OE IRC 22Feb07

From GO Wiki
Jump to: navigation, search
5:20 PM

midori: Hi, John

midori: I've just put in a comment on that "lock view" SF item.

5:25 PM

Alex_MGI has joined the channel

jrichter: Hi, Midori. I'll take a look at it before we meet.

tberardi has joined the channel

jrichter: Okay, Midori, I looked at the comment.

midori: And? Was it pithy and insightful?

5:30 PM

jrichter: Certainly. There's one other complication:

hjd_mgi has joined the channel

jrichter: If the locked panel was the primary selector, we'd need to hand off primary selectorship to another panel. If there wasn't another eligible panel, we'd have to show an error message and let the locking operation fail.

jrichter: I think that's the only way to avoid another slew of unacceptable behaviors.

jrichter: Were you thinking this would go into version 1.1, or is this something we're saving for OBO-Edit 2.0?

midori: That sounds okay, at least for the way I use lock view (or would use "lock screenful").

midori: I don't mind, so let's not hold up 1.1 any more for it. Now that I understand the current behavior better, I can wait.

jrichter: There's an interim fix for OBO-Edit 1.1 where we just show a warning popup if you lock the screen around a non-visible path.

midori: OK.

jrichter: I could do that in 5 minutes.

5:35 PM

jrichter: Alright. I'm going to change the current "lock path" item into a feature request, and create a new bug report for the OBO-Edit 1.1 warning message.

midori: Great! Shall we start the main chat? (Those of you who wonder what on earth we've been talking about, see the SF item: https://sourceforge.net/tracker/index.php?func=detail&aid=1654414&group_id=36855&atid=418257)

tberardi: Let's go!

midori: I know Jane has that feature request, but she's not on yet. I'll drop an email; meanwhile, how are things going? I've just downloaded 1.1 beta20.

hjd_mgi: I just tested beta20 for load, search, select tern, save; all work

jrichter: As far as I know, all the major problems are worked out (unless there are lots of people who share Jen's reasoner problem). We're just putting out brush fires now. I also believe the user's guide is up to date.

midori: I haven't gone anywhere near the reasoner since beta18.

tberardi: Haven't checked the u guide yet for the changes I suggested but will now.

jrichter: The beta18 reasoner is very, VERY sick! Don't trust that reasoner!

j-lo has joined the channel

jrichter: Tanya: I can't remember what changes you suggested. We may need to add these as bug tracker items, since the bug tracker is the beginning and end of my todo list.

midori: Don't panic, I was only looking at its speed, not its output. 

jrichter steps into the next room for a few seconds, please keep chatting

5:40 PM

tberardi: John: OK, I will add them in ASAP.

jrichter: Jane's here now. Was there a feature request we needed to hear about?

j-lo: it's kind of a feature request AND POSSIBLE BUG

j-lo: sorry, I wasn't shouting there

jrichter: GOOD!

j-lo: basically, people who want to view multiple ontologies are having problems

j-lo: Erika (who can't access IRC so can't join us today) has been working on GO

j-lo: and needs to load ChEBI at the same time

j-lo: so she can check stuff in ChEBI

j-lo: but when she saves, she only wants GO

Alex_MGI: Is there any way to have two instances of OBO-Edit running?

j-lo: there is, but it's gets slow

5:45 PM

j-lo: she has to save with a filter to get just GO - is that right?

jrichter: Alex: It also depends on your platform. On Macs, you need to use the platform-independent version to run two instances.

midori: Yes, the save can be done with filters, but that's kind of a hurdle for new users.

jrichter: Jane: Yes, but I'll bet that she's having a problem where Chebi specific synonym types and categories keep ending up in the GO.

j-lo: I've also had this problem - e.g. I want to check SO, and already have GO loaded

midori: Alex - you also need plenty of RAM.

jrichter: I've been watching this discussion, so I think I understand the problem people are having. There are actually at least 3 related problems here.

j-lo: so my suggestion is we have a 'browse only' mode

jrichter: 4 related problems, then.

jrichter: Jane - "browse only" would be specified when you loaded an ontology, then? It would mean something like "never save members of this ontology, and don't allow its contents to be edited"?

j-lo: exactly

5:50 PM

j-lo: is that tricky?

jrichter: Allow me to hold forth for a moment. The general solution is VERY tricky, but maybe some pieces will suffice...

jrichter: We really want 4 different things...

jrichter: 1) Live import capabilities. It's annoying that you have to save your work and load a whole new editing session if you want to bring in an ontology that you didn't load initially.

jrichter: It's about 2 weeks of intensive work to implement this, because there's a lot of tricky work involved in making imports undoable, in resolving dangling links when an ontology is imported, and in dealing with identifier clashes.

jrichter: 2) Fixes to the existing filtered save system. Right now, when you do a filtered save, there's no way to filter which categories and synonym categories end up in the output file. The best way to solve this is to associate namespaces with SynonymCategories and Categories. But that's quite a bit of work.

5:55 PM

jrichter: There's a simpler fix where we add some options to the (already incredibly cluttered) filtered save gui where users can decide to only save synonym categories and categories that are actually used in the ontology. This will filter out the Chebi-specific stuff in Erika's example, because once the Chebi terms are removed, those categories will no longer be in use.

jrichter: The simple fix would take 3-5 days. The more complex fix would take 1-2 weeks, I think.

jrichter: 3) Read-only phase 1: Automatically filtering out "read-only" ontologies when saving. This could probably be done in less than one week, but it has some strange interactions with...

tberardi has to go to lab meeting - Sue is presenting, would be bad form to be late 

jrichter: 4) Read-only phase 2: Making "read-only" ontology members non-editable. Here's where things get really tricky. There's a ton of GUI work to make certain ontology members non-editable. We'd need big changes to at least 25 gui components. We'd also need a way of specifying what a "change" is. Is adding a child to a term a change? What if the new link is in a different namespace from the read-only term being linked?

6:00 PM

jrichter: There's at least a month of work here.

j-lo: hmmm

Alex_MGI: What about writing a browser only app just by getting rid of the editing and reasoner stuff that can run side by side with OBO-Edit, at least as a stop-gap measure?

jrichter: I feel like the only change we can justify in OBO-Edit 1.1 is the simple fix mentioned in (2). The rest of the stuff should probably wait until OBO-Edit 2.0

midori: I agree; the rest are much bigger feature requests.

j-lo: I agree - we can't wait much longer for 1.1

jrichter: 2) is a real problem, because it actually results in bad data ending up in an ontology file. The rest of the stuff is just inconvenient.

hjd_mgi: I think the simple fix would be fine for me. The thing is, if no multiples are loaded, it just defaults to save all synonym categories

hjd_mgi: Could the fix you suggesed also work with MORE than 1 ontology loaded??

hjd_mgi: I mean more than 2

jrichter: Absolutely. I think the best way to do this is only allow this option if the user is doing a filtered save.

jrichter: And we save the generalized fix for OBO-Edit 2.0. That's something we ought to do anyway, because I've been wanting to associated dbxrefs with namespaces for a long time anyway. I can go ahead and do that while I'm adding namespaces to categories and synonym categories.

hjd_mgi: Sounds good then. I can see where during an editing session, I might want to load chebi or anatomy and cell in making a GO term compatible with the terms used in these.

jrichter: ("associate dbxrefs", not "associated dbxrefs")

6:05 PM

midori: All sounds good ...

jrichter: I'll add the filtered category thing to the bug tracker now...

j-lo: great - Erika will be delighted

jrichter: Okay, I think we're resolved on this then. Any other business, Midori?

midori: Have you heard anything more about the paper?

midori: (I haven't.)

jrichter: Not a word.

midori: I guess the only other thing is to keep testing 1.1 beta20. Did you want to use the "testing tasks" wiki page any more?

6:10 PM

jrichter: I feel like we're well beyond those tasks now. We'll use that page again in the future, but I don't think we need to pay too much attention to it anymore.

Alex_MGI: (By the way, I can't get beta20 to launch on my mac after using either the platform specific installer or the platform-independent installer.  Using OS X 10.3)

Alex_MGI: (Reinstalled B16 at one point, and it worked fine.)

midori: Crud. The platform-independent one has launched fine for me; I've got OSX 10.4.

jrichter: I'll do some tests on my mac and see if there's a mac-specific installer problem.

hjd_mgi: I was able to load and use in on OSX 10.4

j-lo: my Mac specific version works fine

j-lo: OSX 10.4 like Harold

jrichter: Alex: see if wiping out your config data helps. I tightened up the layout checking code in this last version, so if your layout has any errors, OBO-Edit might be choking on it.

Alex_MGI: Okay, will try.

Alex_MGI: (where is the config data located?)

jrichter: Run these commands:

jrichter: tar czvf oboedit_config.tar.gz ~/.oboedit

jrichter: rm -r ~/.oboedit

6:15 PM

jrichter: If it works, send me oboedit_config.tar.gz so I can investigate why you couldn't start up.

Alex_MGI: should I reinstall or just try now?

jrichter: Just try it.

jrichter: In the meantime: anything else we need to discuss?

Alex_MGI: The platform independent version, which is currently installed, will still not launch.

midori: John - is there anything we beta-testers should pay particular attention to in beta20?

jrichter: Alex: let's take this offline. Please send me any console output you're seeing from the platform-independent version so I can figure out what's going on.

jrichter: Midori - Make sure the reasoner works quickly, make sure that the dag viewer gives accurate results, make sure that selections work correctly.

Alex_MGI: The org.geneontology.oboedit.launcher.OBOEdit thing launches (it did before too), but it goes no further.

Alex_MGI: (was typing when you wrote to go offline)

midori: OK! When I send the chat transcript around, I'll ask people to test beta20 fairly intensively (pity some of our intrepid crew had to miss chat). I'd love to be able to release it.

jrichter: Well, we can't release beta20 in any case. We need to wait for Erika's requested changes first. I'll try to have a release out by the weekend, but I'm not promising anything.

6:20 PM

midori: Oh yeah.

midori: I assume it's still worth testing in beta20?

jrichter: Absolutely. I need to know that the new changes work.

jrichter: We shouldn't need to test them again in beta21.

midori: I guess that last bit is what I was really asking about; thanks.

jrichter: Are we done, then?

j-lo: think so

midori: That's all I've got.

jrichter: Okay, let's get on with our lives. I'll try to get a release out by tomorrow night.

jrichter: Farewell all!

j-lo: bye!

6:25 PM

jrichter has left the channel saying: Konversation terminated!

Alex_MGI has left the channel

hjd_mgi: beam me up scottie

hjd_mgi has quit the server saying: Quit: Chatzilla 0.9.77 [Firefox 2.0.0.1/2006120417]