OE Webex 5Jul07

From GO Wiki
Jump to: navigation, search
July 5, 2007	    10:36:40 AM	    from Jennifer Clark� to All Participants:	hi


July 5, 2007	    10:37:54 AM	    from John Day-Richter� to All Participants:	Hi, all. There's a solution for the preferences overwriting problem, and I was composing an email to address it, but it involves passing a command-line switch to OBO-Edit.


July 5, 2007	    10:38:10 AM	    from Midori Harris� to All Participants:	Command line I can live with.
July 5, 2007	    10:38:30 AM	    from John Day-Richter� to All Participants:	Since I think most people don't know how to do that, I'm going to go back to the original plan of having beta versions write to their own preferences directory.


July 5, 2007	    10:38:34 AM	    from John Day-Richter� to All Participants:	.oboedit-beta, maybe


July 5, 2007	    10:38:53 AM	    from hurdil� to All Participants:	THAT I like even better
July 5, 2007	    10:39:02 AM	    from Midori Harris� to All Participants:	That would be fine.
July 5, 2007	    10:39:56 AM	    from Midori Harris� to All Participants:	I"ve been reluctant to test because it's so annoying to save a copy of 1.1 .oboedit, and copy everything back.
July 5, 2007	    10:40:14 AM	    from John Day-Richter� to All Participants:	Would you like it if the .oboedit-beta directory always reset itself on re-install? Like, it always wipes out the old .oboedit-beta directory when you install, and copies over your current .oboedit directory as .oboedit-beta?


July 5, 2007	    10:40:42 AM	    from Alexander Diehl� to All Participants:	Can this perhaps be controlled from within the program?
July 5, 2007	    10:40:57 AM	    from John Day-Richter� to All Participants:	I could add a screen to the installer that is only shown in beta installations.


July 5, 2007	    10:41:16 AM	    from Midori Harris� to All Participants:	At the moment, no strong preference; my concern is simply to preserve .oboedit for work I do in 1.1.
July 5, 2007	    10:41:17 AM	    from Alexander Diehl� to All Participants:	We might want to have persistent settings during testing most of the time.
July 5, 2007	    10:41:22 AM	    from Karen R Christie� to All Participants:	Would this only work with the installer? I still use the platform independent versions.
July 5, 2007	    10:41:33 AM	    from Midori Harris� to All Participants:	So do I.
July 5, 2007	    10:42:30 AM	    from Jennifer Clark� to All Participants:	wouldn't it be simpler just to have it do what john said first?


July 5, 2007	    10:42:47 AM	    from hurdil� to All Participants:	I think the simplest thing would be just to have it use a .oboedit-beta directory. We don't need to get fancy.
July 5, 2007	    10:42:47 AM	    from John Day-Richter� to All Participants:	I'm thinking I need to phase out the platform-independent version. Now that I'm trying to control memory settings through the installer, the platform independent version is going to become an expert-only option. If you're comfortable with that, keep on using it. But the installers work well now, and it makes sense to move a lot of initial configuration stuff into the installer.


July 5, 2007	    10:42:47 AM	    from Jennifer Clark� to All Participants:	just overwrite the beta one with 1.101 when you install a new beta


July 5, 2007	    10:43:18 AM	    from Jennifer Clark� to All Participants:	I always use the installer version and have not had any problems for ages


July 5, 2007	    10:43:32 AM	    from Jane Lomax� to All Participants:	nor me
July 5, 2007	    10:43:55 AM	    from Karen R Christie� to All Participants:	I don't mind using the installers if they allow me to have two versions. I used to find it problematic to use the installers when I wanted to install a test version without it replacing the official one.
July 5, 2007	    10:44:19 AM	    from Midori Harris� to All Participants:	I would like to keep using the platform-independent version, but I won't kick & scream.
July 5, 2007	    10:44:20 AM	    from Alexander Diehl� to All Participants:	Why not use a separate setting directory and provide a reset settings button in the Config menu?
July 5, 2007	    10:44:23 AM	    from hurdil� to All Participants:	What exactly goes on when you start the beta (even after installing) that it doesn't use the settings in the .oboedit directory, but essentially overwrites them  
July 5, 2007	    10:44:48 AM	    from John Day-Richter� to All Participants:	Well, that's not really true...


July 5, 2007	    10:44:55 AM	    from hurdil� to All Participants:	Specifically waht gets removed is my path to graphviz and the dbxreffs
July 5, 2007	    10:45:30 AM	    from Midori Harris� to All Participants:	My verification plugin settings get overwritten (or whatever) too.
July 5, 2007	    10:45:43 AM	    from John Day-Richter� to All Participants:	It seems like that's happening often because a lot of times a beta version will represent a certain setting in a different way than the official release, and I haven't yet provided a mechanism to import the old settings.


July 5, 2007	    10:46:14 AM	    from John Day-Richter� to All Participants:	Anyway, if we use the "beta versions use a different settings directory" system, we'll be fine.


July 5, 2007	    10:46:37 AM	    from hurdil� to All Participants:	Yes, that seems like the simplest to impliment
July 5, 2007	    10:46:52 AM	    from Midori Harris� to All Participants:	Makes sense to me.
July 5, 2007	    10:47:03 AM	    from Jennifer Clark� to All Participants:	me too


July 5, 2007	    10:47:12 AM	    from John Day-Richter� to All Participants:	We have 15 minutes before the demo section of this meeting. Can we talk a little about complex filter editors?


July 5, 2007	    10:47:54 AM	    from Midori Harris� to All Participants:	OK ... lemme find the page ...
July 5, 2007	    10:48:04 AM	    from Jennifer Clark� to All Participants:	can you send the url?


July 5, 2007	    10:48:23 AM	    from John Day-Richter� to All Participants:	We could just do it this way.


July 5, 2007	    10:48:27 AM	    from John Day-Richter� to All Participants:	But the URL is http://wiki.geneontology.org/index.php/OBO-Edit:_Proposals_for_Designing_a_New_Filtering_Interface


July 5, 2007	    10:49:37 AM	    from Midori Harris� to All Participants:	freaky - there's an error message about "media object not supported on your platform" in the Webex window, but the filter page openend in my default web browser.
July 5, 2007	    10:49:48 AM	    from Karen R Christie� to All Participants:	same here
July 5, 2007	    10:50:02 AM	    from hurdil� to All Participants:	same here
July 5, 2007	    10:50:04 AM	    from John Day-Richter� to All Participants:	Stupid! Webex must use ActiveX for showing web pages, which, of course, doesn't work on macs.


July 5, 2007	    10:50:12 AM	    from John Day-Richter� to All Participants:	Lemme try it this way...


July 5, 2007	    10:50:39 AM	    from John Day-Richter� to All Participants:	Can you see it now?


July 5, 2007	    10:50:51 AM	    from Midori Harris� to All Participants:	it was working ... I did get the right page before!
July 5, 2007	    10:50:59 AM	    from Midori Harris� to All Participants:	... So eithre way is fine.
July 5, 2007	    10:51:03 AM	    from Alexander Diehl� to All Participants:	Yes, and I could see it before too, in Safari
July 5, 2007	    10:51:27 AM	    from John Day-Richter� to All Participants:	Anyway, there are kind of 3.5 gui options here, and I wanted to get feedback on each of them.


July 5, 2007	    10:51:34 AM	    from hurdil� to All Participants:	I just typed it into my regular broswer to look
July 5, 2007	    10:51:47 AM	    from John Day-Richter� to All Participants:	First: Set-based searching:


July 5, 2007	    10:52:34 AM	    from John Day-Richter� to All Participants:	In this version, you create a simple filter, and then you "narrow" or "broaden" the filter with other filters.


July 5, 2007	    10:53:07 AM	    from John Day-Richter� to All Participants:	It's like you're saying, "Take the results of the last search, and either search within them (narrow) or add some more results (broaden) based on this other filter"


July 5, 2007	    10:54:18 AM	    from John Day-Richter� to All Participants:	The trouble with this option is that there are a lot of filters that can't be expressed this way (see the page for an example).


July 5, 2007	    10:54:47 AM	    from Midori Harris� to All Participants:	This sort of thing would be useful if you do a search, and get what you think are too many or too few results.
July 5, 2007	    10:55:08 AM	    from John Day-Richter� to All Participants:	Exactly. I think medline used to work this way.


July 5, 2007	    10:55:53 AM	    from John Day-Richter� to All Participants:	We can offer several of these filtering tools, so maybe this is one search mode we should offer.


July 5, 2007	    10:55:57 AM	    from Midori Harris� to All Participants:	With a bit of pummeling one can still get PubMed to behave this way.
July 5, 2007	    10:56:10 AM	    from Midori Harris� to All Participants:	For that reason it'll be very familiar to biologists.
July 5, 2007	    10:56:31 AM	    from Jennifer Clark� to All Participants:	I just had a quick read and I like the wiring diagram best


July 5, 2007	    10:56:50 AM	    from John Day-Richter� to All Participants:	Should we look at other options, or are there more comments about this one?


July 5, 2007	    10:57:20 AM	    from Midori Harris� to All Participants:	I was just going to say, it would be good to be able to offer this type [set-based] of search; I suspect it would be used much more frequently than the more complex searches anyway.
July 5, 2007	    10:57:26 AM	    from Chris Mungall� to All Participants:	I don't have anything to add to what I said on the list: the GUI suggestions look really cool, but graphical editing is over the top for the majority of queries. A standard table of constraints (eg as in mac Mail) is fine for 99% of filter queries
July 5, 2007	    10:57:26 AM	    from Alexander Diehl� to All Participants:	Go on to the other options.
July 5, 2007	    10:58:01 AM	    from Jennifer Clark� to All Participants:	I agree with chris


July 5, 2007	    10:58:43 AM	    from John Day-Richter� to All Participants:	Chris - I think you're right, but complex queries are absolutely for breaking an ontology into several files, and that will be done more and more frequently, so I think we need a painless way to design and test these queries.


July 5, 2007	    10:59:17 AM	    from John Day-Richter� to All Participants:	Anyway, next option "Wiring Diagrams"


July 5, 2007	    10:59:28 AM	    from Alexander Diehl� to All Participants:	I agree with both John and Chris.
July 5, 2007	    10:59:35 AM	    from Jennifer Clark� to All Participants:	if we need a gui for the queries I think the wiring diagram is very intuitive.


July 5, 2007	    10:59:42 AM	    from Jennifer Clark� to All Participants:	the other's not so much


July 5, 2007	    10:59:49 AM	    from Jennifer Clark� to All Participants:	others


July 5, 2007	    10:59:55 AM	    from Midori Harris� to All Participants:	I disagree ... but let's hear what John has to say first, and then comment.
July 5, 2007	    11:00:09 AM	    from John Day-Richter� to All Participants:	In this kind of editor, the query is designed as a wiring diagram that shows how your query input (some term) relates to the output (whether the query is true or false).


July 5, 2007	    11:00:22 AM	    from John Day-Richter� to All Participants:	If you ever played the 1980's game "Rocky's Boots", this will look familiar.


July 5, 2007	    11:00:36 AM	    from John Day-Richter� to All Participants:	You start with a blank screen like the one I'm displaying.


July 5, 2007	    11:01:02 AM	    from John Day-Richter� to All Participants:	Then you add all your basic sub-queries (but no ands and ors).


July 5, 2007	    11:01:55 AM	    from John Day-Richter� to All Participants:	You drag "And" and "Or" nodes from the toolbar onto the screen, and then you draw "wires" between the outputs of your subqueries and the inputs of your boolean operators (as shown)


July 5, 2007	    11:02:15 AM	    from John Day-Richter� to All Participants:	You connect the whole mess to the "output" node.


July 5, 2007	    11:02:45 AM	    from John Day-Richter� to All Participants:	You can drop a term into the "input" node to test your query. Nodes light up depending on whether they match the query.


July 5, 2007	    11:03:26 AM	    from Tanya� to All Participants:	sorry off to my competing meeting


July 5, 2007	    11:03:30 AM	    from Alexander Diehl� to All Participants:	Can you have tooltips that tell subresults for each subquery?
July 5, 2007	    11:03:47 AM	    from John Day-Richter� to All Participants:	You can tell by the coloring.


July 5, 2007	    11:04:01 AM	    from John Day-Richter� to All Participants:	If it's blue (in my mockup), the subquery matched.


July 5, 2007	    11:04:17 AM	    from John Day-Richter� to All Participants:	My feeling is that this model will make a lot of sense to people with computer science or EE backgrounds, and will be utterly incomprehensible to everyone else.


July 5, 2007	    11:04:40 AM	    from Jane Lomax� to All Participants:	I dunno - I find this quite intuitive...
July 5, 2007	    11:04:41 AM	    from Jennifer Clark� to All Participants:	it makes sense to me and I'm a botanist.


July 5, 2007	    11:04:51 AM	    from Alexander Diehl� to All Participants:	I like it a lot.
July 5, 2007	    11:04:54 AM	    from Karen R Christie� to All Participants:	well, just from what I see on the diagram you've drawn, it's an improvement to me over the current system, which I still can't use easily
July 5, 2007	    11:04:56 AM	    from Midori Harris� to All Participants:	It's certainly going to be incomprehensible to the vast majority of biologists (my esteemed colleagues notwithstanding).
July 5, 2007	    11:04:59 AM	    from hurdil� to All Participants:	This makes perfect sense to me; I loke it a lot also
July 5, 2007	    11:05:15 AM	    from John Day-Richter� to All Participants:	This is good news, because the other options are all simplifications of this editor.


July 5, 2007	    11:05:20 AM	    from Midori Harris� to All Participants:	So I'm the oddball. No problem.
July 5, 2007	    11:05:31 AM	    from John Day-Richter� to All Participants:	Let's quickly go through some of the simplified versions...


July 5, 2007	    11:06:04 AM	    from John Day-Richter� to All Participants:	It's really easy to build a tree display version of this that hides the "input" and "output" nodes, and shows the wiring diagram as a tree.


July 5, 2007	    11:06:29 AM	    from Midori Harris� to All Participants:	It's not that it's unintuitive to me ... it's more that I can sort of remember what I might've been able to get my head around back when I was a bench biologist. And this wouldn't've made the cut.
July 5, 2007	    11:06:48 AM	    from Jennifer Clark� to All Participants:	this isn't so clear to me at all.


July 5, 2007	    11:07:07 AM	    from John Day-Richter� to All Participants:	The other possible simplification is the nested boxes option, which I thought was the most elegant (although I bet most people will disagree)


July 5, 2007	    11:07:22 AM	    from John Day-Richter� to All Participants:	In this model, you start with your basic queries...


July 5, 2007	    11:07:32 AM	    from Midori Harris� to All Participants:	I actually did like the boxes best.
July 5, 2007	    11:07:47 AM	    from John Day-Richter� to All Participants:	The boolean operators are represented as containers.


July 5, 2007	    11:08:05 AM	    from John Day-Richter� to All Participants:	You drop the sub-queries into a boolean container to create a complex query.


July 5, 2007	    11:08:28 AM	    from John Day-Richter� to All Participants:	You can drop containers into other containers to create really complex nested queries.


July 5, 2007	    11:08:42 AM	    from John Day-Richter� to All Participants:	(see the currently displayed example)


July 5, 2007	    11:08:58 AM	    from Jennifer Clark� to All Participants:	I wouldn't get that if I hadn't seen the wiring diagram.


July 5, 2007	    11:09:05 AM	    from Jane Lomax� to All Participants:	no, nor me
July 5, 2007	    11:09:12 AM	    from Karen R Christie� to All Participants:	this makes equal sense to me as the wiring diagram
July 5, 2007	    11:09:15 AM	    from John Day-Richter� to All Participants:	This model can give the exact same debugging/testing feedback as the wiring diagram, but it's harder to understand why at first.


July 5, 2007	    11:09:33 AM	    from hurdil� to All Participants:	does the "coloring" also mean the same as the wiring diagram?
July 5, 2007	    11:09:36 AM	    from John Day-Richter� to All Participants:	Yeah.


July 5, 2007	    11:09:50 AM	    from Jennifer Clark� to All Participants:	if we wrote our queries in that complicated looking code with nesting and brackets and whotnot would it show up in this gui for us to debug?


July 5, 2007	    11:09:50 AM	    from John Day-Richter� to All Participants:	The good thing about this one is that the editor would be a lot easier to use than the wiring diagram. You could design your query much faster.


July 5, 2007	    11:10:01 AM	    from John Day-Richter� to All Participants:	Yes, we could do that, Jen.


July 5, 2007	    11:10:09 AM	    from Midori Harris� to All Participants:	I didn't need the wiring diagram at all to help explain the other two.
July 5, 2007	    11:10:09 AM	    from John Day-Richter� to All Participants:	As long as your query was parseable, that is.


July 5, 2007	    11:10:48 AM	    from Midori Harris� to All Participants:	One thing I wondered about for all 3 is the amount of precision mousing involved ...
July 5, 2007	    11:10:52 AM	    from John Day-Richter� to All Participants:	Another option is that we could allow the user to switch between the different views, and see their query represented in different ways.


July 5, 2007	    11:10:54 AM	    from Chris Mungall� to Presenter (privately):	[just to you] - it'd be great if Mark could see this. The phenote post-composition interface needs work. I'm sure you're aware that an xp def and a query are no different, a precomposed term with an xp def is really just a named query...
July 5, 2007	    11:11:01 AM	    from Karen R Christie� to All Participants:	better speed in designing queries would be good
July 5, 2007	    11:11:14 AM	    from Jennifer Clark� to All Participants:	quick sounds good but this is very unintuitive to me


July 5, 2007	    11:11:32 AM	    from John Day-Richter� to All Participants:	Jen - is that because the "inside-out" presentation seems strange to you?


July 5, 2007	    11:11:41 AM	    from Jennifer Clark� to All Participants:	yes


July 5, 2007	    11:12:10 AM	    from John Day-Richter� to All Participants:	It sounds like the most popular full-featured options are "wiring diagram" and "boxes".


July 5, 2007	    11:12:25 AM	    from Jennifer Clark� to All Participants:	yes


July 5, 2007	    11:12:28 AM	    from John Day-Richter� to All Participants:	And it sounds like we need to offer both, and offer a way of switching between them.


July 5, 2007	    11:12:32 AM	    from hurdil� to All Participants:	I have a feeling that the wiring diagram mode is easiest
July 5, 2007	    11:12:38 AM	    from Jennifer Clark� to All Participants:	would that take a lot of work?


July 5, 2007	    11:12:50 AM	    from John Day-Richter� to All Participants:	That way people can get started with wiring diagram, and then start using boxes when they realize how it works, and how much faster it is.


July 5, 2007	    11:13:03 AM	    from Jennifer Clark� to All Participants:	that seems like a good idea


July 5, 2007	    11:13:19 AM	    from Alexander Diehl� to All Participants:	I still want tooltips that show the number of matchs at each step -- this would be helpful for debugging a query.
July 5, 2007	    11:13:23 AM	    from Jennifer Clark� to All Participants:	the wiring would help new users in courses


July 5, 2007	    11:13:37 AM	    from Karen R Christie� to All Participants:	for pure simplicity, I really like the first option, if it's not too hard, maybe we can allow building queries in this manner even if we have the wiring diagram
July 5, 2007	    11:13:39 AM	    from Jennifer Clark� to All Participants:	boxes would be good for power users/


July 5, 2007	    11:13:46 AM	    from John Day-Richter� to All Participants:	Alex: That's harder than it sounds...


July 5, 2007	    11:14:11 AM	    from Alexander Diehl� to All Participants:	Come on, Java execution is fast!
July 5, 2007	    11:14:23 AM	    from Midori Harris� to All Participants:	Will any of these support a "search within results" option?
July 5, 2007	    11:14:31 AM	    from John Day-Richter� to All Participants:	That's what narrow/widen is.


July 5, 2007	    11:14:40 AM	    from Chris Mungall� to Presenter (privately):	I agree with Alex - intermediate results would be great
July 5, 2007	    11:14:52 AM	    from Midori Harris� to All Participants:	Good! I would find that more useful than any of the fancy gui things.
July 5, 2007	    11:15:04 AM	    from John Day-Richter� to All Participants:	Alex: It's possible, but it requires tricky coding. I can do it, but we've got to sink in the time.


July 5, 2007	    11:15:22 AM	    from Alexander Diehl� to All Participants:	Okay.
July 5, 2007	    11:15:51 AM	    from John Day-Richter� to All Participants:	And, of course, there's the expression language option too, which I forgot to mention.


July 5, 2007	    11:16:03 AM	    from Jennifer Clark� to All Participants:	I like that a lot


July 5, 2007	    11:16:23 AM	    from John Day-Richter� to All Participants:	Note that this option could allow color-based debugging too. I can light up the sub-phrases that match.


July 5, 2007	    11:16:55 AM	    from Jennifer Clark� to All Participants:	there could be auto indenting too


July 5, 2007	    11:17:19 AM	    from John Day-Richter� to All Participants:	Yup.


July 5, 2007	    11:18:12 AM	    from John Day-Richter� to All Participants:	I think the best way to proceed from here is to send around an email asking people to rank, on a 1 to 10 scale, how we should prioritize development of each of these editors.


July 5, 2007	    11:18:57 AM	    from Chris Mungall� to All Participants:	I think we should start by having people compile their favourite queries, and the queries they wish they could do on the wiki..
July 5, 2007	    11:19:20 AM	    from Chris Mungall� to All Participants:	It's hard to evaluate these without specific queries in mind
July 5, 2007	    11:19:39 AM	    from John Day-Richter� to All Participants:	That's a good idea. I'll add a wiki section.


July 5, 2007	    11:19:54 AM	    from Jennifer Clark� to All Participants:	I've never really understoof link queries so I think there are things that I could be doing that I don't know about


July 5, 2007	    11:19:58 AM	    from Chris Mungall� to All Participants:	Also I'm not sure what you meant by the need to break the ontology down into files; sounds like a one-off query?
July 5, 2007	    11:20:44 AM	    from Chris Mungall� to All Participants:	And I think these should also be evaluated against a bog-standard table of constraints that can be combined with and/or, a la mac mail, old pubmed etc
July 5, 2007	    11:20:59 AM	    from John Day-Richter� to All Participants:	If you're doing cross product editing, you may need a fairly complex multiplexing of ontologies into files. Like maybe some terms go to file A, some to file B, and a set of special links go to file C.


July 5, 2007	    11:21:10 AM	    from John Day-Richter� to All Participants:	Chris, what do you mean by "table of constraints"?


July 5, 2007	    11:21:57 AM	    from Chris Mungall� to All Participants:	I just mean the constraints are shown in a table, non-graphically. Ever used the smart mailbox filter optionon mac mail?
July 5, 2007	    11:22:28 AM	    from Jennifer Clark� to All Participants:	should we get someone to demo it?


July 5, 2007	    11:22:54 AM	    from John Day-Richter� to All Participants:	I think something very similar is used in my mail client. That model of complex querying assumes that you're using only one kind of boolean operation.


July 5, 2007	    11:23:15 AM	    from John Day-Richter� to All Participants:	(Usually "and")


July 5, 2007	    11:23:51 AM	    from Chris Mungall� to All Participants:	yes and I'll bet this satisfies 99% of cases  (usually you can switch between and/or in one set). but i don't know for sure. for the remaining 1% as long as there is some other way to do it
July 5, 2007	    11:24:25 AM	    from John Day-Richter� to All Participants:	You're right. We should offer this option too.


July 5, 2007	    11:24:44 AM	    from Chris Mungall� to All Participants:	(I am actually really keen on the gui, I just think we have to properly justify it)
July 5, 2007	    11:24:59 AM	    from John Day-Richter� to All Participants:	We're way over time for this discussion, so I think I'd like to demo the new installer screen, rather than try to demo any graph stuff.


July 5, 2007	    11:25:16 AM	    from John Day-Richter� to All Participants:	Some people said they didn't see a new installer screen, so I want to show everyone what it looks like.


July 5, 2007	    11:25:26 AM	    from Midori Harris� to All Participants:	OK; we can revisit the graph next week.
July 5, 2007	    11:26:07 AM	    from John Day-Richter� to All Participants:	Here's the beta6 installer. As you go through the installation, everything looks normal for a while...


July 5, 2007	    11:26:25 AM	    from John Day-Richter� to All Participants:	But suddenly, <gasp!>, a new screen!


July 5, 2007	    11:27:11 AM	    from John Day-Richter� to All Participants:	The installer will try to come up with a decent guess for how much memory to give to OBO-Edit, which is about 80% of your system memory, or 512M, whichever is larger.


July 5, 2007	    11:27:45 AM	    from John Day-Richter� to All Participants:	If the installer can't figure out how much system memory you have (if you're using Solaris, for example), it will suggest 512M.


July 5, 2007	    11:28:20 AM	    from John Day-Richter� to All Participants:	Put whatever number you want in there, and then OBO-Edit will automatically use that memory setting whenever you start up.


July 5, 2007	    11:28:50 AM	    from John Day-Richter� to All Participants:	Note that this memory setting behaves differently than all other settings in OBO-Edit. This setting is stored with the installed application, NOT in your .oboedit directory.


July 5, 2007	    11:29:09 AM	    from John Day-Richter� to All Participants:	(At least, this is how OBO-Edit behaves as of beta6)


July 5, 2007	    11:29:47 AM	    from John Day-Richter� to All Participants:	I'd like people to try this out and make sure it works, and make sure it's producing a reasonable guess for your machine.


July 5, 2007	    11:29:58 AM	    from Jennifer Clark� to All Participants:	It worked for me


July 5, 2007	    11:30:12 AM	    from Jane Lomax� to All Participants:	yep ,and me
July 5, 2007	    11:30:39 AM	    from John Day-Richter� to All Participants:	(It guessed very low here because I'm not using a real machine - I'm using windows from inside VMWare on a linux machine - for my real Linux installation it said I should use about 1100M, which I think is a very good guess)


July 5, 2007	    11:30:58 AM	    from John Day-Richter� to All Participants:	I'm glad it's working for people so far. I'


July 5, 2007	    11:31:28 AM	    from John Day-Richter� to All Participants:	ll release a new OBO-Edit soon that uses a different config directory for beta versions, and I'll create a new wiki page for people to post their favorite/trickiest queries.


July 5, 2007	    11:31:33 AM	    from Alexander Diehl� to All Participants:	In the installer it had a number like 1.536 or something, but the program shows 512M in the Configuration manager.
July 5, 2007	    11:32:12 AM	    from Karen R Christie� to All Participants:	regarding the installer, how would I know what is a reasonable guess for my machine
July 5, 2007	    11:32:50 AM	    from John Day-Richter� to All Participants:	It shouldn't be more than the amount of memory you have total, and it should be around what you manually set before we added this feature.


July 5, 2007	    11:32:58 AM	    from Karen R Christie� to All Participants:	I didn'
July 5, 2007	    11:33:07 AM	    from Karen R Christie� to All Participants:	I didn't manually set before, so I have no idea there
July 5, 2007	    11:33:30 AM	    from John Day-Richter� to All Participants:	You must have. Whatever your Configuration Manager memory setting was.


July 5, 2007	    11:33:51 AM	    from Jennifer Clark� to All Participants:	do you know how much ram you have?


July 5, 2007	    11:33:59 AM	    from John Day-Richter� to All Participants:	If you didn't set it, it would have used the default 512M, I guess.


July 5, 2007	    11:34:51 AM	    from Karen R Christie� to All Participants:	About this Mac says I have 2 GB DDR2 SDRAM
July 5, 2007	    11:35:15 AM	    from John Day-Richter� to All Participants:	Karen - it should be guessing around 1600M, then.


July 5, 2007	    11:35:47 AM	    from John Day-Richter� to All Participants:	Time's up, folks. Shall we meet again next week?


July 5, 2007	    11:35:57 AM	    from Jennifer Clark� to All Participants:	yes


July 5, 2007	    11:36:07 AM	    from Karen R Christie� to All Participants:	OK, I'll see what it does when I test, after you release a version that writes its prefs into a beta directory ;)
July 5, 2007	    11:36:14 AM	    from John Day-Richter� to All Participants:	Fair enough.


July 5, 2007	    11:36:20 AM	    from John Day-Richter� to All Participants:	Bye everyone.


July 5, 2007	    11:36:24 AM	    from Karen R Christie� to All Participants:	bye
July 5, 2007	    11:36:26 AM	    from Midori Harris� to All Participants:	Next week is regular meeting time. See you then!
July 5, 2007	    11:36:27 AM	    from Jennifer Clark� to All Participants:	bye


July 5, 2007	    11:36:41 AM	    from Alexander Diehl� to All Participants:	bye