Agenda: 1) FeatureType revision & decision 2) GC timeline and status 3) misc.
jgarnett Hey simboss
simboss ji
simboss hi
jgarnett I was going to try and get a status update on GC stuff with respect to the road map
simboss I saw the emails
jgarnett (http://docs.codehaus.org/display/GEOTOOLS/Road+Map)
simboss but so far i have not had time to answer
jgarnett No prob
simboss tomorrow I should try to drop some lines
simboss I am going to add support
jgarnett pls glance over that page though (it is more detailed the the emails)
simboss for geotiff
simboss r&w
simboss and GriB edition 1
jgarnett I was more wanting to ask what you guys wanted
jgarnett (not specific things like GeoTIFF, but at the API level)
jgarnett Hi polio! Happy lunch break I assume?
polio yah, only have a couple min
jgarnett We have not started but, since you are here (on time) we probably should.
jgarnett I think we have answered most of my questions on email
polio yah, was in vic this weekend
simboss tomorrow I'll check that page and I'll write some lines, I am splitting my time between WCS and GridCoverage so I am very busy!
jgarnett The only nagging doubt is if FeatureType should extend some form of ListAttributeType (given the 90% overlap)
jgarnett Thanks simboss - I would like to put you on the agenda today, because I don't want to cut your work out of geotools 2.1 deliverable, if it only going to be a diffrence of weeks.
polio well, you can just make featureattributetype extend listattributetype
jgarnett (Or the other way, include your work and stress you out)
jgarnett If you think I should do it then I will
simboss deadlines are stressing me
simboss not you!
jgarnett I am reading the OGC Feature overview document now
jgarnett (Geotools 2.1 deadline is at least two months away simboss)
simboss things I have been talking MUST be ready in two weeks!
simboss deadlines for my job!
polio ouch
jgarnett Okay, I got it - we will talk to you in two weeks.
simboss
jgarnett simboss if you need help we may be able to ask in todays meeting.
jgarnett Others around the world want this stuff too (ask for volunteers).
jmacgill be right with you
jgarnett Hi James we have two things on the agenda:
jgarnett 1) FeatureType design decision time with polio
jgarnett 2) simboss needs help (and status update on gc)
jgarnett So back to agenada item #1 ...
simboss well, the only thing I might ask is to check why some tests fail on the rack while they do not fail on any my machines
jgarnett You are cool so far polio?
polio yah, think so
jgarnett Okay simboss (can we get back to that in a second, after topic #1?)
jgarnett Okay .. I will keep on keeping on then.
simboss cool
jgarnett I am trying to find evidence I can throw out multiple inheirtence.
jgarnett Other than the fact that both polio and I don't want to do it.
jgarnett If I can get rid of it I would like to be explicit with a getSuper() method.
polio well, SFS and gml don't support it
jgarnett Thoughs polio?
jgarnett That is almost good enough for me.
jmacgill is back
polio I would aproach this from the other direction, and ask the questyion as to why we should include it
jgarnett For completness
jgarnett this OGC doc of over view doom seems to like it
jgarnett (expresses things in terms of feature vews of some real world entity.
polio yah, in a perfect world we would want them ... use them for styling
jgarnett I would also point out that GML allows sequences to be included, a form of "mix in" if I translate to OO land.
polio but since we can't really produce them
jgarnett I also found the difference between AttributeType and PropertyType
polio yes, but the sequences still map to a single inheritance path
jgarnett in OGC Feature talk property includes opperations and attribtues.
jgarnett Other question was if AttributeTypes need a super?
polio types or type?
jgarnett Type needing a AttributeType.getSuper()
polio type doesn't ... if we assume that it is either simple or complex
jgarnett AttributeTypes meaning the whole category of 'em.
jgarnett And what happens when I think about Feature Type A.attr1 extended by B.attr1
polio then in the simple case you can always flatten
polio in the3 complexcase ... you have a featuretype anyways
AlFa796 Hi all by Alessio
jgarnett I would think B.attr1 almost has A.attr1 as a "super"
jgarnett Hi - will send you the agenada offline.
AlFa796 thx
polio umm, inheritance will take care of that
jgarnett Okay cool.
jgarnett I think that is pretty much my set of design questions left over from email.
polio (just do the diff between the parent and the current to find the inherited fields vs current fields)
jgarnett I will bang that out, and ask for a review mid week.
polio think they also have a namespace that can be used for this
polio cool, just send it to the list
jgarnett ? error parsing your last sentence ?
polio (send the review to the list ... I will comment if you want)
jgarnett by namespace do you mean their silly "GF_" naming prefix?
polio nope
polio though I could get at the xml namespace (same as ft) in attributetypes ... but can't recall atm if that is posible
polio but thinking more i'm also not sure it's appropriate
polio well, I should get back to it (ussually lunch is longer ... but we lost out internet a bit back)
jgarnett no prob.
jgarnett Thanks for attending, ..
jgarnett we should probably but another item on the agenda
jgarnett 3) what happend with the votes from last week
jgarnett
polio np, ttyl
jgarnett 2) simboss needs help (and status update on gc)
jgarnett Translation simboss has a deadline, and rather than ask for a status update
jgarnett with respect to the road map.
jgarnett We should ask what, if anything we can help him with.
jgarnett Over to you simboss.
jgarnett So far I got investigation in to rackmount based failure.
jgarnett (considering it is a "new" rackmount the problem may be anything)
AlFa796 I think Simone is away for a while
jgarnett Well that makes me talking to him rather tough ...
jgarnett .. we may have a record short meeting.
jmacgill ok, I'll mentiona problem I found with M4 then
jmacgill validation depends on experiment which was not included
jgarnett I have an anoucement of the surprise cholmes and dblasby kind
jgarnett Why would validation depend on experiment?
jmacgill no idea
jgarnett Oh right people moved the Repository classes there
jgarnett we moved em back, but the project.xml file does not know.
jgarnett Fixing it now.
jgarnett (that kind of cans 2.1.M4 being a match with what is behind GeoServer 1.3.beta-3 thou.
jgarnett Well the point is to learn for next time ...
jmacgill the binary version works
jmacgill but the src version won;t compile
jmacgill (validation that is, the rest is ok)
jgarnett Too bad I did not catch that ...
jgarnett James we also had a javadoc publication request
jmacgill need to empty local repo
jgarnett and the README was not generated with the correct subst for date, version, modules.
jmacgill yeah, working on it
jmacgill both javadoc and readme that is
jgarnett cool
jgarnett what about empty the local repo?
jgarnett before nightly builds?
jmacgill no, before attempting to build the src version of a release
jmacgill you will have had an M4 version of experiment installed in your local repo
jgarnett oh I see
jgarnett We also ran into some repo problems with lists.refractions.net
jmacgill assuming you did try to build the src version (i.e. unzip the generaed archive to a clean directory then try building it)
jgarnett That must be why I missed it ...
jgarnett here is the deal: http://lists.refractions.net/geotools/gt2
jgarnett is not "normal"
jgarnett it should have a jars directory.
jmacgill but is must not be
jgarnett with all relesaed jars in it: 2.0.0
jgarnett 2.1.M4
jgarnett but not 2.1.x
jgarnett That is other maven projects may want to use geotools ?
jmacgill yeah
jmacgill they can use
jgarnett Right now other projects are doing magic with SNAPSHOT ...
jmacgill http://dist.geotools.org/geotools
jgarnett or just copying everything like 2005-Mar-22 or UDIG1.0.RC4
jmacgill dam, missed
jmacgill hang on
jgarnett lol darn whily urls
jmacgill http://dist.codehaus.org/geotools/
jgarnett What is that dir?
jgarnett Oh I get it ...
jgarnett should we move of lists.refractions.net?
jgarnett Or is this a case of that where you want our geotools "releases" to go?
jmacgill releases and deps must be seperate
jmacgill or builds will get very very confusing
jmacgill so we keep lists.refractions.net (aka repo.geotools.org
jgarnett James can you update the releae guidelines with instructions on how to publish content there?
jgarnett We can change the uDig and GeoServer ant scripsts to fetch content from there as well I guess?
jgarnett Okay james is quiet
jmacgill yes, its new
simboss sorry guys, I am back now...
jgarnett I think that may be it for 2.1.M4 feedback?
jmacgill I'm going to be sending cruisecontrol builds there too
jgarnett Shall we return to agenda item #2?
jmacgill cool, good timing
jgarnett (any chance of javadocs from CC?)
jgarnett So simboss
jgarnett agenda item #2 was basically the following:
jmacgill jgarnett probably
jgarnett Q: Does your GC work want to be included in Geotools 2.1 (aka is the API good enough) in around 1.5 months?
jgarnett Q: You have a deadline, where are you stuck? How can we help ...
simboss I absolutely think so
jgarnett (or even if we can help).
jgarnett Cools that is good enough for PMC to plan with ...
simboss i am going to stop coding new plug in in a week
simboss then me and alessio
simboss will start testing more and more
simboss my plan is to be done
simboss in 3 weeks
simboss at least at a decent stage
simboss because I need those plugin for a decent version of the WCS
jmacgill what renderers can rasters be displayed in?
simboss for the moment
simboss arcgrid
simboss world images
simboss gribs
simboss geotiff
simboss and gtopo30
jgarnett On our side simboss we can place GC support back in uDig (with those formats) and generate a bunch of bugs and testing for you?
jmacgill re-read my question (though that answer is helpfull too)
simboss (well with except for the tiles from antartic)
jgarnett I think the question was?
jmacgill i.e. lite, lite2, j2d
jgarnett Does LiteRenderer2 work, or are you doing something magic?
AlFa796 actually I'm using LiteRenderted2
jmacgill can you post a demo or sample code on how to get a raster up in lite2?
AlFa796 it works very well except for raster symbolyzer (a task that I'm going to do as soon as possible)
jgarnett I trust we would also ending up debugging RasterSymbolizer on our side as well.
jgarnett So if we can get support in uDig again we can both work on RasterSymbolizer together...
AlFa796 LiteRenderer2 works ... simply I build up a Feature with an attribute called "grid"
AlFa796 cool
simboss for the moment we made it wrok as it was
jgarnett (There is a convience method that does just that, simply "add" a GridCoverage and it packages for you).
simboss but from july
simboss we will try to enhance it
simboss addisng support for pyramids and imporivng performance if possible
jgarnett You may need to hack format.
jgarnett Format was done rather strangly in order to support single use when streaming from internet.
simboss (i am being dislessic sorry! )
jgarnett You may want to hack in a random access in order to choose the correct level of the pyramids.
jgarnett (I am not the fellow to call you on it)
simboss we will work extensively on that
simboss because we will have to support very large images for the WMS
simboss for example side scan sonar images
simboss which are like 1G each
simboss we definitely need tiling and pyramids
jgarnett Here is the deal, we have never had any "review" or "revision" on Format.
jgarnett Or any of the GCE API.
jgarnett If you need changes, please go ahead and make them.
jgarnett We can even clear out the GCE API and leave support at the format level.
jgarnett And after your deadline, any feedback you can give us would be great.
simboss actually we have some ideas
CIA-10 jeichar * r13358 udig/plugins/ (2 files in 2 dirs): fixed keybinding scope issues
simboss especially about coverages with more than 2 dimensions
simboss but for the moment
simboss we stopped looking at this issues
jgarnett Okay, if possible I would like to have that discussion with you.
jgarnett Before we hit Geotools 2.1.0
simboss well, I will try
jgarnett (because after that it will be a lot more work to change API to suite your needs).
simboss i have first to gather these ideas
jgarnett I understand.
jgarnett James does this sound workable?
jmacgill yes, very
rschulz hello
jgarnett Hi
rschulz jody, when will svn be back up?
jgarnett It is down?
AlFa796 Do you think is good for you if I start to take a look at the LiteRenderer2?
jeicha1 I'd say that's fine.
rschulz well, I cannt get to it (with svn client or through a web browser)
jeicha1 (Answering question about Lite2)
jgarnett I see
jgarnett can you do the control-refresh?
AlFa796 My intentions are to update the GridCoverageRenderer to support tiling and pyramidalization as far the formats will support that
jgarnett We got a faster machine, so the ip change takes a while to get around.
jeicha1 That'd be great. You know MUCH more about that than I do.
rschulz jody, ok, what is the ip address for a host file
AlFa796 We must even synchronize with the RasterSymbolizer, I have not yet understand the state of that
simboss the point is
simboss the work we have done so far was to get acquainted with geotools GC and geosever
simboss we are trying to get a decent version of both
simboss after our own deadlines will be met
simboss we will start to really improve things
simboss and the rendering side
simboss and we will work and adding more formats
simboss I talk
jmacgill ok. My main interst for now is to include some form of raster capability in 2.1 even if it is not complete
simboss we have that
simboss because we have a decent version of a WCS which is built on top of it
jmacgill yeah, cool.
jmacgill very cool actualy, raster has long been neglegted
simboss for the future I also talked to frank warmerdam (the guys who wrote gdal) he is interested in porting his coverage work to java, we might exploit that as well!
jmacgill better and better
jgarnett Well we have kind of chewed through most of our agenda
jgarnett I was talking to james ealier and he is on track for FeatureCollection / FeatureResults merge.
jgarnett (apparently with a bit of inspiration from JDO)
jgarnett Oh and when dblasby and cholmes are back in the loop they can check out this:
jgarnett http://docs.codehaus.org/display/GEOSDOC/User
jgarnett (Aka a wiki version of their docbook documentation)
rschulz question - when did you guys start the meeting (wiki says 1:30 pdt)
jmacgill erk, meeting time moved by an hour a few weeks ago
jmacgill wiki clearly out of date
rschulz ah, sorry for my interuptions above then
jmacgill ? not at all.
jgarnett Um I update he wiki when we did the time change?
jgarnett Please correct wiki, perhaps I only announced it.
jgarnett So Should I package up the logs? The meeting does appear to be over?
jgarnett Oh hi chris.
chodgson hi
jgarnett Hey rueben chris here is looking at CRS stuff
rschulz jody, do you have an ip for svn.geotools.org?
jgarnett and may have some questions for you ...
jgarnett 64.30.100.236
rschulz ok, shoot (I do not guarente answers thought)
jgarnett 12ms ping time.
chodgson well, I don't exactly have any questions at this point... only problems that I'm trying to figure out... but Jody wanted me to show up and say hi...
jgarnett Yep I am so bad, introducing people so they don't just meet via jira bugs.
chodgson if/when I do have questions, now I'll now who to fire them at...
jgarnett sorry rueben :-P
rschulz you can fire questions at martin also
rschulz thats ok jody
jgarnett rueben did you find svn yet?
rschulz no, still re-learing vi
chodgson does svn.geotools.org not resolve for you? I don't know why it wouldn't...
chodgson hey, I'm an IRC newbie.. there's no way I can look at the conversation before I joined the channel, is there?
jmacgill fraid not
rschulz logs will be posted on the site
chodgson that's what I figured
jmacgill we should get someone to set up a logging bot
jgarnett I can do that now if we are done the meeting/
rschulz jody, I can now access svn. Had an old entry in my hosts file from a long time ago
chodgson huzzah!
rschulz (last svn update (laptop) was at christmas; this is going to take a while)
jgarnett lol