OE IRC 5Oct06

area51.slashnet.org has activated the following mode : No External Message

area51.slashnet.org has activated the following mode : Only Operator Change Topic

5:25 PM

kchris has joined the channel

5:30 PM

jclark has joined the channel

j-lo has joined the channel

j-lo: hello

jrichter has joined the channel

jrichter: Hey, folks.

tberardi has joined the channel

kchris: Hello

tberardi: Hi, sorry I'm late.

jclark: hello

jclark: Midori said to tell you that she's going to be a bit late but is leaving her chat window open to get the transcript. 

jrichter: Okay.

jclark: shall we start?

jrichter: Sure. Are you in charge, Jen?

jclark: sort of

jclark: I don't have the agenda though

jrichter: Ahh... then I'll do my items first, and then Midori can add on any when she arrives.

jclark: good-oh

5:35 PM

jrichter: Item 1: John's status report:

jrichter: a) The paper

jrichter: It looks like we've finally got it ready to go. Once I integrate Jane's edit, it's ready to submit. Unless there are objections from those assembled, I'm going to submit it today. Is that okay?

kchris: good with me 

jclark: yes

jrichter: b) OBO-Edit Documentation

jrichter: It turns out that the new features in OBO-Edit require a LOT of new documentation. I'm going to have to rewrite the entire "Customizing OBO-Edit Section", and the documentation will require extensive proofreading and testing to make sure its comprehensible. I'm thinking I've got at least 2 weeks of work here.

jclark: That was my section wasn't it?

jclark: I'll take a look

jrichter: Excellent!

jrichter: I'd like people to send me their favorite OBO-Edit layout. It would be nice to have a bunch of example layouts for users to try.

jclark: is the plan that you write it and I read it and see if I can make all the stuff work?

tberardi: I like and use Jen's layout.

jrichter: That's right, Jen.

jclark: I'll make an improved one with the new gadgetry

jrichter: Excellent! I think you'll be impressed.

jclark: I checked and I am already

jrichter: Since we have 2 weeks, does anyone have a tutorial they'd like to add to the documentation (maybe based on an email you wrote or something)...

jrichter: ?

5:40 PM

kchris: I still use the default, but would really be happier if I found the instructions Midori gave me for using a layout that allowed resizing the query results window

hjd has joined the channel

tberardi: Is there something in the documentation now about finding is_a orphans?

tberardi: If not, that would be handy.

kchris: For tutorials, we did talk about the idea of providing instructions for doing a query where you limit to one ontology, e.g. only function, I still can't remember how to write that query

jrichter: Hmm... maybe a better way to do this is to solicit the working group list for "Frequently Asked Questions" like "How do you limit a search to a single ontology"

jrichter: People on the list can compose answers, and I'll pick up the slack on any remaining unanswered questions.

jrichter: Does that sound good?

j-lo: yep

jrichter: Moving on to c) OBO-Edit 1.1 Last Minute Features

jrichter: As I was writing some documentation for the layout customization section, I was stunned by how hard it was to use the layout editor. It's highly intolerant of errors, and it's impossible to tell where your mistake is if there's a problem with the layout..

jrichter: The worst part is, all that error information exists, it's just not being displayed.

jclark: funky

5:45 PM

midori: Hi guys ... finally here for real.

jclark: how's the cat?

midori: do you really wanna know? :P

jrichter: I'd like permission to update that part of the Configuration Plugin to display line numbers in the layout editor, and I'd like to add a panel that will show you any errors (with the associate line number) if you try to enter a bad layout.

jclark: yes

jclark: good idea

midori: permission? nay, I grant encouragement!

jrichter: (BTW: I've already dropped in a new editor component that does syntax highlighting and automatic tabbing)

jrichter: Cool!

jrichter: Two more things on my agenda:

jrichter: 2) The install4j installers:

jrichter: I know the popup messages are annoying, but did the new installer system work well for everyone?

midori: I haven't tested it yet ...

kchris: nor I

hjd: I had no problem with beta 5 install

jclark: it works for windows

jrichter: And no memory problems?

midori: the platform-independent beta5 opens & loads GO nicely ...

hjd: I've loaded up to 5 ontologies simulataineously, and it seems to work ok

jclark: oh, actually now you mention it I used the platform independent version

hjd: I'm using the mac installer version

jrichter: It sounds like Harold's the only one who uses the installers. Please give the installers a try on your platform; I think the problems have been solved. If everything works well, I can get money to buy the program.

5:50 PM

jclark: okay 

midori: downloading now ...

jrichter: The main problem to look for would be OutOfMemory issues.

jrichter: Last thing from me... 3) Webex status report?

jclark: we tried it on monday

jrichter: How's that going, Jen/Tanya?

jclark: it worked for me but not for jane or david

jclark: nto sure why

tberardi: We did three locations on Friday.

tberardi: Denver, Hinxton and Palo Alto.

tberardi: Prices:

tberardi: $38.50/month

tberardi: Voice options

tberardi: VoIP: 2 cents/min/person

jrichter: !

tberardi: 800 number: 20 cents/min/person

hjd: I've had a Webex with one person down in Boston (boston - bar harbor);  Didn't know it does voice; does it do video too?

tberardi: 650 number (local for Palo ALto): 5 cents/min/person

tberardi: Yes, it does video with max 5 locations.

tberardi: the voice component is extra.

tberardi: VoIP is cheap but only 2 people max can talk at one time.

tberardi: Others have to wait for the host to pass the 'mike'

tberardi: the phone numbers let everyone talk at once

tberardi: but are more expensive

tberardi: For VoIP and for web meeting, max number of attendees is 200.

tberardi: Think that's enough info?

jclark: yes

jrichter: When are we getting an account?

jclark: the pricing seems problematic

tberardi: I do want to test VoIP with at least three locations.

tberardi: I have an account we can use now at Carnegie.

hjd: I just got a mike, so I'm game

jclark: Do you think if we paid more per month then the voice would be free?

tberardi: Do we need approval to get another one for GO purposes?

tberardi: I think this is the negotiated contract with Stanford as a whole.

tberardi: And we can't mess with it.

jclark: is that able to be opened up to the whole consortium?

5:55 PM

tberardi: I could get it here at Carnegie or Mike C. could get it at the Dept. of Genetics.

tberardi: Well, technically, no.

tberardi: The account goes to one person.

jclark: I think if we have to have a stanford person on every call then it will nto work

jrichter: No, it will nto.

tberardi: Right.

tberardi: It will not work.

jclark: should we find out the cost of a non stanford account?

tberardi: webex.com

jrichter: I know: $300/month.

tberardi: !!!

jrichter: That's why I was pushing convoq.com

jclark: and VoIP?

jrichter: VoIP is always pay per play, I think.

jclark: oh right

tberardi: We didn't have too much luck with convoq, though.

tberardi: (Jen, David and I tried.)

jclark: we need to write down all these costs so I can take them to the menagement call

tberardi: I believe I sent an email to you, Jen and John.

jclark: I think it has to be the $300 one so we can all use it

kchris: I've missed something, why do we need this?

tberardi: Did that disappear into cybervoid?

tberardi: Want to take that, John?

jclark: no actually, I have it, but we need to add on the requirements doc that john is writing

jrichter: It would be helpful for:

jclark: I hadn't managed to type that yet

jrichter: 1) These meetings

jrichter: 2) Webinars

jclark: online content meetings

jrichter: 3) Remote trainings of all kinds

jclark: management calls

tberardi: Harold: I'll email you separately to set up a test.

j-lo: and we want it for content meetings too

j-lo: quite urgently

jrichter: $300/month is WAY too much. There are cheaper solutions.

jrichter: I'm looking up a price breakdown now...

jclark: but do they work?

jrichter: megameeting does

tberardi: Would you like to try having a webex meeting for our next OBO-Edit meeting?

jclark: ok right

jrichter: http://www.megameeting.com/web-conferencing-chart.html

jclark: we need to test options but it doesn't have to be this meeting

jclark: yes

hjd: was that $300 per one site, or each one?  if one site is needed for starting, and everyone else connects in, that's $3600 per year, but what would it be saving in terms of travel costs?

jclark: john should we test that next?

tberardi: No, the cost is only associated to the host.

tberardi: Attending is free.

jrichter: That page I posted shows a comparison of the major products.

jclark: cool

jclark: John can we talk about this in a separate meeting? we need to get details

6:00 PM

jrichter: Okay, let's take this offline.

jclark: we also need to do more testing

jclark: cool

jclark: what's next?

jrichter: Any other agenda items?

j-lo: John - following the content meeting we had with David earlier this week, we have an urgent feature request

jrichter: !

j-lo: it's to prevent us re-using ids

j-lo: when we're editing a section of the ontology 

midori: I don't remember any feature requests that I'd characterize as 'urgent'....

j-lo: parallel to the live ontology

midori: (we have a social workaround for the ID problem)

jclark: yes

hjd: ?

j-lo: well, that hasn't worked so far 

jclark: what did you think oboedit was going to do jane?

jclark: true

j-lo: not sure

jclark: that's the tricky part

j-lo: perhaps assign temporary ids?

tberardi: What's the problem?

jclark: because it's when we merge files that the problem arises

jclark: it's when we edit two files in parallel

jclark: and use the same id twice

jrichter: Okay, here are the possible solutions:

jrichter: 1) Assign temporary ids that are scoped per-file. When a user does some kind of "sync" operation, all the temporary ids are resolved to real, global identifiers.

jrichter: Advantages: Easy to write, doesn't require network access

jrichter: Disadvantages: All the problems of id clashes have been shunted to the "sync" operation. If two people sync without communicating, we're still screwed.

jclark: I think we have to assume that we're oblivious of the id clash

6:05 PM

jrichter: 2) Create a simple ID server on geneontology.org. When a user requests an id, the server never, EVER gives out that id again.

jclark: that would be good

jrichter: Advantages: Works every time, no possibility for problems

jrichter: Disadvantages: We'd rapidly waste our id space, it would require network access to create terms

midori: I wouldn't bother with (1) then, because if communication works, we don't need it, and if communication fails it won't help.

jrichter: We may need some combination of 1 & 2, actually. The temporary ids could be used until the user can communicate with the id server.

j-lo: yes - sometimes it's useful to be able to edit offline

jclark: true

midori: What we thought of was to have obomerge cough up a warning if terms in the changed files share an ID but not the term name.

jclark: that seems more straightforward to me

jrichter: That won't work.

jclark: oh

jrichter: OBO-Edit won't be able to tell if its an id clash or a name change.

hjd: wouldn't it be possible to just assign ranges, much like we do for entire Mods, but maybe split them per curator (mods's take care of assignment)?

midori: We'd get some "false positives" when we really just change term names, but I could live with that.

jclark: sometimes curators aren't bright enough to remember that they're editing two fiels concurrently

jclark: files

jrichter: Okay, it sounds like there are 2 proposals on the table:

jrichter: 1) Warnings for obomerge (easy)

midori: Harold - that's essentially what we decided to do pending any technical solution; we'd assign a range per meeting as well.

jrichter: 2) ID Server System (quite tricky)

midori: (for "we" read: Midori & David will be ID Range Czars)

jrichter: I can sneak 1) into the current release, but I think that 2) needs to wait OBO-Edit 1.2, if we can. Otherwise we're going to have delays.

kchris: To me, the fact that the ID server system is tricky means that it is probably not a good use of our time currently

j-lo: can we just do 1 then?

6:10 PM

jclark: I think 1 is best

j-lo: every big content change we've done has had this id problem, it's a huge time waster

jrichter: Sure. I think the ID server idea is good, though. It deserves a place on the feature request tracker. Suzi has been begging for something like this for about 2 years.

midori: We need the non-technological aspects of ID management anyway, so I'd make 1 the priority ... maybe not even  do 2 if we turn out to get on OK wit 1 plus aggressive ID range management by humans.

kchris: I completely agree with Midori

jclark: me too

jrichter: So voted. I'll add name warnings into obomerge.

midori: (fine to do 2 somewhere along the line if we or other O-E users need it later, but I don't feel the need right now)

j-lo: is that equivalent to the sync function john?

tberardi: (BTW, I just tried the Windows installer and it worked well.  OBO-Edit opened and I was able to open the ontology without a problem.  The pop-ups were a little annoying though so I ended up just clicking on 'Yes to all.')

jrichter: Jane - no. This would just tell you if two files you're trying to merge reuse an id with a different name.

midori: I've tested the Mac installer -- fine for installing, starting up, and loading GO.

jrichter: Should I go ahead with the temporary id feature too?

jrichter: Or was that voted down?

j-lo: I'd prefer the sync idea - I thought that was the vote

jclark: I don't think we need that

j-lo: I do because the obomerge method will still require hand assigning ids when things go wrong

midori: I agree with Jen. I think we've had problems in the past because we didn't manage ID ranges well.

kchris: couldn't you effectively create a temporary ID anyway, just by changing the name space?

6:15 PM

jclark: if obomerge shows duplicate ids we have easy ways to deal with that

jrichter: No! Namespaces don't effect ids!

midori: Karen - we can and do, but ... what John said.

jclark: eh?

j-lo: they're not that easy Jen!

kchris: sorry, I meant the prefix that is part of the ID

jclark: I have fixed one before

jclark: I have a script from Midori

midori: we'd like to be able to just do find/replace to convert temporary prefixes to GO:

jrichter: You know, now that I think about it, I may be able to get a more sophisticated behavior out of obomerge if I'm very, very tricky.

jrichter: Jane - would it be okay if obomerge just reassigns clashing ids?

j-lo: yes

jclark: yes that would be good

midori: reassigns them to what?

jclark: it could ask for a new range

tberardi: Midori, does anyone still fill in the go_numbers file?

midori: ... and we would need to be able to tell obomerge which file keeps its IDs

jclark: I do

j-lo: what about name changes though, John?

j-lo: won't it give them new ids?

kchris: automatic reassignment seems possibly problematic to me, based on the fact that someone, Midori?, already mentioned cases where it is just a name change

midori: Tanya - Jen does, and one of my action items from this week is to bring the rest of it up to date.

jclark: true

jrichter: Now that I think about it more, there's a way I can tell the difference between name changes and new term creation.

jclark: wouldn't it just be better if obomerge just listed the problematic ids and names?

jrichter: Best of obomerge could fix them too, optionally. Then we solve everyone's problem.

tberardi: That would certainly be a useful first step. (What Jen said.)

j-lo: I'd like it automatic

midori: could we try it with just step 1 first?

midori: automatic worries me more, actually ...

jclark: me too

kchris: automatic worries me a lot

6:20 PM

j-lo: doesn't worry me at all!

jrichter: It's equally easy to write either version, I think. Detecting the problem is a lot harder than fixing it.

jrichter: How 'bout I write the full featured version, with automatic reassignment, but we leave the reassignment features undocumented.

j-lo: okay

jrichter: A task force will go about the task of seeing whether the reassignment features are useful and safe.

jclark: sure

jclark: good idea

kchris: instead of automatic reassignment, could it bring them up with some sort of yes/no checkboxes

midori begins to suspect that j-lo is just angling to be relieved of file-merging duty ...

jclark: file merging does suck quite a lot

midori: agree with j, j, & k 

tberardi: I was wondering about that.

jrichter: Karen - it probably couldn't, because obomerge is meant to be a non-interactive script.

j-lo: \j-lo never wants to do that again!

jclark: It's no fun

jrichter: You could run it in "warnings" mode, and then go ahead and do the merge if you aren't bothered by the warnings.

jclark: I did it about 16 times for the cns file

midori: oooh, I like John's latest suggestion!

jclark: there might be quite a lot of warnings

jclark: Could it list the terms in those warnings?

midori: 'tis the nature of the beast, I'm afraid ...

jclark: true

tberardi: (A way to get your heart racing while at the computer.)

jrichter: Jen - absolutely.

midori: good idea to have it show terms

jclark: That would be good

jclark: Then we'd know where we are

jclark: We can manually reassign and then merge

j-lo: it's the error-prone human fixing of said warnings that worries me

jrichter: Would it be useful to have a little command line utility called "reassignoboid" that would take a file and reassign an id in it?

jclark: We have a script that midori wrote

j-lo: it's so easy to introduce new problems when you're doing search and replace on the file

jclark: the script rocks

midori: tell me more, 'cos I suspect the answer (to John) is yes

jrichter: Usage:

jrichter: reassignoboid <oldid> <newid> <inputfile.obo> -o <outputfile.obo>

6:25 PM

midori asks in an aside to Jen - which script is that? I can't remember!

jrichter: or, if you're nasty:

jrichter: reassignoboid <oldidregexp> <replacementidregexp> <inputfile.obo> -o <outputfile.obo>

jclark: it take the obo file and a two column tab delimeted file with old id in first column and new id in second column and replaces

jclark: splendid

j-lo: would you have to do that for each id you've reused?

midori: oh, that one! using it on obo files is your adaptation; I wrote it for something else originally

jclark: oh right.

jclark: It's fantastic

jclark: I like it a lot

midori: j-lo - the regexp option would let you do a batch at once

jrichter: Jane - obomerge could do it automatically. You'd run this script if you wanted to change a specific id.

jrichter: I could write the script so that you could give it an unlimited number of oldid:newid pairs.

jclark: that would be good

midori: yup

jrichter: Should this be done for OBO-Edit 1.1, or wait until 1.2?

jrichter: (I think it's < 1 day)

midori: we'll use it as soon as it's available, so if it's quick , let's go for 1.1

jrichter: Okay, for OBO-Edit 1.1: obomerge improvements, reassignoboid script!

jclark: Yay!

midori: yay too

jrichter: I've got a fussy baby in my lap. Any other items?

jclark: can you tell me the path to the eclipse oboedit file so I can import from cvs?

jrichter: I'll post that on the developers wiki today.

jclark: ta

6:30 PM

jclark: nothing else for me then

jrichter: Oh! One more thing for the grammar cops.

jrichter: Is it a "User's Guide" or a "Users Guide"?

midori: for next time, we should report on 1.1 beta testing (which means I ought to do more testing in the meantime   , if possible ...)

tberardi: Users' Guide

jrichter: (I think 1, on the "Farmer's Almanac Logic

tberardi: Well, are you addressing a single person or all users?

midori: User's Guide (for one user) or Users' Guide (for plural users)

jrichter: I'm addressing "The User"

tberardi: Then, 1.

jclark: User Guide?

midori: in that case, User's Guide

jclark: That would be a guide to users I suppose

jrichter: Good. Nothing changes then.

jrichter: I'm off to write some scripts and so on!

kchris: singular User's Guide seems file, it's a guide to the use who is reading it at the time

jclark: Good luck 

tberardi: Keep me in the webmeeting loop.  I'm happy to try the VoIP option, if we desire.

jclark: ta

tberardi: on webex.

jclark: yes

jrichter: Bye.

jclark: bye

midori: likewise on the web meeting front

tberardi: bye

j-lo: see ya

jrichter has quit the server saying: Client exited

jclark: cheerio

tberardi has quit the server saying: Quit: 

kchris: bye

j-lo has quit the server saying: Quit: 

jclark has quit the server saying: Quit: jclark