OE IRC 9Mar06

From GO Wiki
Jump to: navigation, search
4:55 PM
jclark: Great system though
jrichter has joined the channel
Harold: the gang is assembling
jclark: yes
jclark: lots today
jrichter: I'm still unwell, but I can pull it together enough to text chat. My fingers aren't sick.
jclark: take it easy there
midori: glad you're that bit better!
MelissaH has joined the channel
jrichter: Thanks.
5:00 PM
jrichter: According to the IRC god, it is now 100. Should we start, or are there others we are waiting for?
midori: Jane will be along in just a minute ...
MelissaH: Are you all better?
J-Lo has joined the channel
Alex_MGI: Is there a specific agenda for today?
midori: ... but how 'bout we decide what to chat about?
jrichter: I've got something:
midori: (so, no, there isn't an agenda yet)
jclark: we could ask john what he thinks of last time's conclusions
jrichter: Someone had asked a few weeks ago exactly what tests have been implemented in OBO-Edit.
jrichter: I can talk about those, and what I intend to add.
J-Lo: okay, great
midori: Jen - good idea. John - I remember seeing that list you sent to the list. Is there more?
jrichter: If Jen is talking about the priority list before we can release OBO-Edit, I thought they were great conclusions.
jrichter: There are some more now, midori.
pascale has joined the channel
jrichter: As I understand it, the main sticking point continues to be the documentation.
pascale: hello
midori: OK! If John is happy with last chat's outcome, I think all we need to do is ask: any progress, or an ETC (C for completion of user guide)? Then we'll move on to the tests. OK?
midori: John: yuppity.
jrichter: Has anyone systematically gone through the existing documentation?
jrichter: It might be helpful to debug/clarify the existing docs while I write new documentation.
Harold: No, and I was going to suggest that when we consult it, if we can't find what we are looking for, we should send john an email
midori: I haven't really, but I have given it a once-over just to see what topics have anything at all.
MelissaH: me too.
5:05 PM
pascale: not systematically, but I noticed there was no documentation for "parents"
jrichter: What do you mean, Pascale?
jrichter: The parent plugin isn't finished, but I had no intention of creating a special section for "parents"
jrichter: (I probably should have said "the parent plugin section of the documentation isn't finished" above)
pascale: I looked at the documentation, and it was a topic where there was no documentation
jclark: I was interested to know about the xml layout stuff but that section seemed to be empty.
midori: I did enjoy the 'What is an ontology' section 
pascale: it's not the only one - I know
jrichter: Perhaps it would be helpful to divvy up the documentation amongst the working group members so that everyone controls a few sections.
Harold: right now, I see the entire section on data adapters appears to be empty
jrichter: If you get a section that exists, it's your job to make sure it's clear and free from errors.
jrichter: If you get a non-existent section, it's your job to badger me mercilessly until I include it.
midori: May I suggest: the first thing to do is list the completely empty topics - no point trying to evaluate non-existent docs. 
jrichter: That's what I meant.
jrichter: I don't plan to add any topics that aren't in the current outline.
midori: And we can badger J efficiently as well as mercilessly.
jclark: with a cron job?
midori: OK, who wants to assemble the empty-topic list?
jrichter: It's probably easiest for me to do that.
Harold: ok; also empty is ID generation, The reasoner, Plugins,  command line, customizing, parts of cross-products AND the data adapters.
jrichter: Or not.
5:10 PM
Harold: and filters 
jrichter: Right now I'm working on the filtering section.
MelissaH: and history
midori: We're getting close to a complete list right here!
jrichter: Also "The Parent Plugin" section of Editing is empty.
jrichter: All the rest of Editing is complete.
J-Lo: I'll review the filtering and rendering section when it's done if you like - I've used those features quite a lot.
jrichter: So now that we've got this list of incomplete topics, let's have everyone sign on for one incomplete topic and a couple of complete ones.
MelissaH: Maybe we can assign completed sections to be reviewed, and assign empty sections to be completed.
midori: Undo/Redo looks fine 
jclark: I'd like to take 'customizing OBO-Edit'
midori: I can try to keep track of who's got what.
jrichter: Good. Jen, why don't you take Undo/Redo and Text Editing as well.
kchris has joined the channel
jrichter: Melissa:
jrichter: Would you be comfortable taking on the topics under "Viewing Ontologies"?
MelissaH: yep, and I can do the parent plugin too.
jrichter: Nice.
Harold: I can look over the ID manager once he has it... I can also do  the graph viewer since I use it a lot
Alex_MGI: I can start on data adaptors.
jrichter: Who wants "Structural Editing"?
Harold: I can look over that if you'd like
5:15 PM
jrichter: Thanks, Harold.
jrichter: Alex, will you also take "Organizing Your Ontology" and "Working with Obsoletes"?
midori: Harold - thanks, and make sure you fix the mistake in 'Merging'.
jrichter: What?
midori: If Alex doesn't want obsoletes, I'll take it.
Alex_MGI: sure
jrichter: Midori, I'd like you to take the very first section.
Alex_MGI: Sure I'll take it, I meant
midori: "A 
midori: (oops)
jrichter: Midori - What mistake is in "Merging"?
kchris: Hi, sorry I'm late. It appears that we're divvying up the sections of the User Guide to get it completed?
midori: Wait, already been fixed (sorry, haven't read it in a while)? anyway, that's the point of this exercise. (I was thinking of when merged terms were appearing in the file as obsoletes ... and thought that was what 'destroyed' referred to. my bad)
midori: Karen - yep.
jrichter: Well, wait a sec
jrichter: The point of this is to assign editing/badgering responsibilities for different sections. But I've still got to write 'em.
jrichter: Right?
midori: John - do you mean I should take What is OBO-Edit/What is an ontology/etc.?
jrichter: That's right, Midori.
pascale: that's less scary then!! 
5:20 PM
gwg has joined the channel
midori: replying-to-John-replying-to-me-replying-to-Karen: Mostly, yes, but a few intrepid volunteers have offered to contribute portions of text that you may be able to incorporate.
Harold: amelia will get all of the other sections
gwg: er, yes...
gwg: sections of what?
jrichter: Grapefruit.
gwg: I was hoping for oranges :\
jrichter: Midori - Do these sections of text currently exist?
midori: at least it's still citrus
midori: Some do, some don't.
midori: OK, pause to get a grip ...
jrichter: Well, if anyone has put together text for a section, please email them to me right away.
jrichter: That will help keep the despair at bay.
midori: Does anyone have a list of complete (or incomplete) sections right this minute?
rama has joined the channel
Harold: I'll take a stab at making the text for the graph viewer anyways
gwg: can you use any of what was said at the webinar?
rama has quit the server saying: Quit: rama
J-Lo: I'm writing a description of how to make a GO slim in OBO-Edit for a book chapter - I'm sure it can be modified for the user guide
J-Lo: includes filters and renderers
jrichter: Aha!
rama has joined the channel
jrichter: Now we might want to add a growing collection of "How To"s and FAQs to this document.
jrichter: Should I add sections for those?
MelissaH: Your How-tos seem already incorporated, but I think yes on the FAQ
midori: Maybe - do we have stuff that should be included somewhere but doesn't fit in any of the existing sections?
jclark: you could add 'how to submit bug reports and feature requests'
midori: (replying to John, not Melissa)
jrichter: I don't think so, Midori, but it might be nice to have articles that cover things like creating a GO slim that involve the use of multiple features at once.
midori: A FAQ would be good, and could go on the web somewhere ... 
kchris: GO slims seem popular enough that it might be worth them having a subtitled section to find easily
5:25 PM
midori: John - good point!
J-Lo: a troubleshooting section might be useful - like why won't OBO-Edit launch?
midori: OK, let's look at priorities ...
jrichter: J-Lo: Isn't the answer to that question usually "Broken Version"?
MelissaH: what about a section about working with multiple ontologies at once?  another multiple feature issue.
J-Lo: no, could be wrong java etc
gwg: could also be "not using the platform independent version"
kchris: J-Lo: If it doesn't launch, how would you get to the User Guide?
midori: I think top priority is still completing the existing user's guide.
jrichter: Right!
J-Lo: good question!
gwg: how easy is it to convert the user guide into other formats, John-boy?
pascale: I was think that too, but the help can be somewhere else as well, right?
gwg: (e.g. html, txt, etc.)
jrichter: The User's Guide is in HTML now.
pascale: but that sectioin could be in the read_me file
jrichter: It's just got some special bells and whistles to make it searchable.
MelissaH: Can't you just post the whole thing at the cBIO site?
gwg: so there would be no trouble in having it hosted on the GO website
midori: The DAG-Edit guide is on the godatabase.org site.
jrichter: gwg - No trouble, in theory, but in practice the User's Guide should live in the OBO-Edit cvs repository. So we'd need to build some kind of process for posting the guide online when there's an update.
jrichter: But back to Midori's point: we need to finish the guide before we start adding stuff.
midori: Yes; we should also work on these newly proposed items, just maybe second instead of first.
kchris: What's the summary of what parts are written (and maybe are ready for editing) and what parts need to be written?
midori: There are a bunch of good ideas in this transcript, so we can come back to them any time.
5:30 PM
jrichter: (Someone should go through this chat transcript later and come up with a list of action items - not it)
jrichter puts his finger on his nose
gwg misread "on" as "in"...
midori: OK, boys and girls ...
midori: Action item 1: List completed and incomplete user guide sections.
pfey has joined the channel
midori: Action item 2: Make sure all completed sections are assigned to someone, who will go through and check for accuracy and, well completeness.
jrichter: (and clarity)
pascale: so, Midori does #1 and John #2 ? 
midori: Action item 3: John will write stuff for the incomplete sections, with contributions from other working group volunteers.
jrichter: I'll take on items 1 and 2.
jrichter: I'll send out an email later today with a list of complete and incomplete sections, along with who is assigned to what.
jrichter: There will be some unassigned sections, and we'll ask for volunteers over email.
midori: Action item 4: assign newly written sections for review.
midori: Action item 5: John will act on suggestions from reviewers.
jrichter: I don't know if we need to do 4. If you were assigned an empty section, and I write something in that section, your job goes from badger to edit.
5:35 PM
midori: One possible point of confusion: if someone is assigned an empty section, do they merely badger, or also help write?
jclark: tanya says she can't get onto this chat forum. 
jrichter: Help write if you feel like it, badger if you don't.
jrichter: If you ARE going to help write, please let me know. It would be awful for us to be duplicating work.
pankaj has joined the channel
MelissaH: I have to turn off virus protection to get IRC to work
jrichter: Some places/machines firewall the IRC port. You may need to turn that off.
midori: OK ... so 4 can be folded into 2. no prob.
jrichter: Shall we move on to the test suite?
MelissaH: yes time is running short
midori: OK
jrichter: Okay - I just wanted to explain how the test suite works, and what tests currently exist.
jrichter: The test suite tests are meant to be utterly exhaustive. Each test in the suite is performed several dozen times on randomly selected terms in the ontology.
jrichter: Each test is executed and then undone, and checked against an expected correct result for that test.
jrichter: If there's a mismatch, the test fails.
MelissaH: presumably on all platforms?
5:40 PM
jrichter: No, Melissa.
jrichter: The tests are only run on my compilation platform.
jrichter: This is because these are low level - non-interface tests.
jrichter: Platform-specific issues almost always manifest at the interface level, which I can't test automatically anyway.
midori: the idea is, if something fails for John, no point sending that version on to us
MelissaH: ok, got it.
kchris: John, Mike Cherry uses OBO-edit to convert the OBO-format to old GO format (if I understand correctly). Is there, or can there be, as test for this?
jrichter: The tests also make sure that the HistoryGenerator mechanism gets the same answer for the test as the rest of OBO-Edit. (The HistoryGenerator is the guy who builds the obodiff results, handles text edits, and creates the results in the Term Change Tracker Plugin).
jrichter: Karen - I'm working on that test.
kchris: Cool 
jrichter: That one won't be as exhaustive, but I'll make sure that it is always generating the same flat files.
jrichter: Anyway, here's what I'm testing now:
midori: (drumroll)
jrichter: 1) Comment changes
jrichter: 2) Definition changes
jrichter: 3) Namespace changes
jrichter: 4) Removal of namespaces
jrichter: 5) Removal of Secondary IDs
jrichter: 6) Removal of Synonyms
jrichter: 7) Changes to synonym scope
jrichter: 8) Creation of secondary ids
jrichter: 9) Creation of synonyms
jrichter: 10) Term text changes
jrichter: 11) Basic reasoner testing
jrichter: 12) Tests to the history save/load mechanism
jrichter: 13) Adding new terms/id generation
jrichter: That's it, for now.
midori: excellent.
jrichter: That last one adds 100 new terms and makes sure they have the right ids.
5:45 PM
gwg: do these tests also check the flat file output is OK?
MelissaH: Does the reasoner testing include testing of newly added relationships?
jrichter: Melissa - reasoner testing is done on a specific file, so it only tests the relationships in that file.
jrichter: gwg - There's no test for that, right now. What do we mean by "ok"?
tberardi has joined the channel
jrichter: It would be really easy to add a test where I load a file, save it, and make sure that the saved file is the same as the one we loaded in.
jrichter: Would that do?
gwg: that OBO-Edit is outputting what you'd expect it to - e.g. that the 10,000 newly added terms all have IDs, namespaces, names, etc., etc.
gwg: it would really be better to have it go through a set of tests and save the file, and we'd have another file to compare it against
gwg: sorry, I didn't phrase that very well
jrichter: gwg - I'll think about the best way to make that happen. I think my proposed, simple test would work as a first step, since what we're looking for is striking rearrangements of terms, problems with escape characters, or lost fields.
midori: would be better than nothing
5:50 PM
jrichter: The file writing system isn't tied into the editing system at all, so I think making changes and then saving would only make the test more complex without telling us anything new. We just have to choose a good starting file.
jrichter: We have 9 minutes. Any question/additional issues?
jrichter: questions, I meant
midori: or suggestions for next time?
Harold: Are we still putting something together for the meeting?
midori: yes, I'm working on a poster
J-Lo: \jane raises her hand
jrichter: Yes, Jane.
J-Lo: are we going to make beta 16 a main release?
jrichter reminds Jane that she should use /me to make statements about herself
jrichter: No!
jrichter: There's a serious flaw in the way dbxrefs are written.
jrichter: Which I've fixed.
jrichter: We need one more version.
midori: Anything else it needs (besides a gap-free user guide)?
J-Lo: okay. 
MelissaH: does anyone else's save button not work in b16? 
pascale: me, I can only save_as
jrichter: Also: That serious flaw seems to have polluted the current version of the ontology. We need to figure out a way to put those dbxrefs back together.
midori: we (GO-eds) always use Save As anyway .. force of habit
pascale: but that's not just from b16, it's been a while
gwg: that's OK, I can do it
jrichter: "Save" has never been implemented in any version of OBO-Edit.
pascale: Amelia and her super powers...
gwg: but beware yoghurt...
jrichter: There are subtle reasons why it's hard to know what to do with a simple "Save", and I've never thought it was worth the effort to figure it out.
J-Lo: What ahppened to the dbxrefs?
5:55 PM
jrichter: Dbxrefs with commas in them are getting split into 2 dbxrefs.
gwg: they had commas in them, which OBO-Edit interpreted as being dbxref separators
MelissaH: It takes a long time to save as, which I do every 10 minutes or so...
midori: see bug report SF 1440953
jrichter: Melissa: Plain save would be no faster, I'm afraid.
kchris: Maybe the simple 'Save' shouldn't even be there then, seems confusing for it to be there, but not work.
Harold: I like the save as because it always asks about file overwriting; just makes sure. 
MelissaH: Probably because I always fear saving in the wrong place- one more reason to show the file loaded.
rama: I agree with Karen
jrichter: You're right, Karen. I should take that off the menu.
jrichter: Melissa: Have you submitted that request to the Feature Request tracker?
midori: action item 6 ... 
jrichter: If not, could you...
MelissaH: Will do.
jrichter: Oh! One last thing!
jrichter: There are a number of very old bugs sitting at the bottom of the bug tracker on sourceforge.
midori: indeed.
jrichter: Could everyone take a moment to look at the old bugs. If you submitted a bug, and you know it's no longer valid, please remove it.
jrichter: Or post a comment saying it should be removed.
jclark: many of them seem to be mine
Harold: If the bugs pertain to an older version, should we just remove them now? Or at least, if it's to a version that is more than one-back?
jclark: I'll check
midori: action item 7 goes to Jen (everyone else's entries are fairly recent)
jrichter: They should be removed if they're no longer valid.
jclark: They're probably fixed. I'll take a look.
jrichter: Any bug that's listed for a version earlier than beta16 will be purged in a week.
MelissaH: gotta run soon, one last question, is the reply to email you have on your working group listserve the one to write to?
midori: for the save thing - you can create a tracker item at https://sourceforge.net/tracker/?group_id=36855&atid=418260 (different tracker from the bug one)
jrichter: The email address is geneontology-oboedit-working-group@lists.sourceforge.net
6:00 PM
jrichter: Melissa, I'll add you to the list now.
tberardi has left the channel
MelissaH: I am on that list, I meant yours- I haven't gotten any responses from you.
MelissaH: just sick/moving/prengnant?
jrichter: My email address is john.richter@aya.yale.edu
midori: ... and so the badgering begins!
MelissaH: ok, good.  thanks. 
jrichter: I haven't seen any messages from you in the last week, though. Please email me again. If I don't respond, call me at (303) 519-4760, 'cause you're getting junk filtered.
MelissaH: yes sireeee!!!! so many questions that I can't possibly bother the working group with all of them.....
midori: who wants to send the transcript to the list?
jrichter: (not it!)
jrichter touches his nose
kchris: Unfortunately, having been late, I don't have any of the key early decisions about who's responsible for which sections, so mine won't be that useful.
MelissaH: ok all, bye for now.
MelissaH has quit the server saying: Quit: MelissaH
jrichter: (Actually, I'm checking to see if my IRC client created a good transcript. If it did, I'll mail it out)
midori: I've got it from 4:55 pm GMT, so I can send mine.
jrichter: Midori, do you have timestamps in your transcript?
jrichter: Mine is coming out untimestamped in odious rich text format.
J-Lo: she left, but yes
pankaj has quit the server saying: Quit: pankaj
jclark: she's gone away. Should be back any minute
jrichter begins humming
Harold: mine have time stamps since 11:45AM est
6:05 PM
jrichter can be heard to say "Tall and blonde and young and lovely, the girl from Iponema goes walking..."
jclark: how do you get a transcript from the webinar irc?
gwg wonders if she should tell John-boy that Midori is actually short and dark haired...
jrichter: You wait and wait and wait for me to get off my lazy but and put it online
jclark: oh
Harold: The "conversation" program does a nice job of saving the transcript
kchris: That's what I use too
midori: yes, I have timestamps
jrichter hums "Short and dark of hair and goofy, the girl from Detroit goes walking"
pascale: Jen, I think if you use the applet from the obo-edit web site, you cannot get a transcript
jclark: sad
jrichter: Ahh... now I see what you mean, Jen. Yeah, you can't save files from an applet, usually.
jrichter: That's why I don't like 'em.
midori: Detroit doesn't really scan ...
jrichter: De-e-troit.
gwg: don't like girls from De-e-troit or applets?
jrichter: a) Don't put words in my mouth, gwg b) We've clearly run out of useful things to say.
kchris: OK, I think we're done. Midori has agreed to send her timestamp containing transcript to the list, so till next time
midori: OK, I've got a train to catch. Are we done?
jrichter: Clearly. Farewell, folks.
jclark: yup
jclark: cheerio
Harold: I hear the train a-coming /
J-Lo: goodnight/morning
jrichter has quit the server saying: Quit: jrichter
J-Lo has left the channel
Harold: ka-la
pascale has quit the server saying: Quit: pascale
Harold: I mean ka-pla!
jclark has quit the server saying: Quit: jclark
midori: bye all! (transcript will come tomorrow)
Alex_MGI: Thanks John et al.
pfey has left the channel
Harold has left the channel
gwg: see yas
gwg has quit the server saying: Quit: Hasta la vista, baby.
Alex_MGI has quit the server saying: Quit: Quitting!