Difference between revisions of "OE IRC 8Mar07"
From GO Wiki
(New page: <pre> radon.slashnet.org has activated the following mode : No External Message radon.slashnet.org has activated the following mode : Only Operator Change Topic 5:15 PM jclark has joine...)
|Line 1:||Line 1:|
radon.slashnet.org has activated the following mode : No External Message
radon.slashnet.org has activated the following mode : No External Message
Latest revision as of 14:09, 30 June 2014
radon.slashnet.org has activated the following mode : No External Message radon.slashnet.org has activated the following mode : Only Operator Change Topic 5:15 PM jclark has joined the channel jclark: hi 5:20 PM midori: hi 5:25 PM jclark has quit the server saying: Ping timeout Anon440 has joined the channel Anon440 has quit the server saying: Quit: Anon440 jclark has joined the channel jclark: b21 is working much better for me. midori: Good. I've been getting a lot of unhandled exceptions. jclark: oh dear jclark: me too actually jclark: my layout is working now midori: Let's whinge about it when John turns up. jclark: good plan midori: I've been playing with layouts a bit too. jclark: the verification plugin is broken for me too jclark: they're fun aren't they? jclark: Mines is now better than before jclark: the dbxref library can shirnk jclark: shrink midori: Oh ... I thought the verification plugin was working for me. jclark: so there's more space for definitions and stuff midori: I learned that ChEBI is really really not is_a complete. jclark: It's just when I try to save and it runs the verification plugin and I click links nothing happens jclark: ah! 5:30 PM jclark: chebi has some really funny relationship types too jclark: circular ones jclark: we're strangely short staffed today aren't we? MelissaH has joined the channel midori: I've often wondered what's supposed to happen when you click the things that look like links ... I've had nothing happen in so many betas that I've stopped expecting much ... midori: Hi Melissa! MelissaH: Hi midori, Hi Jen! MelissaH: how are you all doing? midori: ok jclark: I think it's meant to take you to the term if you click on a goid jclark: good thanks. jclark: you? MelissaH: good-a bit swamped here. know any curators looking for a job? jrichter has joined the channel jclark: lots! jrichter: Hi folks. jclark: hi midori: hi MelissaH: really? please send them our way, we are hiring two. MelissaH: hi john jclark: where are you? MelissaH: eugene oregon jclark: I'll have a think midori has activated the following mode : Private midori has de-activated the following mode : Private jclark: what happened there? MelissaH: channel mode? jrichter: Midori's gone mad with power. I think the channel mode refers to whether or not new people can join, but I'm not sure. midori: That was a "what does this button do?" moment - -the button in question has a pair of sunglasses. jclark: cool! midori: No, John, I think Midori's just plain gone mad ... jrichter: See! I'm not the only one who creates sucky, incomprehensible icons. It's a tradition! 5:35 PM jclark: great! jclark: did the lights dim when you clicked the button midori: I'm scared to try the spy sihouette button now ... MelissaH: how come I have no icons? feeling left out. jclark: I have none either jrichter: In OBO-Edit 2.0, the "obsolete term" icon is going to be a cupcake with an umbrella crying a single tear. jclark: some people on e-mail are sad that they can't get in MelissaH: thats just plain weird, john. tberardi has joined the channel jrichter: Hi, Tanya. Alex_MGI has joined the channel tberardi: Hi John Alex_MGI: Hi tberardi: I think Chris and Petra are trying to find their way here. hjd_mgi has joined the channel hjd_mgi: arrrrr midori: why the umbrella? jrichter: Oops. I was supposed to tell Nicole and Mark how to join too! jrichter: The umbrella represents the everpresent possibility of disaster from the heavens. Duh. MelissaH: I'll email them now, just emailed chris and Petra jclark: shouldn't the icon be a lightning strike or something? jclark: wrath of the gods and all jclark: should we begin? hjd_mgi: JOE BTFSPLK : He walks around with a perpetually dark rain cloud a foot over his head. 5:40 PM hjd_mgi: Once he appears on any scene, dreadfully bad luck befalls anyone in his vicinity. jclark: except for the very thirsty midori thinks we should begin ... before JOE BTFSPLK notices us ... cjm has joined the channel cjm: found you! midori: yay! jrichter: Uh oh. A real person is here. Put your beers down and lets get started. MelissaH: hi chris, you missed all the great talk of doom and terror.... MelissaH: shhhh! tberardi: *gulp* midori: but fear not (oops, pun ...), you can catch it all in the transcript. jclark: are we off then? jrichter: Ahem. The latest release should have fixed all the outstanding bugs, except for the weird category bug that only Melissa's machine can produce. cjm: hey, petra is stuck over on macnn.net. She is using conversation. Can someone tell me how she connects & i will pass it on MelissaH: I am good at finding those. cjm: John: I just found two bugs, one appears to have been newly introduced in b21, and it is serious for people running the reasoner from obo2obo jrichter: Let me guess: it locks up in a headless graphical environment. jrichter hits himself in the face and yells "stupid!, stupid!" 5:45 PM jrichter: That's really easy to fix. cjm: I don't think that's it - see http://sourceforge.net/tracker/index.php?func=detail&aid=1676684&group_id=36855&atid=418257 midori: oh wow -- lots of bug reports in the last couple hours hjd_mgi: I'm looking at the OBO-edit list: should there be a notice there about a b21? there was for b20; MelissaH: I got one Alex_MGI: me too midori: I got a message because I'm monitoring oboedit -- that's how we're doing it these days. jrichter: Chris - it's the bug I thought it was. It's just manifesting a little differently than I thought. jrichter: Do people mind if I ask some questions about this last round of bug reports? jclark: no pfey has joined the channel midori: not at all, if it'll get 'em fixed! tberardi: Yay, Petra made it! midori applauds pfey: yes! pfey: by webinar pfey: don't like that so much though.. Anon322 has joined the channel jrichter: Jen's first problem with her layout seemed to be caused by a stray space in one of the plugin ids. So that's not a bug. jclark: the end result was that there is just a mistake in the help docs jclark: easy to fix by search and replace jrichter: Jen - in your second bug report, how were you running the verification plugin when the hyperlinks failed to respond? Did you run it manually, or did it launch automatically during a save? jclark: save 5:50 PM jrichter: Hyperlinks don't work during any of the autolaunch events because of the kind of popup window I have to use to show the verification results. There are a few things I can do to fix this: jrichter: 1) Disable hyperlinks when a verification results popup appears jrichter: 2) Don't use a modal window for hyperlink reports. The drawback is that if you click anything besides the warning results, the warnings window will be placed behind your regular OBO-Edit window, and could easily be forgotten. That would be weird, because now you've got a "Save" process waiting for a response from you, but you've gone on and started fixing your ontology. jclark: that would be okay midori: #1 would be adequate to avoid confusion/disappointment. jrichter: Actually, the more I think about that one, the worse it seems. You could get some really bad behavior if you hit the save menu item several times and a few of those windows built up in the background. jclark: would that mean there would be no hyperlinks? jrichter: (I was referring to option 2, in my last statement) jclark: yes jclark: I meant 1 jclark: I think the hyperlinks are useful if they call up the term in question jrichter: Jen - there would be no hyperlinks unless you opened the Verification Plugin directly. jclark: oh I see jclark: I could just do that then jrichter: Maybe I should remove the hyperlinks but add a message about how you should open the verification plugin to view these results and correct any problems. jclark: good plan jrichter: Action item! jrichter: Any objections? midori: nope jclark: nope 5:55 PM MelissaH: who is anon322? jrichter: Probably just a friendly lurker from the irc server. MelissaH: Hello anon322, come learn about obo-edit bugs with us!!! jclark: that happens to me if I am logged in twice from two clients jclark: I am not but someone else might be MelissaH: next? jrichter: I'm done with bug discussion. What's next on the agenda? midori: Well, I've been getting an awful lot of unhandled exceptions in beta21. Anon322: hi this is mark - is my login anon322? its sposta be gus?? jclark: mark who? jrichter: Gus was probably already taken, so your IRC client made up a new name for you. I think you can change your name with the /nick command. jrichter: Mark of Phenote fame. cjm: Mark Gibson is part of the berkeley/NCBO team jrichter: Phenote is considering forming its own working group, so I invited Mark to lurk on this chat. jclark: good plan midori welcomes Mark by any user name. Anon322 is now called markgibson MelissaH: hey mark glad you made it. markgibson: hi cjm: just a quick plug for phenote: it's main function is for phenotype annotation, but it is highly adaptible and could be used for GO annotation and indeed for GO editorial functions such as editing mapping files and cross products. Anyway, back to our scheduled service.. 6:00 PM jrichter: damned spammers... tberardi is off to her 10 am meeting, will read transcript tberardi has quit the server saying: Quit: jrichter: Midori, is there another scheduled agenda item? Otherwise, I had a few things that might be interesting for the second half... Alex_MGI: John, is there a Java version issue, or am I crazy? jrichter: Is OBO-Edit barfing on Java 1.4? hjd_mgi: I had no problem with b21 just now; I have 10.4 java 1.5 Alex_MGI: (both is an acceptable answer) Alex_MGI: Yes, bug report is in. Alex_MGI: Will not launch, same problem as B20 on OS X10.3 Java 1.42 jrichter: That's a real possibility. I'm going to require Java 1.5 for OBO-Edit 2.0, and I'm developing it simultaneously with the current OBO-Edit. It's possible the compiler is getting confused about what I want. I'll check my Eclipse settings before the next release. Alex_MGI: Great, thanks! midori: Sorry, Tim got home and I got sidetracked ... midori: Anyway, the only other thing is the exceptions; I'll send you a stderr. 6:05 PM jrichter: Hooray! Please post a bug report too. jrichter: Okay, here are a few things I wanted to talk about... markgibson: yea i really recommend java 1.5! jrichter: You're gonna love it Mark. I've already modified most of the datamodels to use that autocasting dealy. jrichter: Anyway: First, a tip. It's totally safe to delete your .oboedit/stderr file if it starts to get too large. You could probably safely delete your stderr file at the beginning of each day, before starting OBO-Edit to save disk space. markgibson: lovely! you mean with collections right? jrichter: Yup. jrichter: Actually, all that stuff is probably in CVS already, in the OBO-Edit 2.0 fork. markgibson: great jrichter: Second thing: Yesterday Melissa and I were working on tracking down a bug, and I had her install the TightVNC server on her machine. It gave me the ability to share control of the machine with her so I could see her bugs as they manifested on own equipment. And TightVNC has an HTTP mode, so it works with most firewalls. MelissaH: warning- it is scary to have a poltergeist fighting for control of you mouse. jrichter: I think this technology + skype would be really useful for diagnosing those hard-to-find bugs that I can't reproduce on my machine. 6:10 PM MelissaH: But I agree that its worth fighting the poltergeist. jclark: would webex not do that? jclark: we now have a licence jrichter: Webex would do that too. But is everyone in the working group allowed to initiate a webex session? Can any webex participant share their desktop? midori: I think either would do ... jclark: you can initiate and anyone an share desktop jclark: it's really easy and we're paying for the support so we may as well use it jclark: you can control oboedit on anyone's computer jclark: we've tried it midori: We'll have five people who can initiate sessions (and John, you'll be one of them, so try not to go mad with power); once the session has started, the host can pass control to any participant. jrichter: Hmm. Jen, let me talk to you offline after this, and then I'll post some Wiki instructions for the Working Group so we can start doing this to diagnose bugs. jclark: david will send you your password jclark: I got mine yesterday jrichter: Sweet! jrichter: Moving on: some interesting issues came up in Mel & I's session yesterday that might impact OBO 1.3 format... jclark: you can also record footage of demos jrichter: Mel wants to create a special kind of slim ontology for PATO. She wants to be able to mark some terms as directly belonging to the slim (like we do now) but she wants to mark other terms in a way that means "this term, and all its descendants, even if they change in the future, belong to the slim". 6:15 PM jrichter: Now it turns out that there's a way to do this now... MelissaH: (Mark- george couldn't get it to work yet but made you the slim manually) jrichter: You have to create TWO categories, one that means "terms directly belonging to this slim" and another that means "terms whose descendants belong to this slim". Then you design the filter in such a way that it knows to handle the two categories differently. jrichter: (I promise it works when I do it.) markgibson: yea this would be a nice feature - clearly would help georges efforts jrichter: So here's the thing: We can do this now, but it's kinda tricky, between the two different categories and a fairly complex filter. jrichter: Do we want to include special scoping modifiers (or something) to the OBO 1.3 spec so that we can mark certain categories as having certain properties? jrichter: More importantly, if we do that, what should those special properties be? SF has joined the channel jclark: do you mean like hereditary and not hereditary jrichter: I feel a little weird about just having a single flag that says "include this term AND its descendants in this category". That feels too specific, like the kind of thing where we'll be forced to add thousands of other flags down the line. 6:20 PM MelissaH: hi nicole SF: hi jrichter: Plus, simple heredity will get us in trouble in the future. You probably don't really want to include ALL descendants. You probably want to include all is_a or part_of descendants, or something. SF: sorry to be so late. jrichter: I don't expect us to solve this problem here, I just want to put it out there so people can think about it: What kind of special features would you like to add to categories? What features would make it easier to design slims in the future? pfey has quit the server saying: Quit: pfey pfey has joined the channel jrichter: But here's a question we can answer right now: to you folks who have designed slims, what tools would have made the process easier? jclark: I think if you could do a search and say 'mark all these as x slim' it would be good midori: automagically including _parents_ would be handy jrichter: Midori - automagically adding *existing* parents, or automagically adding all possible future parents? SF: so i haven't actually used oboedit to do this function, but i imagine having a pannel side-by-side, with the drag and drop feature...to be able to copy/paste a single selected item, and/or an item and its descendants (like the previous discussion). 6:25 PM jrichter: Hmm... that's a neat idea. We'll have to wait for OBO-Edit 2.0 for that, but that could have all kinds of uses. SF: and yeah, i would like to maybe have a different copy command to choose whole 'kinds of' descendants (part of, is a...by relationship) midori: I was thinking of existing parents, but including future parents would help keep a slim in sync with the full ontology, so it would be most cool. SF: i like the automatic parent inclusion. SF: again, would that be by relationship type? jrichter: Here's something to consider: there's no formal specification of how categories map to a slim file. jrichter: Right now, we just mark every single thing that belongs to a slim. But we don't have to do it that way. jrichter: We could just as easily mark all the LEAVES of a slim, and then let the filter automatically add in all the ancestor terms. jrichter: It's really the filter you use that determines what the categories actually mean. midori: that would be much more convenient than having to mark each term jrichter: And we can do it right now, without me changing a single thing. SF: ooh SF: thats actually seems pretty logical midori 's mind boggles jrichter: But that's a GO editorial policy decision, so maybe it needs to be discussed in the appropriate working group before that change happens. MelissaH: (john can, that doesn't mean the rest of us are savy enough, but he'll teach us) SF: except its more work than if you just want to choose a whole category...and if that category is pretty close to the root, then its a lot of wasted effort jrichter: Nicole - I'm not sure what you mean by that. midori: Can we have a description to take to the relevant group (indeed, to identify the relevant group(s))? SF: ...to mark each leaf 6:30 PM jrichter: Here's the proposal - Slims are no longer defined by marking every term in the slim. midori: Would leaf-marking and ancestor-marking have to be mutually exclusive? jrichter: Hmm... I guess not. midori: but go on; didn't mean to interrupt SF: what happens if a term gets moved to have a different ancestor? how does that affect a slim? jrichter: Nicole - the slim would change. The new ancestors would be included in the slim, and any old ancestors that used to be in the slim but are no longer connected to an explicitly marked term would be removed. SF: is that automatic, or does the user have to mark each of them? jrichter: Automatic. SF: excellent jrichter: Let me step back and talk about how slimming works now, and how it would work under this new proposal. jrichter: Right now, slimming is done by applying a filter like this: [Self] [Category] [equals] "MyCategoryName" jrichter: I'm suggesting that we change the filter to: jrichter: OR jrichter: [Self] [Category] [equals] "MyCategoryName" jrichter: [Descendant] [Category] [equals] "MyCategoryName" jrichter: Right now, if you mark a term as belonging to a slim, but none of its parents belong to the slim, that term is either discarded or appears as a root (depending on your root algorithm). 6:35 PM jrichter: In the new scheme, that orphaned slim term would automatically be connected to the ontology. jrichter: The drawbacks of the new system is that EVERY path to a term will appear in the slim. jrichter: ("are" not "is") jrichter: (even though I only listed one drawback) jrichter: Comments? jclark: I think you need to talk to jane jclark: she does lots of slims Alex_MGI: Are there case where we don't want all the paths to a term? midori: agreed - she's done more with GO slims than most of us jrichter: Alex - I can imagine lots of cases like that, particularly when we start adding new relationship types to the GO. Alex_MGI: I see jrichter: But that doesn't mean any cases like that actually exist right now. midori: Alex - -I'm not sure, but it seems that any paths via transitive relationship types would be desirable; over non-transitive relations probably not. jclark: I'm afraid I have to go now. See you soon. Alex_MGI: Could one then filter out the non-transitive relationships? MelissaH: I can't see why not jrichter: Alex - that wouldn't be quite enough, because you could still have some unwanted orphan terms in there, even if you got rid of the non-transitive relationships. jrichter: I think I need to add two new search aspects: "Transitive Descendants" and "Transitive Ancestors" 6:40 PM midori: I like that idea ... jrichter: That's pretty easy to do. Is this something we want to push into the 1.1 release? jrichter: (I'm calling it < 2 days) jrichter: (And in my head I think it will take closer to 2 hours) Alex_MGI: I think you should get Jane's opinion, perhaps. midori: I don't have a strong opinion; and yes, we should hear from Jane. Alex_MGI: (I'm a little concerned with feature creep.) jrichter: Okay, I'll move this discussion to email so we can get some more opinions. We'll also want the editorial group to weigh in, because it's important that all the ontology editors understand the implications of this change. jrichter: And, with that, I think we're out of time. MelissaH: yeah- even though we have this need for PATO, I have little or no experience making slims SF: this will fall in george's lap, or marks midori: sounds like a plan MelissaH: Glad I could be of use creating more work for you john! hjd_mgi: I just have two more questions unrealted to this I want to mention hjd_mgi: related hjd_mgi: 1. Have we every figured out how the dropped namespace happens?? Is an OBOEDIT beta doing this? jrichter: I haven't investigated this question at all. No one seems to know where it's being introduced, as far as I could tell from the discussion. jrichter: (Although I did drift in and out...) midori: nope, I think it's a text editor glitch (which I can now catch with the verification plugin) jrichter: Who is using a text editor and why? 6:45 PM hjd_mgi: As long as we are certain that it isn't OBO-EDIT... that would not be a feature! midori: me, because some people (Alex, you know who you are ) ask for tons of related terms at once, and even with the clone feature it's too damn tedious to do it in the gui hjd_mgi: arrr jrichter: Midori - think about an in-program interface you'd like in OBO-Edit 2.0 for batch term input. Alex_MGI: (It would be great to have a regulation term generator.) jrichter: Now that the OBO-Edit core is mostly stable, we're gonna start adding lots of sugar in the next version. jrichter: Harold, did you have something else? hjd_mgi: 2nd minor issue: Sourceforge monitor might not be working: I *have* been monitoring OBOEDIT download page (I just checked and it says I'm already monitoring), but I don't get any notices of update from the monitor2. The "downloads" tracker also seems to be off; it only has it recorded that there was one download (mac); no downloads for platform independent, etc. hjd_mgi: Thats it from me Alex_MGI: Have you checked your spam filter? Alex_MGI: The emails from from SF not the mailing list. jrichter: I'm not too worried about the download tracker - sourceforge likes to do periodic batch updates for statistics, so we'll probably see a jump in numbers in about 24 hours. jrichter: Also, Harold, you might want to make sure that SF has a usable email address on file for you. They've wiped out some user account data over the last few years by accident. You might have been victimized. hjd_mgi: Not in my spam as far as I can see: when would the notice have gone out about b21 6:50 PM jrichter: Between midnight - 5am your time. hjd_mgi: I do get all of the bug report stuff, etc. from sf midori: mine's timestamped Wed, 07 Mar 2007 19:48:26 -0800 MelissaH: btw- anyone know how to render for terms with more than one parent of a given type? or shall I make this a feature request? hjd_mgi: No, it's definitely not in the spam box then jrichter: Anyway, since everybody else is getting the emails, I think we can surmise that Harold's SF account is specifically to blame here, so I'll start emailing Harold personally until this all gets worked out. jrichter: Mel - that can't be done right now. The filters aren't expressive enough to say things about specific collections of parents. jrichter: But that kind of thing will be doable in OBO-Edit 2.0 via scripted filtering. jrichter: Don't ask. MelissaH: ok cjm: (stepped away for a while.... sorry to be out of sync, just wanted to echo Alex's concern about feature creep. Slim-makers are generally power-users, so they can in general uses post 1.1 betas) jrichter: I won't add any slim-related features to OBO-Edit 1.1 unless the email discussion suggests an immediate, pressing need. jrichter: We're well over time. I say let's wrap this up. jrichter: I love you all! MelissaH: yeah, right....... jrichter: Midori, would you like to sing the goodbye song? hjd_mgi: oooo 6:55 PM midori: na na na na ... jrichter: That was really inspiring. I'll see you all in two weeks, and I'll try to get these bug fixes out as soon as possible. MelissaH: thanks john, see ya hjd_mgi: and remember jrichter: Bye. hjd_mgi: no matter where you go hjd_mgi: there you are MelissaH has quit the server saying: Quit: MelissaH jrichter: HA HA HA! That's the dry Northeastern wit I've heard so much about! hjd_mgi has left the channel SF: thanks john jrichter: Thanks for coming Nicole. jrichter: I'm off, folks. Bye. jrichter has quit the server saying: Quit: Konversation terminated! SF: bye! SF has left the channel markgibson has quit the server saying: Quit: markgibson Alex_MGI: bye Alex_MGI has quit the server saying: Quit: Quitting! pfey has left the channel 7:00 PM