hjd_mgi has joined the channel
hjd_mgi: hey
MelissaH: hi there
Alex_MGI: Hi, I'l be in and out a bit here.
12:35 PM
Alex_MGI: Is John there?
MelissaH: yeah , he's fixing bugs right now.
Alex_MGI: Oh
hjd_mgi: Couldn't even load a signle ontology with b14 unfortunately
MelissaH: really? worked fine for me.
MelissaH: are you mac?
12:40 PM
_jrichter: Harold's problem is probably the same as everyone else. It's fixed in beta15, but you can get beta14 working by wiping out your config directory.
Christian has joined the channel
_jrichter: (But if you wait until beta15, you won't have to)
MelissaH: dunno what I did to make it work.....
hjd_mgi: I think I'll just go with 15!
_jrichter: You can probably also get it to work by disabling all the verification checks.
Alex_MGI: Is B15 imminent?
MelissaH: I didn't do anything, it just worked for me. I am pc.
_jrichter: Today or tomorrow for beta15.
_jrichter: Mel - It's a bug related to the verification plugin's cycle check. If you never enabled it, this problem might not show up for you.
12:45 PM
hjd_mgi: Mel, I am Mac, and I think I do have the verification plugin on
hjd_mgi: I just started it up and when I pulled down plugin manager, it was bold; I clicke don it and it became grayed out, but nothing else happened
hjd_mgi: I assume a menu should have sprung up.
MelissaH: when I open my verification plugin it has cycle check on for reasoner, save and manual.
hjd_mgi: So it must be a mac specific but in the plugin or how it is being handled
hjd_mgi: bug, I mean bug
12:50 PM
_jrichter: Harold - I'm uploading a patch to sourceforge right now. If it works, we know you're getting the same bug as everyone else.
_jrichter: Where are the EBI gals? Are we supposed to start without them?
12:55 PM
_jrichter: Okay, the patch is now on sourceforge on the beta14 release section. Drop the oboedit.jar file into the runtime/ subdirectory of your OBO-Edit installation
_jrichter: Did it work?
hjd_mgi: It now starts up as beta 13, and it loads the GO ok; and the verification plugin now invokes a menu window
_jrichter: beta13?
_jrichter: That's all wrong.
hjd_mgi: Yes, that is what the splash screen says. I downloaded the oboedit.jar file that just appeared on sf
1:00 PM
_jrichter: What file size do you get when you run ls -l oboedit.jar ?
_jrichter: (I think sourceforge is serving up the wrong file)
hjd_mgi: -rw-r--r-- 1 hjd staff 1805692 Dec 21 12:56 oboedit.jar
hjd_mgi: The date stamp might be due to the mac,
_jrichter: Yeah, that's totally wrong. Sourceforge is serving up the wrong file. Let me try something else...
hjd_mgi: I think it is going to mirror OSDN automatically to fetch the file
1:05 PM
hjd_mgi: I think everyone is already on holiday in Hinxton!
_jrichter: We'll just do this quickly then!
_jrichter: Who hasn't signed off yet, and why?
MelissaH: me, but I do so now, with flourish.
1:10 PM
MelissaH: I still have one bug thats not fixed in b14, though.
MelissaH: the color picker in the graph viewer.
_jrichter: Ah, yes. I don't know what that bug means.
Christian: I wanted to talk about the versioning which I created a feature request.
_jrichter: (Try the new patch on sourceforge, Harold. Download it, rename it to oboedit.jar, and put it in your runtime dir)
MelissaH: (I invited Christian along today, hoping that the EBI girls would be here.)
_jrichter: Hi, Christian. Do you want to say more about that while I look up the feature request.
Alex_MGI: I was going to test B14 this morning but got sidetracked by a sick child. Guess I'll wait until B15.
Christian: It appears that neither GO nor ZF AO is using a tag such as version: and I wanted to get a discussion started for all obo files incolouding versioning.
_jrichter: Mark Gibson (the phenote developer) is also really concerned about this issue.
_jrichter: Actually, the GO does keep updated version numbers, they just don't use the correct tag.
Christian: Yes, I would like to see a robust versioning process all users are using.
_jrichter: CVS automatically supports some in-file version tracking via the $ version: VNUM $ tag.
_jrichter: That's why GO files usually have a line like: remark: cvs version: $Revision: 5.33 $
Christian: That is correct but I understand ObOedit does not read that free text format line.
_jrichter: (Sorry, I guess the tag is Revision)
_jrichter: You're right.
MelissaH: And, not all of the obo files at obo use sourceforge cvs.
_jrichter: But I think this is less an OBO-Edit problem than a GO Consortium problem.
1:15 PM
_jrichter: The simplest way to address this problem is to just get the various ontology editing groups to add the data-version: $Revision: whatever$ tag to their files.
Christian: Also, the CVS versioning only works if you are using text files. If we decided to zip files then the cvs versioning qwould not work any more. One more reason to use a separate process.
_jrichter: True.
hjd_mgi has quit the server saying: Client exited
MelissaH: for those of us that are not GO, we would still like a way to read and write version number.
Christian: Do you think it makes sense to have oboedit at least display the version?
hjd_mgi has joined the channel
_jrichter: Yes, I do. But let me say some things about the very difficult issues behind this seemingly simple problem (these difficult issues have caused me to punt on this question time and time again)
_jrichter: Issue 1) Multi-user synchronization...
hjd_mgi: The attempted download hung everything and I had to force close Firefox (along with chatzilla). Can you email the file to me
hjd_mgi: at hjd@informatics.jax.org
hjd_mgi: I think SF is trying to go to a server that the file isn't on yet or something
_jrichter: Let's say we give users an "increment version" number. Now let's say that Melissa and Christian are both editing the ontology at home.
_jrichter: They both make a change, and they both naturally increment the version number.
_jrichter: Oops. Two different versions of the file have the same version number.
_jrichter: There are two ways to prevent this kind of problem: resolve the version number through some kind of check in system (the CVS solution), or create a centralized version number issuing authority.
MelissaH: so right now zfin does both.
1:20 PM
MelissaH: GO gives us a version number, and we have our own internal cvs.
_jrichter: Both solutions are kinda tricky to implement in a generalized way. The centralized versioning authority seems even harder to me, because we'd have to write server software from scratch, and every project would need their own server (or at least their own account on a central server)
_jrichter: But even worse is Issue 2) - The ambiguous meaning of version numbers in OBO format
_jrichter: We're used to working with files that have fairly stable contents. GO terms appear in gene_ontology.obo, ZFIN terms appear in zfin_anatomy.obo.
Christian: I agree synchronization is a problem and it goes beyond OBOEdit.
Christian: However, I think whoever is checking in a new version is somehow
_jrichter: But often a user will load several ontologies with different version numbers, edit some terms, and then save all the ontologies into one file. What's the version number on that file?
Christian: responsible to update the 'version' tag, that may not be automatic but should work, right?
Christian: I agree, doing it in OBOedit is complicated. We should first try to get a more manual process in place.
1:25 PM
Christian: For that we should agree on using a single tag in the header
_jrichter: It would be easy for me to create a "Versioning Plugin" for OBO-Edit 2.0 that basically allows a user to type in a version number for their file (maybe with some automation to easily increment the version number). But we'll need big red warnings that says "This version may be incorrect if multiple users are concurrently editing!"
_jrichter: The correct header tag is data-version.
_jrichter: I think the OBO spec says that "version" is deprecated.
_jrichter: (It wasn't clear whether it referred to the version of the data or the OBO format version)
hjd_mgi: (from the bakcground to jr: ok, used another machine to get the file, renamed, and then put it in the directory. It now opens as b14 and has the same error loading any ontology, and the verificaiton plugin grays out)
Christian: data-version would be fine with me
_jrichter: (Harold - please zip up your .oboedit directory and send it to me)
_jrichter: We should talk to the GO consortium about adding the data-version tag to the GO file (although they may have already done this earlier this week)
_jrichter: We should also talk to the OBO foundry about requiring the use of the data-version tag as one of the criteria for well-formedness of their ontologies.
Christian: I agree we have to start making the GO's and otheres adopt this tag...
_jrichter: Here are some action items for me, then:
_jrichter: 1) Ask GO to use data-version tag
_jrichter: 2) Ask Chris M. if we can require the use of data-version on the OBO site
1:30 PM
_jrichter: 3) Edit my comments here about the complexity of automatic versioning and post them on the OBO-Edit Wiki. Maybe we'll get some comments that will help us figure out a good way of dealing with this.
_jrichter: 4) Build a version editing component for OBO-Edit 2.0
_jrichter: Do these address your immediate concerns, Christian?
hjd_mgi: done
Christian: I think it wold be helpful to write up how we can do versioning in muli-user environment
Christian: Yes, that is a good start.
Christian: Thanks, John!
_jrichter: You're welcome!
MelissaH: Thanks everyone for letting us barge in our chat with versioning issues.
_jrichter: Any other issues before we go? Has anyone used the Question Plugin (I'm looking at you, Mel)
Christian has quit the server saying: Quit: Christian
MelissaH: nope, but promise to do it today.
_jrichter: Let's call this meeting done, then. Harold, could you stay on the line while I try to reproduce your problem?
hjd_mgi: (must be SOME party at Hinxton!): remember that one with the crowns, etc. at your first dat edit thing?
hjd_mgi: yup, I"ll hang on.
Alex_MGI: Will you be working next week, John?
_jrichter: Yeah, except for the 25th, when I'll be doing a one man version of Mel Gibson's "The Nativity Story". It took me a week to mix up all the fake blood I needed!
1:35 PM
Alex_MGI: Okay, I'll get to my testing on B15 sooner rather than later so I can sign off.
_jrichter: Thanks.
MelissaH: alrighty, I'll email you RE the Q plugin. happy holidays with much blood.
_jrichter: Same to you. Bye!
Alex_MGI: I just can't wait 'til the bloody holidays are over, myself!
MelissaH: hee hee, Ho Ho.
MelissaH has quit the server saying: Quit: MelissaH
Alex_MGI: Merry Christmas then. Bye.
Alex_MGI: I'll save the transcript too.
_jrichter: Good news, Harold. Your stderr file looks like you're having the same problem as Jane, which is fixed in beta15. It just looks like the patch didn't take. Try the beta15 release tonight.
1:40 PM