1. OSGEO
2. Filter / Expression API Transition
3. Quality Control
<iant_> I'd like to add open source geosatial foundation discussion
<jdeolive> 1. OSGF
<iant_> I think we're prefering OSGEO
<jdeolive> ooops, sorry
<jdeolive> 1. OSGEO
<jdeolive> 2. Expression / Filter API update
<jgarnett> Yo.
<iant_> hi
<jgarnett> (hey cool - cameron is going to crash our meeting)
<jdeolive> cool
<jgarnett> We are having a meeting right ...
<jdeolive> current agenda:
<jdeolive> 1. OSGEO
<jgarnett> ... you guys are all not waiting around or anything.
<jdeolive> 2. Expression / Filter API update
<jgarnett> sweetness x2
<jdeolive> we were wating a couple of more minutes for stragglers such as yourself
<jgarnett> I got a annoying thing on the user list, so I am going to put on ...
<jgarnett> 3. Quality Control
<iant_> I saw that too
<jgarnett> Hi hobu
<jgarnett> 0. Everyone welcome hobo 
<jdeolive> welcome !!
<jgarnett> So um, hobo - I am a spatial hacker - this is the 12 step program
<jgarnett> who are you?
<hobu> you don't know who I am?
<iant_> foundation spy
<hobu> hah!
<jgarnett> (we are teasing you)
<jgarnett> As long as you are here you can review mhy blog post; I thought I should respond to the java representation thang;
<jgarnett> And you are warned that my writing involves a spelling based drinking game.
<hobu> trying to see how closely geotools discusses the foundation stuff in relation to the RFC I put togetther for mapserver http://mapserver.gis.umn.edu/development/rfc/ms-rfc-10/view
<jgarnett> http://weblogs.java.net/blog/jive/archive/2006/02/_open_source_ge.html
<jgarnett> sweet - well lets move into ageanda item 1 shall we?
<hobu> ie, how does a "coalition of the willing" mandate the movement to anything/anywhere?
<hobu> (US military reference just by coincidence
<jgarnett> easy, if you don't interoperate you look bad
<jgarnett> good old fashion compitition on speed, usability never hurt anybody. Compitition with gvSig has done wonders for our index shapefile support.
<jgarnett> Justin can you herd us along today? There is so much exciting stuff going on I need somebody to cut me off before we run out of time
<jdeolive> alrighty
<jdeolive> 1. OSGEO
<jdeolive> iant_, i beleive the floor is yours to start
<iant_> sorry was just reading Jody's ramblings
<iant_> As you are probably all aware I went to Chicargo on Sat and was involved in disscussing the beginings of the OSGEO foundation
<jgarnett> (ow that hurt, oh wait you have to read my source code as well)
<iant_> (as a side note I have no C key on this machine so they may be missing occasionally)
<iant_> The foundation ame out of some mad plan hatched by Autodesk and the mapserver folks.
<jgarnett> (as you can see hobo the java community is high tech, we are just moving away from punch cards, a few people are confused and pop out keys during the transition)
<iant_> Initially this was poorly received so a second plan was made
<iant_> thus more people were invited to attend including geotools
<iant_> hang on phone call
<jgarnett> I see - have you guys had a chance to read the various material, iant_ email and so on ?
<jgarnett> Does anyone have any questions so far?
<iant_> I'm back
<iant_> sorry about that
<iant_> There was lots of discussion! We chose a name the Open Source Geospatial Foundation (OSGEO)
<iant_> the basic outline of the foundation is based on the Apache model and we had Brian from Apache and Collabonet along to advise us
<iant_> currently everyone who attended the meeting in person are members of the foundation and we are looking for nominations for the next 20 members
<iant_> when this group is sorted we'll elect 4 more board memebers to add to the 5 elected yseterday (includes our own ChrisH/cholmes)
<iant_> The foundation is now looking for projects to support -
<iant_> this is where we come in
<iant_> while exactly what we'd get out of membership is not entirely clear yet, we would get hosting for wiki, mailing lists, svn etc
<iant_> we would also get geotools.osgeo.org as a domain name
<iant_> the idea would be for some sort of common look and feel between projects (like on the Apache site)
<jgarnett> I am more interested in an IP check, common branding, and visibility.
<iant_> but that each projet would keep its own pmc and management style
<iant_> we would need to carry out an IP check and assign a right to redistribute to the foundation but not the actual code
<jgarnett> (does not seem like a burden since we are LGPL?)
<jdeolive> coming up on meetin half time
<iant_> if we give the foundation joint ownership of the code they can help defend us on legal challenges
<iant_> OK we all need to think more on this - there is lots of stuff out there to read. If people are blogging or come across useful stuff can they mail the list
<iant_> I would like to see us join but I don't want anyone to feel pushed in before we know anything
<iant_> quick questions? or should we move on?
<iant_> anyone out there?
<jgarnett> hint: PSC is Project Steering Comitee - aka PMC
<jgarnett> (but apparently friendly)
<jgarnett> iant_ you were there, it was hard from IRC logs to tell what was going on all the time.
<iant_> indeed - it was hard to keep track while being there
<iant_> much is still to be decided by the interim board and members
<jgarnett> Did they decided to go big with respect an open source push, or did they leave that on the table?
<jgarnett> (aka delegate to the new board)
<iant_> the idea is to get a web site/wiki and mailing lists up soon
<iant_> a big open source push?
<CameronShorter> I'm ex-geotools developer, and now with Mapbuilder. We decided to join because we would like to see an Open Source GIS body which promotes integration and can present a united front to compete against the big proprietary products.
<Polio> So, I may have missed this part ... but I was wondering about the branding portion of the proposal
<Polio> I was hoping we could have a short round table on this
<iant_> there will be a brand and a logo which if geotools joined we'd be able to use
<iant_> I don't want to use the whole meeting for this - but I can hang on after for further discussion
<Polio> I was more concerned with the required portions, not the optional ones
<jgarnett> Good questions, I am not sure what the required portions would be. I imagine for legal purposes we would need an IP check.
<iant_> the requiement would be to make all of the different project websites look alike -
<iant_> in the same way that each apache site has the same look and feel
<Polio> Ok, given that requirement ... I'm tending between a 0 and -1 vote.
<iant_> you know where to go to find the download
<Polio> I'm VERY concerned about the branding ... ie being swallowed up
<iant_> If we're first in we get to describe the web site shape to an extent
<jgarnett> So far I have been the confluence dude, I can certaintly do the work if that is what people want.
<jgarnett> Polio are you worried about losing the geotools brand?
<iant_> We'd always be geotools - in time we'd become osgeo-geotools
<Polio> Um ... I'm more concerned with looking like a single dot in the field
<Polio> Yes Jody
<iant_> I'm not sure what you mean polio
<jgarnett> The flip side of this is a race, aka Deegree is looking to matter, and they could easily step in.
<iant_> this is more a case of an oppertunity to get our name more widely spread
<Polio> In trying get the right wording ...
<jgarnett> I understand what he is getting at iant_, the apache commons library is great, but it is not a strong brand on its own. I think apache is different because they started with one good application and bred others.
<iant_> especially with people who currently look to the C stack (i.e. mapserver)
<Polio> I'm worries we would be lost in the sea of projects ...
<CameronShorter> Polio, I think the Geotools footprint would increase by assotiation with the Foundation rather than be swallowed up. Visability of Geotools would increase with association.
<iant_> I guess there is a risk of that - at present there is always a risk people don't really know the difference between udig, geoserver and geotools
<jgarnett> osgeo has several really good brands going in, I trust our brand. What I don't want osgeo to get in the way of these projects competing.
<jgarnett> Which tends to be assumed when you watch the apache project turn.
<Polio> I'm not fairly sure you're right cameron ... but I wanted us to think about this
<jgarnett> when they bring out somethiing new in the same space, you assume it is to replace the last one.
<iant_> I think the GT brand is strong. There is no sign of another java mapping toolkit comming out of the foundation,
<Polio> Jody, hadn't though of it exactly like that (just the consequences)
<CameronShorter> Polio, you may also have similar concerns that we had regarding loosing control? We at Mapbuilder were satisfied that our PMC would retain control of our project.
<iant_> they are more concerned to have a java presence currently
<Polio> Cameron, from that angle I wasn't worried due to our current PMC model and licence
<Polio> I'm looking at this from a maketing perpective, and trying to see issues a few yearrs down the line
<jgarnett> I can think of several good candidates, Deegree is more consistent then us and has a more complete stack, gvSig looks generated and the docs are in spanish but their performance caused the indeced shapefile work.
<Polio> *marketing
<iant_> I think there is always a risk of a library project ending up with a lower profile than a sexy application
<jgarnett> I think Polio is concerned about losing eyeballs
<Polio> eyeballs?
<iant_> I think that is a greater risk if we miss this chance than if we take it.
<iant_> this is a chance to be on the same website as mapserver gdal ogr mapbuilder mapbender ossim etc
<Polio> Ok, well I have more to think about atleast (thanks for everyone putting up with the devils advocate here)
<iant_> cholmes thinks geoserver would probably follow gt in
<jgarnett> I agree iant_ but we should watch how things progress. My concern is that the foundation will be busy on infrastructure and not develope a strong marketting presense
<jgarnett> (although the budget thread does mention attending convernces and promotions as priorities)
<iant_> I think a 250k$ budget should lead to quite alot of marketing
<jgarnett> depends hardware, and bodies cost a bit. They mentioned a person on staff...
<iant_> since the foundation is buying in most of the infrastructure direct
<iant_> there is talk of building an SDI testbed
<iant_> which could get a lot of attention
<jgarnett> In anycase we should progress with the rest of our meeting, I have already cast my +1 vote. I would like to see everyone vote on this if we can manage it.
<iant_> If the foundation does appoint a director they would have to mostly market it
<iant_> I really want to hold off voting untill next week so we can have more discussion
<iant_> plus by then we should have a better idea of whats what
<Polio> Yah, a week to mull it over, ask question (and hopefully have a few lively debates) would be nice
<iant_> there is no rush
<jgarnett> hobo do you know of a timeline?
<jgarnett> aka 2 weeks was mentioned, if that to get in on the gound floor anouncement?
<jgarnett> Last time this was a problem, and I don't care much if geotools is in the first draft or gets its own press release later. But a question of timing is no doubt important for you and the MapServer lot.
<jgarnett> Okay, Ian you were there. Timeline?
<iant_> none yet
<jgarnett> If not I will assume the two weeks as mentioned in the IRC channels.
<iant_> everyone basically had to go back to thier projects and talk it over
<hobu> the "competition" issue is a contentious one for Mapserver as well (re: MapGuide).
<jgarnett> Understood, question for you
<jgarnett> does MapGuide do WFS-T ?
<hobu> mapguide doesn't do squat for OGC at this time
<jgarnett> Really? I wont ask you what it does then ...
<hobu> It renders pretty.
<jgarnett> It may be fun if they are the only player that does not talk to all the others. Do they have plans?
<jgarnett> well we can take this offline ... next agenda item please
<hobu> I think they will agressively particpate. Their software already takes in quite a bit of the "c" stack
<hobu> sorry
<jgarnett> no worries you type fast around here - or someone else speaks.
<jgarnett> 2. Expression / Filter API update
<jgarnett> jdeolive - how is it happening?
<jgarnett> Do you want me to do the geoapi update first?
<jdeolive> its going good
<Polio> I don't see martin here ... but I would like us to think about our involvement with geoapi ... and how we would be situated (ie, I didn't see them included in the proposed set of projects, and should this impact us if we joined).
<jdeolive> i sent some mail to geoapi-devel, if youcould answer that woudl be nice
<Polio> oh, sorry I'll go to email
<jgarnett> I could not build you branch - geoapi-2.1 was mentioned. Can we build based on datestamped snapshots please?
<jdeolive> sorry Polio, that was for jody
<jdeolive> your concern is defintley valid
<jgarnett> Polio - I asked about GeoAPI and was shot down
<Polio> That was why I though we might want to talk about it
<jgarnett> They kinda felt it was under OGC control, I did not have the heart to mention that James/Martin started it and I was willing to take it back
<Polio> do we need to worry about opposing presures as it were
<jdeolive> i uploaded the timestamped jar
<jgarnett> Well I said flat out that GeoAPI fit the osgeo interoptability mandate to a T, and I would like to see it included.
<jgarnett> I better email that list, see what happens. Would that help polio?
<Polio> Yes thankk you
<jgarnett> Polio the other pressure is that geoapi does not move, if this experiment of justins does not go ahead we will need to formally split ways with geoapi on the Feature/Geometry etc... issues.
<jgarnett> emailing now.
<iant_> I saw no problem with geoapi joining foundation but since they are more closly coupled with the OGC they might not need to join
<Polio> I was just concerned about being the implementation pawn between two interface projects in the future, and what that could do to the project
<Polio> Jody, I'm fine with that ... so long as we have a longterm plan
<jgarnett> I do, but it is not always politic so I have been keeping it to myself.
<Polio> But I should run, thanks all for the productive banter (feel much better about this proposal now)
<jgarnett> later, thanks Polio
")
<jgarnett> jdeolive - anything else.
<jgarnett> when do we start features?
<jdeolive> hopefully this week
<jgarnett> ack - okay I can code up the interfaces tomorrow night.
<jgarnett> (I thought I had till the weekend)
<jgarnett> can you move us to the last agenda item ....
<jdeolive> you probably do, i have some preparation to do for my trip so I am not really full time this week
<jgarnett> t 3. Quality Control
<jgarnett> iant_ what was your take on the user lists?
<jgarnett> Not sure everyone is signed up let me summarize:
<iant_> he seemed unhappy
<jgarnett> A group from itally tried usin geotools, in particular OracleDatastore. Patched it to death, got no support and went away.
<jgarnett> I am not impressed, indeed I am not impressed with a few of the datastores that have had blockers for a couple releases now.
<jgarnett> So how can we do better?
<iant_> there are some problems with oracle as it has no real maintainer
<jgarnett> I see so this is a toss it out of the build, or PMC hunt for someone to maintain it situtation?
<iant_> but he seemed unhappy with the whole datastore stack (especially jdbc*)
<jgarnett> I need it and I can fix it, but JDBCDataStores are a mess, as generations of developes on a quick project insert optimizations.
<jgarnett> His opionon is justified.
<iant_> unless we can find someone who actually uses oracle it is hard to look after
<jgarnett> Understood, my time would be better spent cleaning up JDBCDataStore so that maintaining the silly things is not so hard.
<jgarnett> If we can find a volunteer let me know, I would love to tag team the problem and get a solution that would last.
<jgarnett> But we still have a quality issue, we have metrics (60% test coverage) which so far only Justin has met.
<iant_> plus we tell people how to build with out running the tests
<iant_> I think the main trust was we should actually fix some bugs before we add more interfaces etc
<jgarnett> And we have been unable to follow our own code formatting. Yes we are not doing so good 
<jgarnett> That one cuts close to home, and I agree. But I cannot fix some of the problems the way things are laid out now.
<jgarnett> ( I am trying to work through the stack, we did filter, style, and we are on filters)
<jgarnett> data store will probably be next on my list ...
<iant_> I have more time now so I'll see what I can do as well
<jgarnett> that would be great, one of the best things we could do (as PMC even) is help the module maintainers get thier Jira lists under control. Jira is full to the point of being ignored...
<iant_> indeed
<jgarnett> Any ideas would be good, the other is not include modules in the release that are not ready (even if it means we take a capabailities cut publically)
<jgarnett> I don't mind open development, and if we drop some datastores due to lack of a module maintain I am okay with that.
<jgarnett> In uding we call these "unsupported" plug-ins...
<jgarnett> uding = udig (sigh - it is late sorry).
<iant_> I think that might be better than releasing with bad code
<jgarnett> We have done it before with our legacy module, and we can put a positive spin on it ( vounteer and make this a supported plugin).
<jgarnett> But please if anyone has any ideas we need help.
<jgarnett> (my other idea is kill the PMC and return to module maintainers - when keeping a module working is your reason to be here we may get better results)
<iant_> works for me - I get to go home
<jgarnett> Sorry for the doom and gloom guys, meeting attendence is up (and meetings are shorter), and our api is getting much better one layer at a time.
<iant_>
<jgarnett> heck we would make you module maintainer of main (that will keep him out of trouble)
<jgarnett> seriously iant_ cholmes is module maintainer of several orphaned modules - even just helping him with Jira would be great.
<iant_> I'll try to make a start on that
<iant_> now I've finished moving I have much more time
<jgarnett> Are we done?
<iant_> jody can you mail me those geoserver transaction docs?
<jdeolive> I think so
<jgarnett> confused iant_ what do you mean?
<jgarnett> as in how to do it? Geoserver is normal struts there is a scary file somewhere in there with all the names broken down...
<iant_> you said you had some review docs on transactions ( I think it was you)
<jgarnett> ah transaction (sorry very late)
<jgarnett> you know that blog entry - it is in there
<jgarnett> http://weblogs.java.net/blog/jive/archive/2006/01/geoserver_wfst.html
<jgarnett> I think you will be the second person to read it
<iant_> thanks
<jgarnett> you say that now
<jgarnett> please send questions to geoserver devel
<jgarnett> hobo - did you want to chat. Meeting seems over.
<jgarnett> How is the MapServer community doing, same concerns as here?
<jgarnett> Hey justin did you want that blog entry for your documentation
<jgarnett> jdeolve - replied to your geoapi email...