1. Containers and Dependencies
2. Feature Model Update

<jdeolive> hi all
<jdeolive> meeting time?
<Jesse_Eichar77> not a great showing today.
<jdeolive> doesn't look like it
<jdeolive> lets wait for stragglers
<CGI774> hi, sorry I'm late
<jdeolive> no problem, we are just waiting for straggllers
<jdeolive> shalll we address you as something other then CG174 ?
<jdeolive> hi ian
<jdeolive> we are just waiting for stragglers
<jdeolive> does anyone have any agenda items while we wait
<jdeolive> i have 1
<jdeolive> 1) Containers, dependencies, and such
<Jesse_Eichar77> 2) Feature Model updates?
<jdeolive> sure, i can speak to that
<Jesse_Eichar77> 3) coverage branch updates?
<Jesse_Eichar77> can anybody talk about that?
<Jesse_Eichar77> ok drop that one then
<jdeolive> ha
<jdeolive> ok well i was hoping jody was going to show but doesn't seem likley
<jdeolive> so lets start, this should be a quick one
<jdeolive> 1. Containers and such
<jdeolive> doing hte massive feature model change, i thought it might be the right time to come up with a container system
<jdeolive> for managing dependencies among classes, etc...
<jdeolive> for those who are like "what is he on"
<jdeolive> check out this little doc i started
<jdeolive> http://www.geotools.org/display/GEOTOOLS/Container+Design
<jdeolive> comments welcome
<jdeolive> i might bring this one up on the list for more discussion
<jdeolive> anyone have any comments
<jdeolive> ian if you want to read we can wait
<jdeolive> if not, i dont really have anything else to say about it
<iant_> ok I've skimmed it - I think I understand what you are doing
<jdeolive> in doing the modle changeover, i am seeing a lot of code that does this
<jdeolive> FilterFactory factory = new FilterFactoryImpl()
<jdeolive> which would be nice to avoid
<jdeolive> also
<jdeolive> part of the new feature model is a two level api for building types / features
<jdeolive> its a pretty standard build wrapped around a factory
<jdeolive> so i would really like to avoid doing this
<jdeolive> FeatureTypeBuilder builder = new FeagtureTypeBuilderImpl(new FeatureFactoryImpl());
<jdeolive> but now i am ranting
<jdeolive> so like i said, probably best to send my doc to the list and see what people think and start brainstroming up a good schme
<jdeolive> any other comments
<jdeolive> ?
<iant_> I get worried anytime I see *Impl used in code
<jdeolive> none of it is really news, jody has ranted about this before many times
<jdeolive> yeah, its a definitley flag that something is going wrong
<jdeolive> well maybe wrong is the wrong word
<jdeolive> sorry for the double use
<jdeolive> but anytime an assumption is made it has repurcussions
<jdeolive> anywho
<jdeolive> 2. Feature model updates
<jdeolive> so i have been handed the slave task of transition geotools over to the new feature model (defined in geoapi) for geotools 2.3.x
<jdeolive> its going well, although it breaks pretty much every api in geotools from DataStore, to Catalog
<jdeolive> however
<jdeolive> for DataStore I beleive that 95% can remain backward compatable
<jdeolive> there will be a few breakages but it should be ok
<jdeolive> although i should probably do the deprecate for release cycle thing, so there might be no breakage in higher level api
<jdeolive> anyways
<jdeolive> the milestone i am trying to reach is getting main to compile against the new modle
<jdeolive> model
<jdeolive> after that, all the indiviudual plugins will have to be transitioned over
<jdeolive> at which point, i will need to write some docs an dask for help for
<jdeolive> sorry for the rant
<jdeolive> but that is teh skinny
<jdeolive> any qyuestions?
<iant_> How soon do you think you can reach that point?
<jdeolive> i am hoping another week, maybe 2
<jdeolive> its mostly grunt work transitioning over
<iant_> I should have some time to help by then
<jdeolive> plus i want to save the wealth of test cases we have
<jdeolive> so they have to be transitioned over to the new api as well
<jdeolive> but yes, after i get main ready to go, help will be much appreciated
<jdeolive> so if no more questions i think that is the meeting
<Jesse_Eichar77> feature model update?
<Jesse_Eichar77> hows that going?
<jdeolive> ha, did you just miss my 5 minute rant
<Jesse_Eichar77> oh
<jdeolive> the short of it is that i hope to have main transitition to the new moddel in a week
<jdeolive> maybe 2
<Jesse_Eichar77> I read it just missed the not that you had started
<Jesse_Eichar77> thought you were talking about containers
<Jesse_Eichar77> haha
<jdeolive> sorry
<jdeolive>
<Jesse_Eichar77> ok good enough for me
<jdeolive> i got a roll and couldnt stop
<Jesse_Eichar77> So from a application point of view (such as udig) the changes shouldn't be too painful to migrate to?
<Jesse_Eichar77> (assuming I'm concerned with just getting things going and not leveraging the new model too much)
<jdeolive> no they shouldn't be
<jdeolive> for a couple of reasons
<jdeolive> 1. A lot of the interaction is through higher level api (ie DataStore, FeatureStore) which we can mostly save
<jdeolive> 2. the new model has the notion of a SimpleFeature which pretty much just mirrors the old goetools Feature
<jdeolive> however
<jdeolive> there will be some grunt work changing to new classes, and changing alot of method names
<jdeolive> its not really something you can use EClipse refactor to save yourself doing
<Jesse_Eichar77> ok
<jdeolive> so there is going to be some work changing names around and stuff, but there should be nothing show stopping from a functioality point of view
<jdeolive> ie. things you could do before that you cant do now
<jdeolive> and like I said, every other api is going to remain backwards compatable
<Jesse_Eichar77> what do you thing. week of changes? another for testing?
<Jesse_Eichar77> thing=think
<jdeolive> well my first mielstone is to have main and api transitioned over
<jdeolive> which I think i can probably do in another week, maybe two
<Jesse_Eichar77> (I mean to change udig)
<jdeolive> oh
<jdeolive> sorry
<jdeolive> yeah that sounds about right
<jdeolive> week to trnasition
<jdeolive> week to test
<Jesse_Eichar77> ok cool
<jdeolive> also I am working on a transition doc
<jdeolive> so it should make your life easier
<Jesse_Eichar77> Very excellent.
<jdeolive> and it should really just be mechanical
<jdeolive> no hard thinking
<iant_> That's my sort of coding
<jdeolive> haha, i hear you
<jdeolive> how i ever got roped into doing this i dont know
<jdeolive> jody is good
<jdeolive> at least i understand his crazy new model and can actually save others the pain of getting it explained to them by him
<jdeolive> good thing he is not here
<jdeolive> haha, just kidding jody, for when you read these logs