04:32 AM from David OS: FaCT++ can be very zippy
12/16/2008 10:04:37 AM from Jen Deegan: but I'm not sure if that was totally right, what with the file problems.
12/16/2008 10:04:39 AM from Jen Deegan: yes
12/16/2008 10:05:03 AM from Chris Mungall: I am investigating and will put the results on the benchmarks page
12/16/2008 10:05:03 AM from Jen Deegan: Robert stevens said that our resoners and those in p4 would go much faster if we had lots and lots of disjoint relations
12/16/2008 10:05:10 AM from Jen Deegan: thanks Chris.
12/16/2008 10:05:25 AM from David OS: Did you try the OBO2OWl converter built into P4 or did you use Chris'?
12/16/2008 10:05:31 AM from Jen Deegan: he said we could do things like making obsolete disjoint with other ontologies.
12/16/2008 10:05:45 AM from Chris Mungall: I think we should add more disjoints because they help detect errors - but if they help reasoners that's a good secondary bonus
12/16/2008 10:05:52 AM from Jen Deegan: yes
12/16/2008 10:06:01 AM from Jen Deegan: they seemed to think it would really help a lot.
12/16/2008 10:06:07 AM from Chris Mungall: David - I used the NCBO translation which apparently is the cause of the problems
12/16/2008 10:06:09 AM from David OS: Robert Stevens gave me the same advice. one reason I've been using disjoints so myuch more.
12/16/2008 10:06:13 AM from Jen Deegan: also cc could be disjoint from process and function.
12/16/2008 10:06:46 AM from Jen Deegan: also he said things like x metabolism should be disjiont from x catabolism.
12/16/2008 10:06:51 AM from midori: we did suggest making BP, CC, and MF mutually disjoint a while ago ... it just stalled, as I recall
12/16/2008 10:06:53 AM from Jen Deegan: no
12/16/2008 10:06:59 AM from Jen Deegan: anabolism from catabolism
12/16/2008 10:07:03 AM from Chris Mungall: http://wiki.geneontology.org/index.php/Cellular_component_disjoint_classes
12/16/2008 10:07:08 AM from Jen Deegan: yes I thought you had.
12/16/2008 10:07:16 AM from Jen Deegan: I didn't realise the benefits at the time.
12/16/2008 10:07:45 AM from Jen Deegan: if it would help get incrememntal reasoning on the road it would be worth it, as that is needed for the graph viewer as well.
12/16/2008 10:07:59 AM from Jen Deegan: It would free up a lot of our current problems.
12/16/2008 10:08:02 AM from Jen Deegan: or solve.
12/16/2008 10:08:24 AM from Jen Deegan: I noticed some other good things in protege.
12/16/2008 10:08:25 AM from David OS: Will more disjoints have any effect on the speed of the rule-based reasoner?
12/16/2008 10:08:40 AM from Jen Deegan: they do a lot of their editing in a much more elaborate version of our parent editor.
12/16/2008 10:09:01 AM from Jen Deegan: they do cross products and normal relationships in the same big parent editor panel.
12/16/2008 10:09:05 AM from Chris Mungall: David - nope
12/16/2008 10:09:06 AM from Jen Deegan: I can demo it if you would like to see.
12/16/2008 10:09:27 AM from David OS: Sure
12/16/2008 10:09:36 AM from Amina Abdulla: Yeah I'd be interested in taking a look
12/16/2008 10:09:49 AM from Jen Deegan: Also wheneve a term name is mentioned in the text of a term then the name is hyperlinked to the term, and there is a back and forward button as in a brower.
12/16/2008 10:10:16 AM from Jen Deegan: what do other think? Should I demo?
12/16/2008 10:11:09 AM from midori: is there anything else for us to discuss about OE? if not, go ahead and demo. (I don't have anything.)
12/16/2008 10:12:03 AM from Jen Deegan: just looking for their pizza ontology
12/16/2008 10:12:24 AM from midori: mmmm .... pizza ...
12/16/2008 10:12:33 AM from Jen Deegan: mmmm
12/16/2008 10:12:42 AM from David OS: Funny - was just thinking that...
12/16/2008 10:12:59 AM from midori: (nearly dinnertime here ... does it show?)
12/16/2008 10:13:19 AM from Jen Deegan: yoicks! The version I loaded is not in English
12/16/2008 10:13:23 AM from Jen Deegan: would you like to see?
12/16/2008 10:13:36 AM from Jen Deegan: here goes.
12/16/2008 10:13:54 AM from Jen Deegan: can you still see the chat?
12/16/2008 10:14:00 AM from Harold: yep
12/16/2008 10:14:18 AM from Jen Deegan: the panel at the bottom is like our parent editor.
12/16/2008 10:14:37 AM from Jen Deegan: and those are just regular relationships
12/16/2008 10:14:54 AM from Jen Deegan: but the cross products also show in that same panel.
12/16/2008 10:15:01 AM from Jen Deegan: I will find one.
12/16/2008 10:15:31 AM from Jen Deegan: when I click veneziana
12/16/2008 10:15:43 AM from Jen Deegan: you just see 'hasTopping some blah...'
12/16/2008 10:15:50 AM from Jen Deegan: that is a regular relationship
12/16/2008 10:16:06 AM from Jen Deegan: when I click PizzaDeCarne
12/16/2008 10:16:11 AM from Jen Deegan: that has a cross product
12/16/2008 10:16:34 AM from Jen Deegan: Pizza and hasTopping some blah...
12/16/2008 10:16:45 AM from Jen Deegan: that's a cross product as I understand it.
12/16/2008 10:16:53 AM from Jen Deegan: are the mac people seeing this?
12/16/2008 10:16:58 AM from Jen Deegan: chat I mean?
12/16/2008 10:17:10 AM from David OS: I'm on a mac today. Loks fine
12/16/2008 10:17:11 AM from midori: I can toggle between the chat and the demo
12/16/2008 10:17:14 AM from Harold: yes; and JDR would be amazed at how the pizaa ontology has expanded!
12/16/2008 10:17:14 AM from midori: no problem
12/16/2008 10:17:15 AM from Amina Abdulla: yes
12/16/2008 10:17:15 AM from Jen Deegan: oh good.
12/16/2008 10:17:26 AM from Jen Deegan: does that make sense, what I said?
12/16/2008 10:17:28 AM from midori: but sorry, which relation is the xp?
12/16/2008 10:17:45 AM from David OS: Its the one in the Equivalent class box
12/16/2008 10:17:46 AM from Jen Deegan: the one that is under 'equivalent classes'
12/16/2008 10:17:57 AM from Chris Mungall: it is an xp if it appears under "equivalent classes"
12/16/2008 10:18:09 AM from Jen Deegan: so pizza is the genus and the the topping is the differentium I think.
12/16/2008 10:18:22 AM from David OS: has_topping blah is the diff
12/16/2008 10:18:33 AM from midori: oh. ok. NOT intuitive! but I get it now that you've explained.
12/16/2008 10:18:46 AM from Jen Deegan: would you like to see how to add a relationship?
12/16/2008 10:18:53 AM from Harold: your "toppingS" is in a different ontology or node, right?
12/16/2008 10:19:18 AM from Jen Deegan: are those the toppings?
12/16/2008 10:19:34 AM from David OS: actually - has_topping *some* blah. Midori - it becomes easy to folow if you're used to OWL-Dl Macnhetser Synta x
12/16/2008 10:19:37 AM from Jen Deegan: yes I think that's them.
12/16/2008 10:19:42 AM from David OS: Manchester....
12/16/2008 10:20:11 AM from Jen Deegan: you can tell which terms have XPs as they have a little equals sign with three bars in their brown dot.
12/16/2008 10:20:22 AM from Tanya: where's the topping ontology
12/16/2008 10:20:32 AM from midori: I'm familiar with the Manchester syntax; it's the Protege labels I was moaning about.
12/16/2008 10:20:50 AM from Jen Deegan: eek! I forgot how he did it.
12/16/2008 10:20:51 AM from midori: "Equivalent classes" didn't say "cross-product" to me :P
12/16/2008 10:20:57 AM from Harold: I think "has_topping" is a relationhsip[, not "has"
12/16/2008 10:21:00 AM from David OS: Ahh - OK. Those used to be even more confusing
12/16/2008 10:21:22 AM from Jen Deegan: did you see how if I click a name it goes to that term?
12/16/2008 10:21:26 AM from David OS: has_topping is an OWL property = instance level relationship in OBO
12/16/2008 10:21:28 AM from Jen Deegan: then there is a back button top left.
12/16/2008 10:21:57 AM from Jen Deegan: they said they'd be happy to make a skin with our terminology if we wanted it.
12/16/2008 10:22:51 AM from Jen Deegan: I really can't remember how he made that XP
12/16/2008 10:22:57 AM from Jen Deegan: shall I stop demoing?
12/16/2008 10:23:19 AM from David OS: I think you may be able to drag a relationship into equivalent class
12/16/2008 10:23:25 AM from Jen Deegan: the layout is totally configurable like ours.
12/16/2008 10:24:02 AM from Tanya: this is a time I'd love to have audio
12/16/2008 10:24:55 AM from Jen Deegan: me too
12/16/2008 10:25:02 AM from Jen Deegan: I made a relation but sort of by accident.
12/16/2008 10:25:20 AM from Jen Deegan: I'll stop demoing now.
12/16/2008 10:25:24 AM from midori: maybe we should try adding skype to some of these meetings
12/16/2008 10:25:28 AM from Jen Deegan: yes
12/16/2008 10:25:35 AM from Jen Deegan: I think that could be good
12/16/2008 10:25:53 AM from Harold: yes;
12/16/2008 10:26:03 AM from midori: on days when skype decides to play nice for me, I can starta call for up to 10
12/16/2008 10:26:11 AM from Jen Deegan: I gather that it might be possible to have OBO-Edit almost as part of protege so as to have all our cool stuff + all of theirs. Not sure.
12/16/2008 10:26:21 AM from midori: (on other days, it goes glitchy for no apparent reason)
12/16/2008 10:26:47 AM from Jen Deegan: As a plugin maybe.
12/16/2008 10:27:08 AM from Amina Abdulla: I like the idea of having a skype chat
12/16/2008 10:27:13 AM from Jen Deegan: Yes me too.
12/16/2008 10:27:19 AM from Tanya: shall we try sometime?
12/16/2008 10:27:22 AM from Jen Deegan: yes
12/16/2008 10:27:27 AM from Tanya: It beats the heck out of typing.
12/16/2008 10:27:45 AM from Tanya: there are 8 of us today
12/16/2008 10:27:58 AM from Jen Deegan: would get more discussed, more quickly I think.
12/16/2008 10:28:01 AM from Amina Abdulla: Yeah Im sure it'll be easier to get things accross and speed up testing
12/16/2008 10:28:02 AM from Tanya: SKype should be able to handle that
12/16/2008 10:28:11 AM from Tanya: on the dual core computers
12/16/2008 10:28:24 AM from Tanya: for the host
12/16/2008 10:28:24 AM from Jen Deegan: as long as it's started from dual core machine right?
12/16/2008 10:28:27 AM from Jen Deegan: yes
12/16/2008 10:28:33 AM from Jen Deegan: that seems like a plan then.
12/16/2008 10:28:45 AM from Tanya: so amina can host the webex and someone with a dc 'puter
12/16/2008 10:28:47 AM from Tanya: can host the call
12/16/2008 10:28:50 AM from Jen Deegan: yes
12/16/2008 10:28:59 AM from Jen Deegan: shall we do that next time then?
12/16/2008 10:29:02 AM from Amina Abdulla: I can host the call too
12/16/2008 10:29:04 AM from Tanya: yes please
12/16/2008 10:29:17 AM from Amina Abdulla: I have a souped up machine
12/16/2008 10:29:20 AM from Jen Deegan: it can get flaky if one computer does too many things.
12/16/2008 10:29:33 AM from midori: when is the next call -- are we going to try to squeeze one in next week, or is everyone away b y then?
12/16/2008 10:29:34 AM from Jen Deegan: right, that's a plan then.
12/16/2008 10:29:38 AM from Jen Deegan: should we stop now?
12/16/2008 10:29:51 AM from Amina Abdulla: David that reminds me - were you able to get the reasoner statistics for the - too many redundant flags bug?
12/16/2008 10:29:57 AM from Tanya: are we meeting next week?
12/16/2008 10:30:04 AM from Jen Deegan: I'm away
12/16/2008 10:30:17 AM from Amina Abdulla: Yes lets meet next week
12/16/2008 10:30:22 AM from Nomi Harris: I'll be away too
12/16/2008 10:30:23 AM from midori: I'm around, but technically on vacation.
12/16/2008 10:30:37 AM from David OS: Amina - not quite sure what you mean. Perhaps I'm forgetting something...
12/16/2008 10:30:48 AM from Amina Abdulla: Next week will be the last meeting for the year
12/16/2008 10:30:59 AM from Tanya: ok
12/16/2008 10:31:03 AM from midori: ok, I'll try to remember to tune in.
12/16/2008 10:31:04 AM from Amina Abdulla: I posted a comment on the tracker last week
12/16/2008 10:31:07 AM from Tanya: we can do a pilot SKype call then
12/16/2008 10:31:12 AM from Tanya: fewer peope
12/16/2008 10:31:14 AM from Tanya: people
12/16/2008 10:31:33 AM from Jen Deegan: Will there be minutes, for those of us who can't make it?
12/16/2008 10:31:40 AM from Amina Abdulla: Yeah lets do skype too next week
12/16/2008 10:32:02 AM from Amina Abdulla: Yes we'll rotate b/w the wg members for minutes
12/16/2008 10:32:10 AM from Jen Deegan: great
12/16/2008 10:32:24 AM from Jen Deegan: maybe also a rotating chair/agenda writer?
12/16/2008 10:32:32 AM from Jen Deegan: we do that on other calls.
12/16/2008 10:32:52 AM from Tanya: maybe the new system will prove so efficient that we'll either
12/16/2008 10:32:52 AM from Jen Deegan: or not, if you prefer to do it Amina.
12/16/2008 10:33:00 AM from Tanya: (1) meet for shorter times each week
12/16/2008 10:33:01 AM from Tanya: or
12/16/2008 10:33:05 AM from Amina Abdulla: No thats a great idea Jen
12/16/2008 10:33:05 AM from Tanya: (2) meet every other week
12/16/2008 10:33:21 AM from Jen Deegan: (3) actually get the work done
12/16/2008 10:33:28 AM from Amina Abdulla: I'd like the wg to get more involved with the agenda
12/16/2008 10:34:07 AM from Jen Deegan: maybe send a mail to ask for items on the Friday and we can put them straight on the wiki?
12/16/2008 10:34:37 AM from Jen Deegan: shall we shove off now? We're over the hour.
12/16/2008 10:34:39 AM from Tanya: no agenda = no meeting
12/16/2008 10:34:54 AM from Tanya: yup
12/16/2008 10:35:07 AM from midori: all righty then
12/16/2008 10:35:12 AM from Jen Deegan: have a good christmas if I don't see you before.
12/16/2008 10:35:16 AM from Amina Abdulla: ok. David I'll msg you on skype
12/16/2008 10:35:25 AM from Tanya: to everyone else too!
12/16/2008 10:35:27 AM from Tanya: bye!
12/16/2008 10:35:29 AM from David OS: Tanya - I agree - a clearer agenda would be good. Bi-weekly meeting too.
12/16/2008 10:35:30 AM from Jen Deegan: bye!
12/16/2008 10:35:33 AM from Amina Abdulla: Hapy Holidays everyone whos not here next week
12/16/2008 10:35:53 AM from David OS: Ok Bye. Happy Xmas - probably won't make it mext week
12/16/2008 10:36:06 AM from Nomi Harris: Happy Solstice to all
12/16/2008 10:36:11 AM from Amina Abdulla: Yeah bi-weekly when there aren't pressing issues is a good idea
12/16/2008 10:36:33 AM from midori: see some of you next week; happy ... oh, let's pick Saturnalia ... to the rest
12/16/2008 10:36:46 AM from midori: bye
12/16/2008 10:37:04 AM from Harold: see ya
12/16/2008 10:37:06 AM from Amina Abdulla: Bye All