OE Webex 19Jul07

From GO Wiki
Jump to navigation Jump to search
07/19/2007    17:29:02 PM    from melissa to All Attendees:
	hi gals!

07/19/2007    17:29:13 PM    from Midori Harris to All Attendees:
	hi!
07/19/2007    17:29:33 PM    from Karen Christie to All Attendees:
	hi :)
07/19/2007    17:31:48 PM    from Jen Deegan to All Attendees:
	Hi

07/19/2007    17:32:09 PM    from Erika to All Attendees:
	hello all

07/19/2007    17:32:23 PM    from Chris Mungall to All Attendees:
	hi
07/19/2007    17:33:27 PM    from Jen Deegan to All Attendees:
	Is your weather hot enough Erika?

07/19/2007    17:34:03 PM    from Erika to All Attendees:
	37 degrees are enough at 630 pm?

07/19/2007    17:34:24 PM    from Jen Deegan to All Participants:
	Yes that seems hot enough. :-)

07/19/2007    17:34:35 PM    from John Day-Richter to All Participants:
	I like where this chat is going!

07/19/2007    17:34:43 PM    from John Day-Richter to All Participants:
	A reminder: if you haven't set "

07/19/2007    17:34:46 PM    from Erika to All Attendees:
	Hi John :-)

07/19/2007    17:34:46 PM    from Midori Harris to All Participants:
	John's here -- everyone switch to "all participants" or he'll get lonely ...
07/19/2007    17:34:58 PM    from John Day-Richter to All Participants:
	Send to" to "all participants", please do or I'll get lonely.

07/19/2007    17:35:00 PM    from Erika to All Participants:
	Hi John

07/19/2007    17:35:02 PM    from John Day-Richter to All Participants:
	That is not hot enough.

07/19/2007    17:35:06 PM    from John Day-Richter to All Participants:
	Hi, Erika.

07/19/2007    17:35:43 PM    from Jen Deegan to All Participants:
	We're just agreeing that the weather is hot enough in Padova if it is 37 degrees.

07/19/2007    17:36:30 PM    from John Day-Richter to All Participants:
	Must be one of them there crazy Euro-measurements. If it's 37 degrees 'round these parts, it's time to burn some books to keep warm.

07/19/2007    17:36:54 PM    from Midori Harris to All Participants:
	37 °C = 98.6 °F
07/19/2007    17:37:25 PM    from John Day-Richter to All Participants:
	Agreed. That is hot enough.

07/19/2007    17:37:36 PM    from John Day-Richter to All Participants:
	Anyway, I think it's time to start...

07/19/2007    17:38:02 PM    from Midori Harris to All Participants:
	and  Fahrenheit 451 = 232.8 °C
07/19/2007    17:38:35 PM    from Karen Christie to All Participants:
	the things that get stuck in Midori's brain constantly amaze me ;)
07/19/2007    17:38:48 PM    from John Day-Richter to All Participants:
	(A sci-fi joke and a units conversion joke - we should play Magic the Gathering too)

07/19/2007    17:38:50 PM    from John Day-Richter to All Participants:
	Anyway -

07/19/2007    17:38:54 PM    from John Day-Richter to All Participants:
	I've been trying to slowly plow my way through all the feature requests and bug reports.

07/19/2007    17:39:16 PM    from John Day-Richter to All Participants:
	I plan to send out a "rank these features" email later this week, but I wanted to put one on the table right now...

07/19/2007    17:39:48 PM    from John Day-Richter to All Participants:
	David Hill (I think) wants a view that shows all the definitions of all the is_a parents of the currently selected term.

07/19/2007    17:40:12 PM    from Jen Deegan to All Participants:
	I can see where he's coming from.

07/19/2007    17:40:14 PM    from Chris Mungall to All Attendees:
	David S
07/19/2007    17:40:26 PM    from Midori Harris to All Participants:
	Being able to pull up definitions for displayed terms generally sounds really useful.
07/19/2007    17:40:32 PM    from John Day-Richter to All Participants:
	I think I can implement this in the graphical dag view (maybe a drop down that lets you choose whether to show term names, definitions, or whatever in the graph)

07/19/2007    17:40:47 PM    from Jen Deegan to All Participants:
	Is this for aristotelian defs?

07/19/2007    17:40:53 PM    from John Day-Richter to All Participants:
	No, this is text defs.

07/19/2007    17:40:58 PM    from Jen Deegan to All Participants:
	yes

07/19/2007    17:41:10 PM    from Jen Deegan to All Participants:
	I mean if you want to make a def by starting with the is_a parent

07/19/2007    17:41:20 PM    from Jen Deegan to All Participants:
	it would be handy to be able to see all of them

07/19/2007    17:41:22 PM    from Chris Mungall to All Attendees:
	Aristotelian defs can be expressed as text or computable xp defs
07/19/2007    17:41:32 PM    from John Day-Richter to All Participants:
	Yeah, I think that's what David is getting at.

07/19/2007    17:41:32 PM    from Midori Harris to All Participants:
	But one of the thinks it will help with is making sure the text defs are Aristotelian, or at least Aristotle-friendly.
07/19/2007    17:41:43 PM    from Jen Deegan to All Participants:
	yes

07/19/2007    17:41:50 PM    from Chris Mungall to All Attendees:
	It's just a fancy way of saying the def can be written in the form "An X that Y", where X is already defined
07/19/2007    17:41:56 PM    from Jen Deegan to All Participants:
	yes

07/19/2007    17:42:02 PM    from John Day-Richter to All Participants:
	So does this sound like something I should prioritize? I would probably build it as the first OBO-Edit 2.0 plugin, as an excuse to finally nail down the new plugin system.

07/19/2007    17:42:03 PM    from Midori Harris to All Participants:
	(once again someone says essentially the same thing in the time I take to type ...)
07/19/2007    17:42:23 PM    from John Day-Richter to All Participants:
	I think whoever Jen is talking to doesn't have "all participants" set. I just see her crying "Yes! Yes!"

07/19/2007    17:42:25 PM    from Jen Deegan to All Participants:
	I'd be keen to see the list of other positbilities before deciding

07/19/2007    17:42:28 PM    from Midori Harris to All Participants:
	Anyway, yes, I would vote for this one.
07/19/2007    17:42:34 PM    from Jen Deegan to All Participants:
	it's chris

07/19/2007    17:42:58 PM    from Jen Deegan to All Participants:
	chris, set to all participants.

07/19/2007    17:43:01 PM    from John Day-Richter to All Participants:
	Chris - set "send to" to "All participants" so I can see what you're saying.

07/19/2007    17:43:09 PM    from Midori Harris to All Participants:
	He's been sending little explanatory comments about defs (things John already knows).
07/19/2007    17:43:21 PM    from Karen Christie to All Participants:
	I'd also like to see the whole list before setting priorities
07/19/2007    17:43:35 PM    from Midori Harris to All Participants:
	Jen's a very positive person. ;)
07/19/2007    17:43:35 PM    from Chris Mungall to All Participants:
	Personally I think it would be useful to have the normal Ontology Editor Panel optionally show additional information, such as for example text defs or xp defs
07/19/2007    17:43:40 PM    from Jen Deegan to All Participants:
	Also we already have an awful lot of new feature in 2.000

07/19/2007    17:43:50 PM    from John Day-Richter to All Participants:
	Chris - how would that work?

07/19/2007    17:44:00 PM    from John Day-Richter to All Participants:
	In a tooltip or something?

07/19/2007    17:44:11 PM    from John Day-Richter to All Participants:
	Or merged into the view? (which can be done)

07/19/2007    17:44:18 PM    from nicole to All Participants:
	optional expand/collapse?
07/19/2007    17:44:20 PM    from Jen Deegan to All Participants:
	It would be hard to cut and paste from a tooltip

07/19/2007    17:44:37 PM    from Jen Deegan to All Participants:
	I have a thought

07/19/2007    17:44:52 PM    from Jen Deegan to All Participants:
	What if there was a button that says 'copy def from is_a parents.'

07/19/2007    17:45:05 PM    from Jen Deegan to All Participants:
	then you just edit right there in the def box

07/19/2007    17:45:19 PM    from Chris Mungall to All Participants:
	As well as the default:
aaa
+bbb
 +ccc
you could choose to see:
aaa : "blah"
+bbb : "foo"
 +ccc : "wibble"
07/19/2007    17:45:32 PM    from Jen Deegan to All Participants:
	:-)

07/19/2007    17:45:36 PM    from John Day-Richter to All Participants:
	Jen - Does it give you a newline separated list of all the is_a parent defs?

07/19/2007    17:45:48 PM    from Jen Deegan to All Participants:
	maybe two newlines so we get ablank space

07/19/2007    17:45:52 PM    from Jen Deegan to All Participants:
	yes

07/19/2007    17:45:58 PM    from Jen Deegan to All Participants:
	that would rock

07/19/2007    17:46:11 PM    from Jen Deegan to All Participants:
	really a  lot

07/19/2007    17:46:27 PM    from Jen Deegan to All Participants:
	currently we have to footer around finding the is_A parent and it would save all that effort

07/19/2007    17:46:28 PM    from John Day-Richter to All Participants:
	Chris - Can you add this as a feature request? I've had a general idea along these lines for quite a while...

07/19/2007    17:46:36 PM    from Midori Harris to All Participants:
	It would also be convenient to have a way to copy the def dbxrefs.
07/19/2007    17:46:57 PM    from Jen Deegan to All Participants:
	this is turning into the clone button

07/19/2007    17:47:04 PM    from Jen Deegan to All Participants:
	which is almost always how I do it anyway

07/19/2007    17:47:14 PM    from John Day-Richter to All Participants:
	This copy+paste thing would be really easy to add. Are there any votes for just going ahead with it, or should we wait to set priorities until I send out the big list?

07/19/2007    17:47:31 PM    from Karen Christie to All Participants:
	Let's see the whole list before setting priorities
07/19/2007    17:47:35 PM    from Jen Deegan to All Participants:
	yes

07/19/2007    17:47:54 PM    from Midori Harris to All Participants:
	The differences I see are (a) clone copies all parents and children, which isn't always desirable, and (b) what if the term already exists, but it's undefined?
07/19/2007    17:48:19 PM    from Jen Deegan to All Participants:
	true

07/19/2007    17:48:28 PM    from John Day-Richter to All Participants:
	Midori - I think I'd implement this as a right-click menu for every text editing control.

07/19/2007    17:48:31 PM    from Jen Deegan to All Participants:
	also copying def from multiple is_a parents

07/19/2007    17:48:35 PM    from Midori Harris to All Participants:
	I second (or third) the look-at-the-feature-request-list motion.
07/19/2007    17:48:45 PM    from Midori Harris to All Participants:
	John - that sounds fine.
07/19/2007    17:48:48 PM    from Jen Deegan to All Participants:
	yes

07/19/2007    17:48:49 PM    from John Day-Richter to All Participants:
	You could choose - "use value from parent" or "use all values from is_a ancestors"

07/19/2007    17:48:59 PM    from Jen Deegan to All Participants:
	that would be great

07/19/2007    17:49:02 PM    from Midori Harris to All Participants:
	oooh, options :)
07/19/2007    17:49:04 PM    from John Day-Richter to All Participants:
	(That second one wouldn't be available for all fields)

07/19/2007    17:49:37 PM    from John Day-Richter to All Participants:
	Okay - I'll work on a feature request summary this afternoon.

07/19/2007    17:49:38 PM    from Jen Deegan to All Participants:
	If there's a button it should have picture of a tiny bearded gentleman on it.

07/19/2007    17:49:55 PM    from John Day-Richter to All Participants:
	Good idea - we know how I love inscrutable, poorly drawn icons.

07/19/2007    17:51:24 PM    from Jen Deegan to All Participants:
	What's next?

07/19/2007    17:51:32 PM    from John Day-Richter to All Participants:
	Progress report: the only new OBO-Edit feature in the new release is the mapquesty zoom-level control.

07/19/2007    17:51:50 PM    from John Day-Richter to All Participants:
	It turned out to be much harder to do than I thought, but it's part of today's demo now.

07/19/2007    17:51:57 PM    from Jen Deegan to All Participants:
	great :-)

07/19/2007    17:52:05 PM    from Midori Harris to All Participants:
	Update to long-standing feature request: Any hope of keyboard shortcuts to use with the graph editor?
07/19/2007    17:52:44 PM    from John Day-Richter to All Participants:
	Absolutely. I rank that as a high priority as well.

07/19/2007    17:52:44 PM    from Midori Harris to All Participants:
	(I spent the 15-20 min before chat struggling with the "show ancestors" menus ... and I think it's not working!
07/19/2007    17:52:57 PM    from Midori Harris to All Participants:
	great!!
07/19/2007    17:53:02 PM    from John Day-Richter to All Participants:
	It only works if the reasoner is on, but it certainly works for me.

07/19/2007    17:53:19 PM    from John Day-Richter to All Participants:
	I think that right-click on macs works poorly in the graph editor. I'm working on a keyboard-only alternative.

07/19/2007    17:53:40 PM    from Midori Harris to All Participants:
	ah ... I think the reasoner is off, cos I forgot that the show-ancestors bit needs it. still fiddly to use the menus, tho.
07/19/2007    17:53:47 PM    from nicole to All Participants:
	is that right-click with a proper 2-button mouse, or with the ctrl-click?
07/19/2007    17:54:05 PM    from John Day-Richter to All Participants:
	Nicole - ctrl+click. It always confuses Java.

07/19/2007    17:54:32 PM    from John Day-Richter to All Participants:
	Any other business to chat about, or should we get on with the continuing graph editor demos?

07/19/2007    17:54:42 PM    from Midori Harris to All Participants:
	Demo!
07/19/2007    17:54:47 PM    from Jane Lomax to All Participants:
	can i ask  quick q?
07/19/2007    17:54:48 PM    from Erika to All Participants:
	I am ready for demo

07/19/2007    17:55:07 PM    from John Day-Richter to All Participants:
	Please do, Jane.

07/19/2007    17:55:25 PM    from Jane Lomax to All Participants:
	there are some xrefs with whitespaces
07/19/2007    17:55:40 PM    from Jane Lomax to All Participants:
	that keep getting flagged by the verification plugin
07/19/2007    17:55:54 PM    from Jane Lomax to All Participants:
	i tried changing the spaces to %20
07/19/2007    17:56:19 PM    from Jane Lomax to All Participants:
	but the checker still said these were non-URI characters
07/19/2007    17:56:32 PM    from Jane Lomax to All Participants:
	so what can I change the whitespaces for?
07/19/2007    17:56:48 PM    from Midori Harris to All Participants:
	We figured out that they were colons in the source ... but that doesn't answer the question of how to make them verification-friendly.
07/19/2007    17:57:00 PM    from John Day-Richter to All Participants:
	That a bug. %20 should be fine.

07/19/2007    17:57:06 PM    from John Day-Richter to All Participants:
	I'll check this out.

07/19/2007    17:57:10 PM    from Midori Harris to All Participants:
	What about colons?
07/19/2007    17:57:42 PM    from John Day-Richter to All Participants:
	You could use the url escape sequence for that too...

07/19/2007    17:57:54 PM    from John Day-Richter to All Participants:
	I don't know what it is off the top of my head.

07/19/2007    17:58:01 PM    from Midori Harris to All Participants:
	we can look it up
07/19/2007    17:58:15 PM    from Jen Deegan to All Participants:
	John, Did you get my OBO-Edit utility write-ups? I e-mailed them.

07/19/2007    17:58:34 PM    from Jane Lomax to All Participants:
	okay - I'll add a bug report for %20
07/19/2007    17:58:35 PM    from John Day-Richter to All Participants:
	Jen - I did, but I didn't have time to do anything with them. I'll write a note to myself right now.

07/19/2007    17:58:44 PM    from Jen Deegan to All Participants:
	cool, just wondered.

07/19/2007    17:59:17 PM    from John Day-Richter to All Participants:
	I made a post-it for you, Jen.

07/19/2007    17:59:25 PM    from John Day-Richter to All Participants:
	Okay - demo time.

07/19/2007    17:59:38 PM    from John Day-Richter to All Participants:
	You can see that I've got the GO loaded right now.

07/19/2007    18:00:05 PM    from John Day-Richter to All Participants:
	I've decided that I only want to work with the component ontology today, so I'll use the right-click menu to hide everything else.

07/19/2007    18:00:46 PM    from John Day-Richter to All Participants:
	Notice the new zoom control on the left.

07/19/2007    18:00:54 PM    from nicole to All Participants:
	nice
07/19/2007    18:00:55 PM    from John Day-Richter to All Participants:
	(I can put it somewhere else if you don't like it there)

07/19/2007    18:00:58 PM    from John Day-Richter to All Participants:
	(btw)

07/19/2007    18:00:59 PM    from Erika to All Participants:
	Yes, it is fantastic

07/19/2007    18:01:07 PM    from John Day-Richter to All Participants:
	It's transparent.

07/19/2007    18:01:20 PM    from Midori Harris to All Participants:
	it's fine there ... that google meme
07/19/2007    18:01:36 PM    from John Day-Richter to All Participants:
	It shows the current zoom level, and it tracks zooms you do with the right mouse button...

07/19/2007    18:01:54 PM    from John Day-Richter to All Participants:
	Notice that it won't let me zoom out very far right now.

07/19/2007    18:01:54 PM    from nicole to All Participants:
	(does it work with the mouse wheel?)
07/19/2007    18:02:12 PM    from John Day-Richter to All Participants:
	(nicole - it could, I haven't enabled that)

07/19/2007    18:02:20 PM    from John Day-Richter to All Participants:
	(but that's a good idea)

07/19/2007    18:02:43 PM    from John Day-Richter to All Participants:
	Right now I'm zoomed all the way out. Look what happens when I expand some more children...

07/19/2007    18:03:10 PM    from John Day-Richter to All Participants:
	I didn't zoom at all, but the zoom control changed position to reflect the new maximum zoom.

07/19/2007    18:03:33 PM    from nicole to All Participants:
	(is there a shortcut button to make them all fit on the screen?)
07/19/2007    18:03:56 PM    from John Day-Richter to All Participants:
	(you can press the V key to cycle through the bounds guarantors and choose Zoom to All)

07/19/2007    18:04:08 PM    from nicole to All Participants:
	right right
07/19/2007    18:04:28 PM    from John Day-Richter to All Participants:
	You can change the zoom level with the zoom control by dragging...

07/19/2007    18:04:35 PM    from Karen Christie to All Participants:
	hmm, seems like you have it reversed from Google, they have most zoomed out on the bottom and most zoomed in on the top, it'd be nice if it was the same
07/19/2007    18:05:11 PM    from John Day-Richter to All Participants:
	Oh, I can flip that over. My thought was down=deeper, but I should use the common idiom.

07/19/2007    18:05:28 PM    from melissa to All Participants:
	john, is your zoom in the beta9 release? how do I make it appear?

07/19/2007    18:05:36 PM    from nicole to All Participants:
	maybe a + or - to refresh our memory
07/19/2007    18:05:50 PM    from Midori Harris to All Participants:
	Mel - I second that!
07/19/2007    18:05:56 PM    from John Day-Richter to All Participants:
	Melissa, it's only available in the Graph Editor right now. It should appear by default.

07/19/2007    18:06:14 PM    from John Day-Richter to All Participants:
	You can also zoom by clicking on the control.

07/19/2007    18:06:19 PM    from Midori Harris to All Participants:
	I have it, but it's fainter than on John's machine.
07/19/2007    18:06:22 PM    from melissa to All Participants:
	ok, got it

07/19/2007    18:06:35 PM    from Jen Deegan to All Participants:
	In windows where I'm running it the dot on the scale can go below the end of the scale which is actually quite cool

07/19/2007    18:06:45 PM    from John Day-Richter to All Participants:
	Jen - that's bad.

07/19/2007    18:06:51 PM    from melissa to All Participants:
	yeah me too.

07/19/2007    18:06:53 PM    from Erika to All Participants:
	for me too!

07/19/2007    18:06:57 PM    from Jen Deegan to All Participants:
	It's kind of nice actually

07/19/2007    18:07:12 PM    from John Day-Richter to All Participants:
	How are you getting it to do that?

07/19/2007    18:07:25 PM    from Jen Deegan to All Participants:
	I just click on the scale and it goes right down off the end

07/19/2007    18:07:30 PM    from Midori Harris to All Participants:
	On my Mac the dot is hardly even visible -- blends in with the rest of the scale.
07/19/2007    18:08:12 PM    from John Day-Richter to All Participants:
	Huh. When I get my new mac, these problems will all be addressed.

07/19/2007    18:08:22 PM    from John Day-Richter to All Participants:
	Maybe I need to turn off the transparency for now.

07/19/2007    18:08:38 PM    from nicole to All Participants:
	i think that's ok
07/19/2007    18:08:54 PM    from Jen Deegan to All Participants:
	The scale is good and makes zooming much more controllable

07/19/2007    18:09:13 PM    from nicole to All Participants:
	yes, i really like the zooming -scale thingy
07/19/2007    18:09:19 PM    from Jen Deegan to All Participants:
	Also means you can zoom in steps rather than smoothly

07/19/2007    18:09:30 PM    from John Day-Richter to All Participants:
	Now that I've figured out how to add things to the view that don't move around, I was thinking of adding some more stuff...

07/19/2007    18:09:57 PM    from John Day-Richter to All Participants:
	Would people like a little overview box that always shows the fully-zoomed out graph, with a little rectangle that shows what you're looking at?

07/19/2007    18:10:11 PM    from Jen Deegan to All Participants:
	yes

07/19/2007    18:10:15 PM    from nicole to All Participants:
	wowzers!
07/19/2007    18:10:20 PM    from melissa to All Participants:
	yeah!

07/19/2007    18:10:23 PM    from Jen Deegan to All Participants:
	And an arrow that points north :-)

07/19/2007    18:10:27 PM    from Jen Deegan to All Participants:
	not really

07/19/2007    18:10:32 PM    from Midori Harris to All Participants:
	I think I'd like to have it as an option.
07/19/2007    18:10:44 PM    from John Day-Richter to All Participants:
	I'll add that to feature requests, then.

07/19/2007    18:10:50 PM    from nicole to All Participants:
	yeah, option good.  don't always need it for small graphs
07/19/2007    18:11:01 PM    from Midori Harris to All Participants:
	(the overview, I mean ... tho I'm not opposed to north)
07/19/2007    18:11:24 PM    from John Day-Richter to All Participants:
	Excellent.

07/19/2007    18:11:58 PM    from John Day-Richter to All Participants:
	Melissa said that the graph editor interface still wasn't making sense to her, so I wanted to talk about the concepts a little more.

07/19/2007    18:12:13 PM    from Jen Deegan to All Participants:
	could you remind me how to turn off animation? I lost the button

07/19/2007    18:12:43 PM    from John Day-Richter to All Participants:
	Click the little wrench, and you'll get a config screen

07/19/2007    18:12:54 PM    from Jen Deegan to All Participants:
	got it, thanks

07/19/2007    18:13:13 PM    from John Day-Richter to All Participants:
	Oh, while I set up the example...

07/19/2007    18:13:26 PM    from John Day-Richter to All Participants:
	See how every time I expand a node, everything shrinks automatically?

07/19/2007    18:13:49 PM    from John Day-Richter to All Participants:
	(Yikes, I need to deal with these long names)

07/19/2007    18:13:57 PM    from nicole to All Participants:
	(it would be great to have some hover-help capability)
07/19/2007    18:14:08 PM    from Jen Deegan to All Participants:
	we are getting some awfully long names these days

07/19/2007    18:14:12 PM    from Karen Christie to All Participants:
	I was just going to say something about those long names not wrapping to stay in their boxes
07/19/2007    18:14:15 PM    from John Day-Richter to All Participants:
	Everything expands automatically because I've got the "Zoom to all" view mode set.

07/19/2007    18:14:20 PM    from nicole to All Participants:
	(like if i have turned help on, if i hover over something, it tells me what would happen if i click)
07/19/2007    18:14:32 PM    from John Day-Richter to All Participants:
	(Nicole - it usually does that)

07/19/2007    18:14:37 PM    from nicole to All Participants:
	oh!
07/19/2007    18:14:50 PM    from nicole to All Participants:
	never mind.  my bad
07/19/2007    18:15:12 PM    from Midori Harris to All Participants:
	I noticed the long names too ...
07/19/2007    18:15:53 PM    from John Day-Richter to All Participants:
	(Minor problem here, need to restart - not a bug, an as yet unimplemented feature)

07/19/2007    18:16:26 PM    from John Day-Richter to All Participants:
	Anyway, here's one of the concepts that I think might be confusing at first...

07/19/2007    18:16:52 PM    from John Day-Richter to All Participants:
	In the graph editor, you only control which terms are visible. You have no control over which links are visible.

07/19/2007    18:17:17 PM    from John Day-Richter to All Participants:
	OBO-Edit automatically draws *every* true relationship between the currently visible terms, and you can't stop it.

07/19/2007    18:17:33 PM    from John Day-Richter to All Participants:
	That's why sometimes you see links that you haven't expanded.

07/19/2007    18:17:50 PM    from melissa to All Participants:
	that seems ok, as long as you could limit it to a certain link type

07/19/2007    18:18:12 PM    from John Day-Richter to All Participants:
	You will be able to, when I add filtering to the graph editor (which is easy, I just haven't done it yet)

07/19/2007    18:18:30 PM    from melissa to All Participants:
	excellent

07/19/2007    18:19:15 PM    from John Day-Richter to All Participants:
	This feature causes some confusion though, especially in the graph dag viewer.

07/19/2007    18:19:15 PM    from Jen Deegan to All Participants:
	would it be possible to make it so that when you open a new link in the graph editor the graph doesn't move?

07/19/2007    18:19:38 PM    from Jen Deegan to All Participants:
	as in the old text graph view

07/19/2007    18:19:54 PM    from John Day-Richter to All Participants:
	Jen - it moves because it's trying to put the term you clicked in the middle of the screen.

07/19/2007    18:20:09 PM    from John Day-Richter to All Participants:
	Is this bothering other people too, or should this only be disabled when animations are turned off?

07/19/2007    18:20:14 PM    from Jen Deegan to All Participants:
	yes, I understand but I think it would be more intuitive not to

07/19/2007    18:20:20 PM    from Jen Deegan to All Participants:
	good point

07/19/2007    18:20:53 PM    from Midori Harris to All Participants:
	I'd have to do more testing to form an opinion.
07/19/2007    18:20:54 PM    from Jen Deegan to All Participants:
	partly it's very noticeable just now as the bubbles are so huge

07/19/2007    18:20:59 PM    from Karen Christie to All Participants:
	maybe the center on selected term can be an optional behavior, I can see times where I'd want it and times when I would not
07/19/2007    18:21:12 PM    from Jen Deegan to All Participants:
	yes that's true

07/19/2007    18:21:27 PM    from Jen Deegan to All Participants:
	I guess we are feeling our way with this.

07/19/2007    18:21:41 PM    from John Day-Richter to All Participants:
	Hmm... let's leave it on for now, and I'll disable it in "no animations" mode. If people start to get bothered in regular mode, we'll add a control for it.

07/19/2007    18:21:50 PM    from Jen Deegan to All Participants:
	thanks

07/19/2007    18:22:05 PM    from John Day-Richter to All Participants:
	(I'm trying to keep this thing as non-configurable as possible, since it's already so complex)

07/19/2007    18:22:10 PM    from Jen Deegan to All Participants:
	can you also have the speech bubble just appear in no animation mode?

07/19/2007    18:22:24 PM    from John Day-Richter to All Participants:
	What's it doing now?

07/19/2007    18:22:34 PM    from Jen Deegan to All Participants:
	it slides across the screen

07/19/2007    18:22:39 PM    from Jen Deegan to All Participants:
	in windows

07/19/2007    18:22:51 PM    from John Day-Richter to All Participants:
	Oops. I thought I fixed that. I'll look at it again.

07/19/2007    18:22:54 PM    from Jen Deegan to All Participants:
	ta :-)

07/19/2007    18:23:42 PM    from John Day-Richter to All Participants:
	Anyway, the fact that OBO-Edit always shows all links in the graph viewer can confuse people in the graph dag viewer.

07/19/2007    18:24:25 PM    from John Day-Richter to All Participants:
	For example, in the part_of dag view, we're seeing a lot of is_a links.

07/19/2007    18:24:36 PM    from melissa to All Participants:
	yes that is confusing.

07/19/2007    18:24:45 PM    from John Day-Richter to All Participants:
	But it's actually what you want, I think.

07/19/2007    18:24:58 PM    from melissa to All Participants:
	why?

07/19/2007    18:25:06 PM    from John Day-Richter to All Participants:
	OBO-Edit draws these dag views by saying "what are all the true part_of parents of cell envelope".

07/19/2007    18:25:15 PM    from John Day-Richter to All Participants:
	Then it shows those nodes.

07/19/2007    18:25:24 PM    from John Day-Richter to All Participants:
	Then it automatically draws all the links between them.

07/19/2007    18:25:29 PM    from Jen Deegan to All Participants:
	also because we're not part_of complete?

07/19/2007    18:25:58 PM    from John Day-Richter to All Participants:
	The additional information is useful, because it shows why cell envelope is part_of cellular component. The transitivity is right in front of you.

07/19/2007    18:26:04 PM    from nicole to All Participants:
	so why then display implied is_a links in part_of.  why not only direct is_a links?
07/19/2007    18:26:18 PM    from Midori Harris to All Participants:
	Aside: the graph DAG viewer is taking a long time to show up (it appears eventually, but it's not prompt).
07/19/2007    18:26:53 PM    from Jane Lomax to All Participants:
	it's showing the wrong relation tho - cell part is_a cellulaR COMPONENT?
07/19/2007    18:27:02 PM    from John Day-Richter to All Participants:
	(Midori - I plan to thread this so that the application doesn't freeze while the dag view is being calculated)

07/19/2007    18:27:12 PM    from John Day-Richter to All Participants:
	Jane - that's what it says... let's try to figure out why.

07/19/2007    18:27:18 PM    from Midori Harris to All Participants:
	ok ... that'll help
07/19/2007    18:27:52 PM    from John Day-Richter to All Participants:
	Did everyone see that?

07/19/2007    18:27:55 PM    from Midori Harris to All Participants:
	yes
07/19/2007    18:27:59 PM    from Erika to All Participants:
	yes

07/19/2007    18:28:17 PM    from Jen Deegan to All Participants:
	see what?

07/19/2007    18:28:21 PM    from John Day-Richter to All Participants:
	(I may have an out-of-date version of GO, by the way)

07/19/2007    18:28:37 PM    from John Day-Richter to All Participants:
	I'll do it again - when I hover over the suspicious is_a link, I get an explanation...

07/19/2007    18:28:39 PM    from Karen Christie to All Participants:
	bye, gotta go, have a 10:30 meeting to attend here, till next time :)
07/19/2007    18:28:50 PM    from Jane Lomax to All Participants:
	oh right - cell part part_of cell and cell is_a cellular component - i see...
07/19/2007    18:28:55 PM    from Jen Deegan to All Participants:
	yes I see that

07/19/2007    18:29:10 PM    from Jen Deegan to All Participants:
	nice

07/19/2007    18:29:40 PM    from John Day-Richter to All Participants:
	Anyway, that's part of why I think it's a good idea to include all the links in the dag view - it lets you see and ferret out suspicious links.

07/19/2007    18:30:06 PM    from melissa to All Participants:
	ok, so why is cell part not is_a cellular component? is this suspicious?

07/19/2007    18:30:14 PM    from nicole to All Participants:
	(thanks melissa)
07/19/2007    18:30:19 PM    from Midori Harris to All Participants:
	not to me
07/19/2007    18:30:44 PM    from Jane Lomax to All Participants:
	it should be both is_a and part_of cell component
07/19/2007    18:30:55 PM    from John Day-Richter to All Participants:
	Cell part is_a cellular component. But the two links are being drawn right on top of each other, so it's hard to see the is_a link.

07/19/2007    18:31:02 PM    from Midori Harris to All Participants:
	oops, sorry, ignore my last comment ...
07/19/2007    18:31:03 PM    from John Day-Richter to All Participants:
	I'm trying to figure out how to fix that problem.

07/19/2007    18:31:27 PM    from melissa to All Participants:
	I still would rather see only is_a links in the is_a graph I think.

07/19/2007    18:31:49 PM    from Jane Lomax to All Participants:
	me too - i think
07/19/2007    18:32:01 PM    from melissa to All Participants:
	and what happens to orphans? can they all go in a box somewhere?

07/19/2007    18:32:02 PM    from Midori Harris to All Participants:
	I think I agree, at least for ontologies that are is_a complete
07/19/2007    18:32:28 PM    from melissa to All Participants:
	eg. how can we make it easier to make ontologies is_a complete?

07/19/2007    18:32:31 PM    from nicole to All Participants:
	so for cell envelop, you'd only see the implied link to cellular_component then, right?
07/19/2007    18:32:31 PM    from John Day-Richter to All Participants:
	Perhaps I need to add a checkbox "Only show current link type in panels"

07/19/2007    18:32:50 PM    from nicole to All Participants:
	then if you clicked on the implied link you could expand to see how that came about?
07/19/2007    18:33:03 PM    from nicole to All Participants:
	which includes a part_of link
07/19/2007    18:33:03 PM    from Jane Lomax to All Participants:
	that would work
07/19/2007    18:33:11 PM    from melissa to All Participants:
	I still don't get why the part_of link is showing up in the is_a panel. I do understand the reverse in the the part_of panel.

07/19/2007    18:33:31 PM    from Midori Harris to All Participants:
	seconding melissa again!
07/19/2007    18:33:32 PM    from John Day-Richter to All Participants:
	Nicole - I think better to show you every is_a ancestor still, but it would only show is_a links in the graph.

07/19/2007    18:34:03 PM    from John Day-Richter to All Participants:
	Melissa - because it's showing you EVERY TRUE LINK. It only figured out what terms to show, the links are being drawn automatically after that.

07/19/2007    18:34:31 PM    from Midori Harris to All Participants:
	Wait, I'm more confused now.
07/19/2007    18:34:40 PM    from nicole to All Participants:
	John - would you then still show cell, but no link?  that would be odd.
07/19/2007    18:34:48 PM    from melissa to All Participants:
	but I don't want to see EVERY true link, I only want to see is_a links. whine whine.

07/19/2007    18:35:18 PM    from nicole to All Participants:
	melissa - what would you want to see for the cell envelope graph?
07/19/2007    18:35:23 PM    from Karen Christie to All Participants:
	back due to room conflict issues, maybe the links of the other type could be grayed out so they're less prominant
07/19/2007    18:35:24 PM    from Jen Deegan to All Participants:
	what if there were a whole series of checkboxes to let you choose?

07/19/2007    18:35:34 PM    from John Day-Richter to All Participants:
	Jen - choose what?

07/19/2007    18:35:47 PM    from Jen Deegan to All Participants:
	what to see

07/19/2007    18:35:50 PM    from Midori Harris to All Participants:
	Maybe what I'm missing is how something gets to be an is_a ancestor other than by is_a relationships ...
07/19/2007    18:36:00 PM    from melissa to All Participants:
	yes!!!

07/19/2007    18:36:11 PM    from John Day-Richter to All Participants:
	Midori - that's the only way it happens.

07/19/2007    18:36:24 PM    from John Day-Richter to All Participants:
	What are you seeing that makes you think is_a relationships are being added some other way?

07/19/2007    18:36:26 PM    from Midori Harris to All Participants:
	That's what I thought.
07/19/2007    18:37:21 PM    from Midori Harris to All Participants:
	Um, the part_of relationship in the is_a panel ... plus your comment about "every true link" and choosing terms first. Taht one really threw me.
07/19/2007    18:37:40 PM    from Jen Deegan to All Participants:
	could you have checkboxes: 1) show is_a links 2) show part_of links 3) show implied part_of links

07/19/2007    18:37:54 PM    from nicole to All Participants:
	yes jen!
07/19/2007    18:38:41 PM    from John Day-Richter to All Participants:
	Jen - we could, but I think maybe a simpler solution would be to only show the link type of interest (ie only is_as in the is_a panel), plus other links types if and only if they are needed to support the link type of interest.

07/19/2007    18:38:54 PM    from John Day-Richter to All Participants:
	This change would remove the following links from what we're seeing now:

07/19/2007    18:38:56 PM    from melissa to All Participants:
	i like that idea

07/19/2007    18:39:01 PM    from Jen Deegan to All Participants:
	me too

07/19/2007    18:39:04 PM    from John Day-Richter to All Participants:
	cell part -part_of-> cellular_component

07/19/2007    18:39:15 PM    from John Day-Richter to All Participants:
	cell envelope -is_a-> cellular_component

07/19/2007    18:39:24 PM    from Midori Harris to All Participants:
	I think that's more like what I would expect (if I 'd consciously formed expectations in the first place).
07/19/2007    18:39:25 PM    from nicole to All Participants:
	sounds good
07/19/2007    18:39:28 PM    from Jen Deegan to All Participants:
	me too

07/19/2007    18:39:50 PM    from Jen Deegan to All Participants:
	that's what 1.101 does isn'tit?

07/19/2007    18:40:19 PM    from John Day-Richter to All Participants:
	Not really - 1.101 does a graph-walking thing that doesn't use the reasoner at all.

07/19/2007    18:40:26 PM    from Midori Harris to All Participants:
	the infamous "flower button", you mean?
07/19/2007    18:40:30 PM    from Jen Deegan to All Participants:
	yes

07/19/2007    18:40:31 PM    from melissa to All Participants:
	yeah that seems more intuitive

07/19/2007    18:40:36 PM    from John Day-Richter to All Participants:
	I'm going to make this change right away, because I think it's stopping people from effectively testing this component.

07/19/2007    18:40:53 PM    from John Day-Richter to All Participants:
	One last question before we break:

07/19/2007    18:41:12 PM    from John Day-Richter to All Participants:
	It's clear that OBO-Edit's choice to show all true links was confusing in the Graph DAG Viewer.

07/19/2007    18:41:18 PM    from John Day-Richter to All Participants:
	Is it confusing in the Graph Editor?

07/19/2007    18:41:30 PM    from Jen Deegan to All Participants:
	yes

07/19/2007    18:41:55 PM    from Jane Lomax to All Participants:
	no - I'd expect to see all relations there I think...
07/19/2007    18:42:01 PM    from John Day-Richter to All Participants:
	So it's not clear why OBO-Edit is showing a link between "organelle envelope" and "cell part", Jen?

07/19/2007    18:42:11 PM    from Jen Deegan to All Participants:
	though I guess if we just resolved all the implied links it would solve the problem

07/19/2007    18:42:22 PM    from melissa to All Participants:
	no because the graph editor isn't filtered for a specfic relationship type

07/19/2007    18:42:22 PM    from Jen Deegan to All Participants:
	it just seems over busy having it all in there

07/19/2007    18:42:36 PM    from John Day-Richter to All Participants:
	Jen - you can resolve implied links on the right-click menu...

07/19/2007    18:42:44 PM    from Midori Harris to All Participants:
	It bothers me less in the graph editor because the graph editor doesn't lead me to expect not to see part_ofs, wheresa the graph dag view is_a panel does.
07/19/2007    18:42:55 PM    from Midori Harris to All Participants:
	(hope that wasn't too convoluted)
07/19/2007    18:43:05 PM    from Jen Deegan to All Participants:
	are we making a policy decision to resolve such links?

07/19/2007    18:43:08 PM    from Midori Harris to All Participants:
	er, whereas, not wheresa
07/19/2007    18:43:09 PM    from Jen Deegan to All Participants:
	that might make sense

07/19/2007    18:43:21 PM    from John Day-Richter to All Participants:
	Jen - what do you mean by that?

07/19/2007    18:43:24 PM    from melissa to All Participants:
	but I'm wondering if in the long term the editor and viewer won't be the same thing...if you allow limiting to specific relationship types in the editor?

07/19/2007    18:43:48 PM    from Jen Deegan to All Participants:
	I mean if OBO-Edit finds links that should be there why don't I just add them.

07/19/2007    18:43:59 PM    from John Day-Richter to All Participants:
	Aha! That's not what implied links mean!

07/19/2007    18:44:02 PM    from Jen Deegan to All Participants:
	oh

07/19/2007    18:44:06 PM    from Jen Deegan to All Participants:
	ooops

07/19/2007    18:44:21 PM    from John Day-Richter to All Participants:
	They mean that there's some relationship that's implied by existing links (usually by transitivity).

07/19/2007    18:44:21 PM    from Jen Deegan to All Participants:
	what does it mean?

07/19/2007    18:44:33 PM    from Jen Deegan to All Participants:
	but that we don'tneed to add them?

07/19/2007    18:44:47 PM    from John Day-Richter to All Participants:
	No, because they're there already. Let's look at this example...

07/19/2007    18:44:52 PM    from Midori Harris to All Participants:
	So adding them explicitly would be redundant much of the time.
07/19/2007    18:44:53 PM    from John Day-Richter to All Participants:
	There's an implied link between these two terms.

07/19/2007    18:45:01 PM    from John Day-Richter to All Participants:
	Midori - it would be redundant ALL the time.

07/19/2007    18:45:03 PM    from Jen Deegan to All Participants:
	oh I see

07/19/2007    18:45:17 PM    from Jen Deegan to All Participants:
	so shy are we showing them?

07/19/2007    18:45:18 PM    from John Day-Richter to All Participants:
	Watch what happens when I right click the link and choose "Expand all supporting links"

07/19/2007    18:45:20 PM    from Jen Deegan to All Participants:
	why I mean

07/19/2007    18:45:39 PM    from Midori Harris to All Participants:
	(thought so, but couldn't shake the fear of getting it wrong)
07/19/2007    18:45:54 PM    from John Day-Richter to All Participants:
	It showed us the links that were used to figure out the implied link.

07/19/2007    18:46:02 PM    from Jen Deegan to All Participants:
	in that case wouldn't all terms have direct is_a to the root node for example

07/19/2007    18:46:24 PM    from melissa to All Participants:
	so implied is both collapsed and implied through transitivity?

07/19/2007    18:46:27 PM    from Midori Harris to All Participants:
	it could, but they get trimmed
07/19/2007    18:46:31 PM    from John Day-Richter to All Participants:
	Midori is right.

07/19/2007    18:46:56 PM    from Jen Deegan to All Participants:
	this could be a topic for next weekwhen people have had their dinner perhaps?

07/19/2007    18:46:57 PM    from Midori Harris to All Participants:
	(well, that's a refreshing novelty!)
07/19/2007    18:47:15 PM    from John Day-Richter to All Participants:
	Jen - yes. This should be the top priority.

07/19/2007    18:47:22 PM    from Midori Harris to All Participants:
	Jen, you and I will miss next week's fun.
07/19/2007    18:47:26 PM    from Jen Deegan to All Participants:
	oh dear

07/19/2007    18:47:38 PM    from John Day-Richter to All Participants:
	Let's close up shop for today.

07/19/2007    18:47:43 PM    from John Day-Richter to All Participants:
	Melissa, can you stick around?

07/19/2007    18:47:55 PM    from John Day-Richter to All Participants:
	Maybe our discussion can be added to the transcript, and it will help clarify this for everyone.

07/19/2007    18:48:00 PM    from Erika to All Participants:
	Jen, Midori we could try to do it from Italy

07/19/2007    18:48:10 PM    from melissa to All Participants:
	yep

07/19/2007    18:48:11 PM    from Jen Deegan to All Participants:
	I think our brains will be full then

07/19/2007    18:48:18 PM    from Erika to All Participants:
	ok ;-)

07/19/2007    18:48:23 PM    from Jen Deegan to All Participants:
	maybe we will catch up later

07/19/2007    18:48:38 PM    from John Day-Richter to All Participants:
	Maybe Mel and I will figure it all out before next week.

07/19/2007    18:48:41 PM    from Jen Deegan to All Participants:
	great

07/19/2007    18:48:44 PM    from Jen Deegan to All Participants:
	thanks :-)

07/19/2007    18:48:46 PM    from John Day-Richter to All Participants:
	Bye everybody but Melissa.

07/19/2007    18:48:47 PM    from Midori Harris to All Participants:
	Will we be done with the post-meeting event in time? If so, we can try to join. Otherwise, we can read the transcript :)
07/19/2007    18:48:53 PM    from Midori Harris to All Participants:
	bye!
07/19/2007    18:48:53 PM    from melissa to All Participants:
	anyone who wants to can stay...we are just going to try to un-confuse me....

07/19/2007    18:49:00 PM    from Jen Deegan to All Participants:
	bye

07/19/2007    18:49:04 PM    from nicole to All Participants:
	thanks john that's fun
07/19/2007    18:49:18 PM    from Midori Harris to All Participants:
	getting late here, so I'll just be confused for another week or two.
07/19/2007    18:49:28 PM    from Erika to All Participants:
	I shoul go. It is 8 pm here! Time for dinner

07/19/2007    18:49:41 PM    from Erika to All Participants:
	bye

07/19/2007    18:49:47 PM    from melissa to All Participants:
	bye!

07/19/2007    18:51:06 PM    from John Day-Richter to All Participants:
	Okay, Melissa. Do you want to keep talking about this implied links thing, or did you want to discuss something else first?

07/19/2007    18:51:24 PM    from melissa to All Participants:
	sort of.... 

07/19/2007    18:51:48 PM    from melissa to All Participants:
	what you mean by implied links is often one of two things, right? not showing, or implied through transitivity. right?

07/19/2007    18:52:22 PM    from melissa to All Participants:
	so earlier,external encapsulating structure was hidden and you said there was an implied link.

07/19/2007    18:52:28 PM    from John Day-Richter to All Participants:
	 I only mean one thing, but I see why you would think I mean two...

07/19/2007    18:53:03 PM    from John Day-Richter to All Participants:
	I think the confusion comes from my failure to communicate some important things about how the reasoner works, and how that interacts with the display...

07/19/2007    18:53:13 PM    from John Day-Richter to All Participants:
	So lemme try to communicate those things now.

07/19/2007    18:54:34 PM    from John Day-Richter to All Participants:
	When the reasoner runs, one of the things it does is to calculate the transitive closure of the ontology. Like if A-is_a->B and B-is_a->C, the reasoner adds the implied link A-is_a->C to the ontology. If you ask A, who are your is_a parents, A will say "B & C".

07/19/2007    18:55:01 PM    from melissa to All Participants:
	I get that.

07/19/2007    18:55:10 PM    from John Day-Richter to All Participants:
	Now most of the time, you don't want to see that link between A and C. It would clutter the display with thousands and thousands of obvious links.

07/19/2007    18:55:40 PM    from John Day-Richter to All Participants:
	So all the OBO-Edit displays "trim" most implied links from the display...

07/19/2007    18:56:14 PM    from melissa to All Participants:
	except those that are required to form links between all visible terms?

07/19/2007    18:56:23 PM    from John Day-Richter to All Participants:
	OBO-Edit basically says "if all the links that constitute the evidence for an implied link are showing, hide that implied link"

07/19/2007    18:56:43 PM    from John Day-Richter to All Participants:
	Melissa - yes!

07/19/2007    18:56:56 PM    from melissa to All Participants:
	ok, so this makes perfect sense and is what I would expect.

07/19/2007    18:56:57 PM    from John Day-Richter to All Participants:
	But those links won't be trimmed, because their evidence links aren't showing.

07/19/2007    18:57:06 PM    from melissa to All Participants:
	right.

07/19/2007    18:57:15 PM    from John Day-Richter to All Participants:
	Allow me to do a little demo to show how this works.

07/19/2007    18:58:27 PM    from melissa to All Participants:
	ok, so in the current display...

07/19/2007    18:58:38 PM    from John Day-Richter to All Participants:
	Okay, right now there's a part_of link between periplasmic space and cell envelope, and an is_a link between cell envelope and external encapsulating structure

07/19/2007    18:59:18 PM    from John Day-Richter to All Participants:
	That means that the reasoner ALSO contains a part_of link between periplasmic space and external encapsulating structure, but it's being trimmed from the display, because all the supporting links are already showing.

07/19/2007    18:59:42 PM    from John Day-Richter to All Participants:
	But look what happens when I hide "cell envelope", causing the supporting links to disappear...

07/19/2007    19:00:13 PM    from melissa to All Participants:
	ok, this all makes sense, except why is there an implied is_a link showing?

07/19/2007    19:00:34 PM    from John Day-Richter to All Participants:
	There must be some chain of links that imply that periplasmic space has an is_a relationship to cell part.

07/19/2007    19:00:44 PM    from John Day-Richter to All Participants:
	We can ask the graph editor to show us that chain of links.

07/19/2007    19:00:57 PM    from John Day-Richter to All Participants:
	To simplify things, I'm going to hide external encapsulating structure first.

07/19/2007    19:01:18 PM    from melissa to All Participants:
	because cellular component should be is_a complete, there shouldn't be any need for implied is_a links, only collapsed ones.

07/19/2007    19:01:39 PM    from John Day-Richter to All Participants:
	What do you mean? There's no difference between implied and collapsed links.

07/19/2007    19:01:54 PM    from melissa to All Participants:
	thus my confusion, I guess.

07/19/2007    19:02:26 PM    from melissa to All Participants:
	there is a difference bewtween implying a link through transitivity and simply showing or not showing an explicit link. right?

07/19/2007    19:03:01 PM    from John Day-Richter to All Participants:
	There's a big difference - they kind of have nothing to do with each other.

07/19/2007    19:03:12 PM    from melissa to All Participants:
	ok good. (not crazy)

07/19/2007    19:03:26 PM    from melissa to All Participants:
	so go back to what you had before.

07/19/2007    19:03:52 PM    from John Day-Richter to All Participants:
	Let me go back to something similar...

07/19/2007    19:04:04 PM    from melissa to All Participants:
	ok, thats good.

07/19/2007    19:04:23 PM    from John Day-Richter to All Participants:
	Okay. There's an implied is_a link between periplasmic space and cell part.

07/19/2007    19:04:42 PM    from melissa to All Participants:
	so here, it looks like there are implied links between periplasmic space and cell part, rather than just not showing what I know are true explicit part of links.

07/19/2007    19:04:52 PM    from melissa to All Participants:
	I mean is_a sorry

07/19/2007    19:05:35 PM    from John Day-Richter to All Participants:
	There is an implied link that exists because there are real explicit is_a links. When I mouse over the implied link, the reasoner will tell me why it's there.

07/19/2007    19:05:38 PM    from melissa to All Participants:
	do you see my confusion? I know that periplasmic space has is_a links back to the root. 

07/19/2007    19:05:44 PM    from melissa to All Participants:
	maybe show the root.

07/19/2007    19:06:37 PM    from melissa to All Participants:
	so right now, it looks like perisplasmic space has no explicit is_a parent. is this true?

07/19/2007    19:06:56 PM    from John Day-Richter to All Participants:
	No, it does, but it's not currently visible because one or more nodes are hidden.

07/19/2007    19:07:12 PM    from John Day-Richter to All Participants:
	And those nodes are hidden because I (the user) explicitly hid them.

07/19/2007    19:07:14 PM    from melissa to All Participants:
	hidden in the dotted line i link?

07/19/2007    19:07:52 PM    from John Day-Richter to All Participants:
	Just hidden because I said not to show them. The dotted line is there because the reasoner knows that, for some reason (probably transitivity) there's an implied is_a link between periplasmic space and cell part.

07/19/2007    19:08:12 PM    from John Day-Richter to All Participants:
	If we hover the mouse over the implied link, the reasoner will tell us why it's there...

07/19/2007    19:08:42 PM    from melissa to All Participants:
	sorry hang on here

07/19/2007    19:09:20 PM    from melissa to All Participants:
	so why does it not show a solid line between perip space and cell part?

07/19/2007    19:09:35 PM    from melissa to All Participants:
	(am I driving you crazy yet?)

07/19/2007    19:09:43 PM    from John Day-Richter to All Participants:
	No, this is really useful.

07/19/2007    19:09:53 PM    from John Day-Richter to All Participants:
	Dotted lines mean "this link was added by the reasoner, not a human"

07/19/2007    19:10:15 PM    from John Day-Richter to All Participants:
	Which is the meaning of an "implied link". A link that was added by the reasoner, not a person.

07/19/2007    19:10:17 PM    from melissa to All Participants:
	but a human did add a real explicit is_a link between perisplasmic space and cell part, right?

07/19/2007    19:10:41 PM    from John Day-Richter to All Participants:
	No. They created some other set of explicit links.

07/19/2007    19:10:46 PM    from John Day-Richter to All Participants:
	I'll show them now.

07/19/2007    19:11:05 PM    from melissa to All Participants:
	I agree, thats how it should be. but I happen to know that every term in the GO has an is_a parent, so I don't expect to ever see implied is_a links, only hidden ones

07/19/2007    19:11:30 PM    from melissa to All Participants:
	(but then again, I am not used to working with ontologies with expansive multiple inheritance

07/19/2007    19:12:06 PM    from John Day-Richter to All Participants:
	I think there's still a false distinction being made between "implied" and "hidden".

07/19/2007    19:12:18 PM    from John Day-Richter to All Participants:
	Consider the screen right now:

07/19/2007    19:12:29 PM    from melissa to All Participants:
	or mabye what I want is to not show implied is_a links EVER, if the ontology is is_a complete.

07/19/2007    19:12:32 PM    from melissa to All Participants:
	ok

07/19/2007    19:12:34 PM    from John Day-Richter to All Participants:
	Every link on the screen was explicitly created by a person.

07/19/2007    19:12:42 PM    from John Day-Richter to All Participants:
	I'm going to select two nodes to hide.

07/19/2007    19:13:15 PM    from John Day-Richter to All Participants:
	When I hide those nodes, the links that go in and out of those nodes will be hidden too.

07/19/2007    19:14:14 PM    from melissa to All Participants:
	but now, the dotted line represents real explicit links, not ones implied bythe reasoner through transitivity. 

07/19/2007    19:14:15 PM    from melissa to All Participants:
	right?

07/19/2007    19:14:22 PM    from John Day-Richter to All Participants:
	Nope...

07/19/2007    19:14:29 PM    from melissa to All Participants:
	arrgghhh.

07/19/2007    19:14:29 PM    from John Day-Richter to All Participants:
	Once those nodes were hidden, the viewer decided to show the implied is_a link between periplasmic space and cell part.

07/19/2007    19:14:46 PM    from John Day-Richter to All Participants:
	That implied link was always there - it was just being hidden before to keep the display clean

07/19/2007    19:15:02 PM    from melissa to All Participants:
	so because the immediate is_a parent is hidden, the reasoner draws it as an implied link.

07/19/2007    19:15:56 PM    from John Day-Richter to All Participants:
	There's a subtle problem with that statement, and it captures the whole miscommunication, I think.

07/19/2007    19:16:14 PM    from melissa to All Participants:
	I guess what I am now confused by is the difference between reasoner behavior in the editor - showing or not showing links based on what is there

07/19/2007    19:16:35 PM    from John Day-Richter to All Participants:
	You said "because the immediate is_a parent is hidden, the reasoner draws it as an implied link"

07/19/2007    19:16:37 PM    from melissa to All Participants:
	and reasoning on the whole ontology, creating new links based on exisitng ones through transitivity.

07/19/2007    19:17:21 PM    from John Day-Richter to All Participants:
	But the reasoner didn't draw the link as an implied link - it's drawing a different, implied link that was always there, but was hidden until now.

07/19/2007    19:17:50 PM    from John Day-Richter to All Participants:
	If this makes no sense, let me (seemingly) change topics for a second and we can come back to this.

07/19/2007    19:18:42 PM    from melissa to All Participants:
	I think I get it, but I still expect to see a solid is_a link to something, even if its a collapsed something. ok to change topics....

07/19/2007    19:19:14 PM    from John Day-Richter to All Participants:
	I think part of the confusion has to do with the way that the graph editor lets you show and hide nodes.

07/19/2007    19:19:28 PM    from melissa to All Participants:
	yes I think you are right.

07/19/2007    19:19:47 PM    from John Day-Richter to All Participants:
	In the old ontology editor panel, for every node there is ALWAYS a visible path back to the root.

07/19/2007    19:20:10 PM    from melissa to All Participants:
	like maybe if nodes are hidden that are required for is_a completeness, maybe the i should be big or a differnet color, ....

07/19/2007    19:20:19 PM    from melissa to All Participants:
	right.

07/19/2007    19:20:43 PM    from John Day-Richter to All Participants:
	I've got "periplasmic space (sense Proteobacteria)" seleced in the old ontology editor panel now.

07/19/2007    19:20:56 PM    from John Day-Richter to All Participants:
	We're seeing it under "periplasmic space".

07/19/2007    19:21:01 PM    from melissa to All Participants:
	yep

07/19/2007    19:21:19 PM    from John Day-Richter to All Participants:
	In the old ontology editor panel, there's no way for me to right click periplasmic space and say "just hide this node"

07/19/2007    19:21:33 PM    from John Day-Richter to All Participants:
	You're not allowed to hide nodes in the middle of the path.

07/19/2007    19:21:42 PM    from John Day-Richter to All Participants:
	But in the graph editor panel, you can.

07/19/2007    19:22:18 PM    from melissa to All Participants:
	so you clicked on a link and said show hidden terms? it went by fast

07/19/2007    19:23:03 PM    from John Day-Richter to All Participants:
	I clicked the implied link and chose "Show all supporting links", which means "Show all explicitly user-created links that were used to derive this implied link"

07/19/2007    19:23:22 PM    from John Day-Richter to All Participants:
	The implied link disappeared when I did that, because now that the evidence links are showing, the viewer trimmed out the implied link.

07/19/2007    19:23:54 PM    from John Day-Richter to All Participants:
	Right now, the graph editor is only showing explicit links, so it's showing the kind of graph that you might see in the ontology editor panel.

07/19/2007    19:23:55 PM    from melissa to All Participants:
	which is the is_a path between perip(sensu) and cell part. which is why I feel that they are collapsed and not implied.  sigh.

07/19/2007    19:24:40 PM    from John Day-Richter to All Participants:
	Part of the problem is that in the GO, implied links are always implied by transitivity. But in other ontologies, there are other ways to get an implied link.

07/19/2007    19:24:53 PM    from melissa to All Participants:
	I guess I understand implied links better when they are not is_a.

07/19/2007    19:25:32 PM    from John Day-Richter to All Participants:
	In SO, for example, there are links that are implied by intersections. In those cases, there's no collection of "collapsed" links behind them.

07/19/2007    19:26:11 PM    from melissa to All Participants:
	yes, I think we will have those too... once we coordinate better with the cell ontology.

07/19/2007    19:26:22 PM    from John Day-Richter to All Participants:
	Anyway, we're straying back into the previous conversation. I want to try to continue with my viewer comparison a little more.

07/19/2007    19:26:45 PM    from melissa to All Participants:
	ok

07/19/2007    19:26:53 PM    from John Day-Richter to All Participants:
	Before, I was saying that there's no way to click "periplasmic space" in the ontology editor panel and say "just hide this term".

07/19/2007    19:27:29 PM    from John Day-Richter to All Participants:
	But let's say I could do that. What would happen to "periplasmic space (sensu Proteobacteria)"?

07/19/2007    19:28:04 PM    from John Day-Richter to All Participants:
	One way to handle it would be to make "perplasmic space (sensu Proteobacteria)" appear to be a root node.

07/19/2007    19:28:37 PM    from John Day-Richter to All Participants:
	(This is how the ontology editor panel currently handles nodes that can't be connected to the current graph)

07/19/2007    19:28:53 PM    from melissa to All Participants:
	i get the reason you show the implied link, just not why its not of a different type.

07/19/2007    19:29:21 PM    from John Day-Richter to All Participants:
	What do you mean by "different type"? Why it has a dotted line?

07/19/2007    19:29:29 PM    from melissa to All Participants:
	yes

07/19/2007    19:29:56 PM    from melissa to All Participants:
	for comparison, lets look at the part_of graph on the dag viewer

07/19/2007    19:30:20 PM    from melissa to All Participants:
	the dotted line here represents some implied links through is_a transitivity, right?

07/19/2007    19:30:24 PM    from John Day-Richter to All Participants:
	Yes.

07/19/2007    19:30:37 PM    from melissa to All Participants:
	this is when I expect to see a dotted line.

07/19/2007    19:30:51 PM    from John Day-Richter to All Participants:
	Okay, now let's go back to the graph view.

07/19/2007    19:31:25 PM    from John Day-Richter to All Participants:
	When I hide the two selected nodes, we'll see an implied is_a link, for exactly the same reason we saw one in the graph dag viewer.

07/19/2007    19:31:49 PM    from John Day-Richter to All Participants:
	That is an is_a link implied by transitivity.

07/19/2007    19:32:23 PM    from John Day-Richter to All Participants:
	And the links that were used to generate the implied link are exactly those links that I just hid.

07/19/2007    19:33:11 PM    from melissa to All Participants:
	I get why its happening, its just that there are still two kinds of implied links in my mind. 

07/19/2007    19:33:34 PM    from melissa to All Participants:
	those that have a real explicit link in the ontology, and those that do not.

07/19/2007    19:34:21 PM    from John Day-Richter to All Participants:
	But no human being every created an is_a link between periplasmic space (sensu Proteobacteria) and cell part. That's why it's an implied link.

07/19/2007    19:34:26 PM    from John Day-Richter to All Participants:
	ever, not every

07/19/2007    19:34:44 PM    from melissa to All Participants:
	right, but they did create an is_a path to it.

07/19/2007    19:35:08 PM    from melissa to All Participants:
	maybe we need an example where we have part_of across is_a

07/19/2007    19:35:22 PM    from John Day-Richter to All Participants:
	Every single implied link in the go has a user-created path of links underlying it.

07/19/2007    19:35:27 PM    from melissa to All Participants:
	or maybe these types of dotted lines should be different

07/19/2007    19:35:53 PM    from melissa to All Participants:
	I think I am just confusing implied links via one link type vs. those across multiple link types

07/19/2007    19:36:40 PM    from John Day-Richter to All Participants:
	Maybe. Remember that transitivity can hold over a single link type, or a single link type + is_a.

07/19/2007    19:36:50 PM    from John Day-Richter to All Participants:
	But part_of isn't transitive over develops_from, for example.

07/19/2007    19:37:15 PM    from melissa to All Participants:
	right, but it is over is_a

07/19/2007    19:37:29 PM    from melissa to All Participants:
	right?

07/19/2007    19:37:41 PM    from John Day-Richter to All Participants:
	Exactly. So there's no difference between transitivity over a single type and transitivity over multiple types. It's all the same rule.

07/19/2007    19:37:58 PM    from John Day-Richter to All Participants:
	I will hide the "cell envelope" node to illustrate...

07/19/2007    19:38:03 PM    from melissa to All Participants:
	Ah! I think we are getting to the bottom of the problem here. there is a huge difference in my mind.

07/19/2007    19:39:06 PM    from John Day-Richter to All Participants:
	Great!

07/19/2007    19:39:09 PM    from melissa to All Participants:
	And this is why its unintuitive to see part_of links in the is_a hierarchy in the dag viewer (given is_a completeness)

07/19/2007    19:39:33 PM    from John Day-Richter to All Participants:
	I agree that it's unintuitive to see them there.

07/19/2007    19:39:35 PM    from John Day-Richter to All Participants:
	Oh, I get it!

07/19/2007    19:40:01 PM    from John Day-Richter to All Participants:
	You thought we were seeing them because those part_of links somehow contributed to the collection of is_a links!

07/19/2007    19:40:02 PM    from melissa to All Participants:
	ok I feel less stupid now. thank you.

07/19/2007    19:40:15 PM    from melissa to All Participants:
	yes!

07/19/2007    19:40:21 PM    from John Day-Richter to All Participants:
	But they were just there because the viewer was showing everything.

07/19/2007    19:40:47 PM    from melissa to All Participants:
	(in the dag viewer, yes)

07/19/2007    19:40:48 PM    from John Day-Richter to All Participants:
	This also explains Midori's comment about is_a links only being implied by other is_a links!

07/19/2007    19:40:56 PM    from melissa to All Participants:
	ye ye yes!

07/19/2007    19:41:39 PM    from John Day-Richter to All Participants:
	Okay, great. Then realize that the out of context links don't mean anything in the current version of the graph dag viewer - it's just the viewer trying to be thorough.

07/19/2007    19:41:54 PM    from John Day-Richter to All Participants:
	I'll fix this right away, and have a new version out this afternoon.

07/19/2007    19:42:01 PM    from melissa to All Participants:
	wow. 

07/19/2007    19:42:21 PM    from John Day-Richter to All Participants:
	I need to go take Logan to school now, but can we chat again this afternoon when the new version is ready.

07/19/2007    19:42:24 PM    from melissa to All Participants:
	somehow I didn't think we were going to find ourselves at the end of this problem..... hooray!

07/19/2007    19:42:25 PM    from John Day-Richter to All Participants:
	?

07/19/2007    19:42:42 PM    from melissa to All Participants:
	ok, fine. he's in school already? 

07/19/2007    19:42:49 PM    from melissa to All Participants:
	just skype me 

07/19/2007    19:43:07 PM    from John Day-Richter to All Participants:
	I'll skype you. Logan goes to Montessori (or however you spell that) in the afternoons.

07/19/2007    19:43:17 PM    from John Day-Richter to All Participants:
	Talk to you soon. I'll just leave this meeting open.

07/19/2007    19:43:28 PM    from melissa to All Participants:
	doesn't he have to be 2 1/2 to go to mont?

07/19/2007    19:43:33 PM    from melissa to All Participants:
	they do here

07/19/2007    19:43:44 PM    from melissa to All Participants:
	anyway I'll let you go.

07/19/2007    19:43:49 PM    from melissa to All Participants:
	talk to you later.