1: Thanks to Frank for his mentorship
2: legal update
3: feature model discussion
<gtbot> Agenda Items:
<gtbot> End of agenda items.
<jdeolive> i have one but will add it as the last
<acuster> gtbot add Thanks to Frank for his mentorship
<gtbot> Added agenda item '1: Thanks to Frank for his mentorship' to the list.
<jdeolive> since it will probably take up the majority of the meeting
<acuster> gtbot add legal update
<gtbot> Added agenda item '2: legal update' to the list.
<acuster> well we can start with 0. what is up
– pramsey (n=pramsey@S01060014515fec41.gv.shawcable.net) has joined #geotools
<acuster> acuster – Looking at the renderering systems; legal stuff
<jdeolive> jdeolive – feature model
<acuster> gtbot start
<gtbot> Logging started.
<jdeolive> gtbot add feature model discussion
<gtbot> Added agenda item '3: feature model discussion' to the list.
<groldan> groldan – almost away from geotools, immerse in udig's edit tools
<acuster> jdeolive, what are you doing more precisely?
<acuster> groldan, any conclusion on the 'modality' stuff?
<jdeolive> jdeolive – waiting for feedback between andrea, gabriel, jgarnett, and myself to settle down in order to cimpile to concrete geoapi api changes
<jdeolive> add acuster to that list of people as well 
<groldan> so far the experiment is doing well, was able to add modes for ortho, snap to vertex, snap to line, and parallel
<jgarnett> hello; I am having some #foss4g lab drama so I am not sure how attentive I will be to the meeting today
<acuster> jdeolive, so the feature model has to land in geoapi to land in geotools as well?
<groldan> no conclusion though, as I'd love a code review
<jdeolive> yes
<jgarnett> acuster - usual split geoapi interfaces, geotools implementation (or in this case more than one)
<acuster> jdeolive, what's your personal perspective on temporal support? An infrequent issue to be dealt with later?
<jdeolive> acuster, hmmm... yeah i would say it should come later, i think we have bigger fish to fry
<acuster> jgarnett, do you have time to consider the legal stuff in the next 24hours or should I push it through without you?
<jdeolive> like just coming up with a core model which which will work for complex stuff but not too complex that it kills us
<acuster> yeah, I like the direction andrea is pushing
<acuster> okay, 10 after lets start
<acuster> 1) Frank is resigning as mentor
<acuster> We've been having a private discussion over legal matters
<acuster> which has been painful for everyone
<acuster> and Frank finally had enough
<acuster> he wants to enjoy greener pastures, like his emacs or vi screen (not sure which he uses)
<acuster> I wanted to thank him:
<acuster> Thanks for all your hard work Frank
<acuster> and let everyone know
<acuster> Replacements and other issues have not yet been decided
<acuster> questions?
<acuster> 2) Legal update
<acuster> This is 'do or die'
<acuster> I have yet another draft for the copyright assignment agreement
<acuster> http://docs.codehaus.org/display/GEOTOOLS/Geotools+Legal+Review
<acuster> along with some other docs including background
<acuster> if anyone wants to comment they have 24hours
<acuster> after that, I'll send it to a board member asking that it be sent on to the lawyers for a review
<acuster> if this version doesn't yield a document we can sign, then I give up
<acuster> so either FOSS4G will have a document
<acuster> or we will get to start all over again
<acuster> questions?
<acuster> jdeolive, the floor is yours
<jdeolive> ok... we need aaime, groldan, jgarnett, acuster, and myself for this topic
<jdeolive> without all of us i will just skip it
<acuster> andrea is out
<acuster> jgarnett, is distracted (not that that is new)
– aaime (n=aaime@host227-95-dynamic.50-82-r.retail.telecomitalia.it) has joined #geotools
<acuster> oh!
<groldan> and jody is complicated
<jdeolive> speak of the devil
<simboss> here is andrea
<aaime> Hi
<groldan> hi
<acuster> so jdeolive gets to write up his topic after all
<jdeolive> ok
<jdeolive> lets try without jody
<jdeolive> so... the topic is geoapi feature model
<jdeolive> andrea did a great review of it
<aaime> yikes"
<jdeolive> and since then there has been alot of back and forth
<jdeolive> so... my question is for you andrea... are there any major things that have not been unaswered for you?
<acuster> ? what is the work that actually needs to happen? e.g. do all the datastores have to be re-written?
<acuster> ? is the prototype SimpleFeature system written and/or have a javadoc kicking around?
<groldan> before deciding that we need an up to date data access api,
<groldan> guess jdeolive is just talking about the fm
<groldan> sure both things are closely related
<groldan> yet, I guess he's not trying to get that far?
<jdeolive> i would like to focus on the actual model for now
<jdeolive> we can talk about gt datastores later
<aaime> (e.g. when we have resources to tackle them too)
<jdeolive> ping aaime
<acuster> and that model is all on geoapi trunk?
<jdeolive> yes
<aaime> jdeolive
<jdeolive> sorry, i had a question for you
<jdeolive> will repeat it
<aaime> There are no unanswered questions
<aaime> that I can remember of
<jdeolive> so... my question is for you andrea... are there any major things that have not been unaswered for you?
<aaime> Nope
<jgarnett> hi - it seems there a are a lot of questions.
<aaime> I'm not completely comfortable with how multiple instance attributes are managed
<jdeolive> ok... so based on the feaure model review thread i can go trhough and take my first chop at geoapi
<aaime> I guess so
<groldan> Andrea do you think the bad feeling you have could be left to implementation details, or there is some strong api problem?
<jgarnett> justin I have a document I would like you to read; I am sorry I was too busy to take a part in the email thread.
<aaime> (that is, as an OO developer I would feel more comfortable with them being managed in their own collections)
<acuster> jdeolive, that's what I was asking. What's a first chop? Geotools refactoring?
<jdeolive> nope, geoapi interface changes
<jdeolive> they are overly complex, and inconsistent in some places
<jdeolive> i would like to cut that all out
<jdeolive> to make it easier to ingest
<groldan> do you have a feeling of what changes are those going to be?
<jdeolive> my rationale being
<acuster> so what prototype is andrea working from?
<aaime> Andrea is not working, he's just looking at the geoapi trunk
<aaime> I haven't touched anything
<aaime> I was just being asked for an opinion
<aaime>
<jdeolive> right, andrea has just been reviewing
<jdeolive> and my concern is this
<jdeolive> if a java developer as advanced and experienced as andrea has issues with it
<jdeolive> we have serious problems
<aaime> The main problem is that I love OO and hate (hate!!!) XML
<groldan> being the one that perhaps had to deal more with it (the model, not andrea ), I recognize its sometimes hard to work with
<aaime> so if you asked another person the review outcome would have been different
<jgarnett> guys I am going to have to bail on this meeting; too much else going on ... justin I will send you what I have for feedback via email. If you have time for a breakout IRC or coffee on the subject it would be nice. If not I missed the boat.
– jgarnett (n=Jody@mail.refractions.net) has left #geotools
<groldan> but the pain is meant to be softened by the family of builders
<jdeolive> jgarnett, ok, i need to wade through that massive email thread and reorganizse
<jdeolive> he is gone
<acuster> who will the 'model users' be?
<jdeolive> geotools developers
<aaime> Well, whoever plays with geotools
<jdeolive> anyone who writes code having to do with Feature
<jdeolive> which is pretty much all geotools client code
<groldan> except Martin
<jdeolive> haha
<groldan> (lucky guy)
<aaime> ha ha, right
<jdeolive> fair enough
<acuster> and what are we trying to help them do? be quick? understand quickly? do lots?
<aaime> do more
<aaime> handle complex features
– acuster is imagining an ecologist trying to use feature
<aaime> basically being able to represent whatever xml madness you can find in
<aaime> gml schemas
<aaime> 1:1 inside the gt2 features
<groldan> disagree andrea
<acuster> right we have Rob A. at one end, and jane scientist at the other
<groldan> we got rid of most of the xml madness
<groldan> and in the community schema modules
<groldan> xml specifics are being treated separately
<acuster> are there 'test features'?
<aaime> Well, mapping childs like an unsorted bag of stuff smells like xml madness to me
<groldan> aka, concerns tend to be quite separated, and that's why we're carring out the emf model as metadata
<jdeolive> well... fair enough.. the model is not tied to xml directly
<groldan> actually, doing it as lists would be more xml'ized
<jdeolive> but its definitley influenced by it and gml
<jdeolive> and has been made flexible enough to handle model xml schema and gml more exactly
<aaime> Afaik one of the objectives was to map the xml 1:1 into the feature no?
<groldan> two years ago... yes. With this last model I would say no
<groldan> its more biased to map 1:1 with the ogc general feature model
<jdeolive> which is highly influenced by xml / gml no?
<groldan> and that's why an attribute's property bag is represented as a Collection and not a list
<aaime> yep, tell it what you like, if the stuff you're handling cannot be represented by a javabean it's not very OO to me
<groldan> jdeolive: don't know, guess the GFM is an abstract model, gml dont
<groldan> and the former exists prior to gml? (not sure)
<jdeolive> well regardless... the new model is definitley more xml like then the old one
<jdeolive> ie, having descriptors mapping to schema particles, etc...
<groldan> fair enough, the old one is more JDBC like
<aaime> not really, more shapefile like
<aaime> jdbc woud imply natural support for joins
– acuster already has problems with the terminology 'A feature is an attribute?'
<aaime> (and thus complex features made the "sane way", that is, with association managed with collections and the like)
<aaime> Adrian, yes, I don't like the semantic of that hiearchy either, but from a structural point of view it works
<acuster> we do have to content with fourty odd years of GIS here though
<acuster> we can't pick names we like
<aaime> I fear that hiearchy maps the OGC feature description quite verbatim
<acuster> Feature.java is a mess
<acuster> ah, ok. Then we need to state that
<acuster> A Feature, of arbitrary complexity, with at a minimum Geometry and CRS information.
<aaime> Ha ha, my first reaction as well, you have to start with FeatureType to have some understannding
<acuster> does it have to have a geom? does it have to have a CRS?
<aaime> No and no (not necessarily)
<acuster> who are the authors, what's the license?
<aaime> they should be optional
– acuster moves to featureType
<aaime> Start from property type, go down
<aaime> inspect all the hiearcy, then have a look at descriptor
<aaime> or even better, have a look at my review
<aaime> there is a diagram and a summary for both the type model
<aaime> and the attribute/feature part
– jgarnett (n=Jody@mail.refractions.net) has joined #geotools
<acuster> Declaration of attribute type.
<acuster> how does that kind of javadoc help anyone?
<jdeolive> yes, the javadocs are quite poor
<aaime> The javadoc is not helping yeah
<jdeolive> right, so its very poorly documented
<jdeolive> thats a bad thing
<acuster> The javadoc is how we get users to become developers and do work for us
<jdeolive> and maybe decent docs would solve some of our problems with it being over complex
<aaime> Javadoc won't solve the complexity problem
<aaime> only xml example of what you can and cannot do with the various options start to give the model some sense
<aaime> if you tackle it from a pure OO perspective it looks like madness
<aaime> (especially if you've already bean exposed to pure OO metamodels like the MOF)
– jdeolive is not familiar with MOF
<jdeolive> but i agree
<aaime> Meta Object Framework, it's what UML is based on
<jdeolive> some examples of what different classes mean and relating to an xml schema structure woudl probably be helpful
<acuster> so a developer can't start at PropertyType and read the code to figure out what's going on?
<aaime> Imho no, you have to start from xml, if you look at it from code only it does not make sense
<jdeolive> no, this model is 100% not ready for production, we know that
<jdeolive> we are tryign to get it there
<jdeolive> this was painful for andrea
<jdeolive> and i apologize
<aaime> you have to see what there is in xml that you cannot map in a pure oo model to understand
– acuster will stop harping on the javadocs
<jdeolive> ok
<jdeolive> well to be honest, i am quite pessimistic based on peoples feedback
<jdeolive> but i think the next step is for me to go through and try to sum up things so far
<aaime> jdeolive, first contact with that feature model is going to hurt because you cannot recognize the usual mental model of an OO model
<aaime> and you don't get any help understanding why
<aaime> it's made like this
<aaime> So yes, there are simplifications needed
<aaime> the javabean convention has to be respected
<aaime> some interfaces need to be factored out
<aaime> but in the end it would just make it less messy
<jdeolive> right, this is the list i am trying to come up with
<jdeolive> exactly what needs to be done
<aaime> the hard part would stay there, that is, descriptors, attributes as bags of free floating attributes
<aaime> and that needs some good docs to become understandable
<jdeolive> agreed
<groldan> cool, whatever needs to be done, remember we have two real world models, borehole and timeseries, that will help out to know if the model is good enough
<jdeolive> a good run of the javadocs should have been done before you looked at it
<jgarnett> hi guys; I am back ...
<jdeolive> good point gabriel
<aaime> groldan, that is good as long as it's out of your mind and down into a document
<aaime> otherwise only you will keep having good night sleeps
<jgarnett> yeah; I would really like to see those turned into code examples we can all understand.
<aaime> whilst everybody else is left wondering
<groldan> gotcha
– cholmes_ (n=cholmes@cpe-66-108-80-238.nyc.res.rr.com) has joined #geotools
<jgarnett> justin have you given any thoughts to the migration plan?
<jgarnett> ie migration from FetureType.create to SimpleFeatureBuilder ... what happens when and what does it return?
<jdeolive> well i did but that is on the backburner now
<jdeolive> getting this model usable is paramount
<jgarnett> understood.
<jdeolive> using builders and that is a very very small part of the success, and the easy part if you ask me
<jgarnett> I have a good idea about that; but I would like to draw a picture over coffee. Perhaps you will hate it?
<jdeolive> the problems andrea and adrian point out are the important part
<jdeolive> ok... so does anything have anythign else?
<aaime> I agree. If we concentrate on SimpleFeature everything works, but we get Feature and FeatureType in a trojan horses
<aaime> and we may not like the consequences
<jgarnett> nothing you will like ...
<aaime> ?
<jgarnett> I would like to see the javadocs cleaned up .. but I see others have made that point.
<acuster> Where are the data sets?
<jdeolive> yes, i would like to do that jody
<jdeolive> take a run at the entire set of javadocs and make them consistent across the board
<jgarnett> gottcha
<groldan> acuster: you have two on the community-schema config for the community-schema modues in geoserver, and we could certainly ask Rob Atkinson for some more
<acuster> ah, in geoserver
<aaime> (not in the official geoserver)
<jgarnett> so is there anything else this meeting? we have some legal stuff to sort out before foss4g if I understand my inbox.
<groldan> and in the geotool's module test case
<acuster> jgarnett, I'm getting pissed off enough that I may drop the whole thing
<jgarnett> acuster++
<jgarnett> I am sorry we have been less than helpful; I got to the same point last year.
<groldan> any conclusion on the topic? jdeolive do you have what you need? gonna move this to the ml?
<jgarnett> There is an OSGeo meeting on the monday where I hope to explain that the iccubation process is not being supportive for us.
<cholmes_> acuster, does that mean that you wouldn't consider signing any contributor agreement that isn't one that you worked on?
<jgarnett> ie. Sorting this kind of thing out is one of the reasons we were interested in a foundation.
<acuster> well, in 24hours I'll send out what I've done to the OSGeo board and see what they do with it
<jdeolive> groldan: i am more going to take stuff from mail into something more coherent
<jdeolive> and hopefully come up with some recommendations about thigns we can change
<jdeolive> to make thigns more simple
<jdeolive> the other part will be documenting
<jgarnett> let me be blunt; we are all sick of this. If you are happy with what you have we will all stand behind it.
<acuster> cholmes_, yes, although not at all for the reasons you are concoting
<cholmes_> What would you like the board to do with it? I mean, you're not a lawyer so whatever text you come up with will need to be looked at again?
<jgarnett> but we all feel "out of our depth" and pissed off.
<jdeolive> javadocs... and hopefully some code examples
<cholmes_> I'm not concoting any reasons...
<jgarnett> I see; so chris what should we be doing differently?
<acuster> yes you are. and accusing me of being the reason frank got sick of legal shit. Legal shit is shit and we are all sick of it. Frank did very well to get out. Don't blame me for struggling with it longer than others have.
<acuster> I like tOPP's aggreement a lot
<acuster> probably the most out of the lot
<cholmes_> So if we were to just get SFLC to rewrite TOPP's agreement for OSGeo/GeoTools you'd likely be happy with it?
<acuster> $5000 dollars does not strike me as a fuckload of money
<acuster> it's a bunch less than the time I have poured into this
<acuster> partly because of bad lawyers
<cholmes_> Considering the amount of fundraising that osgeo has actually accomplished it is.
<acuster> and partly because of interference between the people working on things (me) and the lawyers
<acuster> I'm even negotiating with other lawyers to help me out because of the logistics of this situation
<cholmes_> Ok, so the best route forward to me seems to be for me to contact SFLC, and get you in touch with them directly with your concerns, and then take their draft and I can advocate that to the board?
<cholmes_> WHich should be a pretty easy sell, since they are very established.
<acuster> no
<cholmes_> why not?
<acuster> I'm done. I have written, to the tinest detail, all the things I could think of.
<acuster> They are in my draft
<acuster> after that, what anyone does with it is up to them
<cholmes_> Ok, so I'll forward your draft to SFLC, and I'll do all that I can to try to make sure that all your concerns are covered by what they come up with?
<acuster> in 24hours you will get a mail, for the OSGeo board, with the results of my past weeks of work
<acuster> hopefully that goes to the OSGeo lawyers whomever they may be
<cholmes_> Well you can't be completely done, since at some point you will have to sign something.
<acuster> and some agreement comes out of it.
<cholmes_> And I would like that to be an agreement that would be acceptable to you.
<acuster> yes, i will have to read the final draft and decide if i want to sign it
<acuster> we also have to modify headers,
<acuster> decide on a policy
<cholmes_> Ok, But you don't want to give any feedback before that final draft?
<acuster> decide on how to handle contributors that don't want to sign
<acuster> what feedback?
<cholmes_> if you'll sign it or not.
<acuster> Have you looked at my doc?
<acuster> I will sign almost anything that is not:
<acuster> 1) wrong
<cholmes_> I looked at it, haven't had a chance to read it in full depth though
<acuster> 2) talk about "intelectual property" as if that were something
<acuster> so you know that my doc is a continuous feedback: this is why this is here, this is what it's trying to do, this is who we need to think about
<acuster> 80% comments and 20% text
<acuster> and yes, I do care about french law because if this company is successful at all it's going to be sued up and down by some big established players in this country
<cholmes_> Ok, I'll do the best I can, I will try to get something before foss4g but it's unlikely, as before this week I was going to be working flat out, and I just learned that I have another full time job worth of other work to also do.
<cholmes_> Which company?
– acuster holds out on that for a few months
<acuster> are we done?
<jgarnett> I guess so
<jgarnett> was that the agenda?
<aaime> Only talking about the feature model afaik
<jgarnett> looks to be; thanks again to FrankW for helping us over the last year and a half.
<acuster> gtbot stop
<gtbot> Logging stopped.
<gtbot> Posting news to GeoTools confluence...
<gtbot> Unable to post news to Confluence. Saving logs to a local location.
<gtbot> Logs saved in /home/rgould/IRC Meeting - 10 September 2007.log