Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 105 Next »

Geotools2 is an open source/ free java GIS toolkit. It seeks to implement OpenGIS Consortium (OGC) specifications as they are developed. This wiki is intended as a home for guides, how-tos, tutorials and the developers guide. It allows everyone to collaborate on new ideas and documentation.

If you are new to Geotools2 please read the Developers Guide.

Many of the decisions regarding future directions are disucessed and decided during a weekly IRC session. The logs of recent sessions are available below:

Could not access the content at the URL because it is not from an allowed source.

You may contact your site administrator and request that this URL be added to the list of allowed sources.

Recents news:

GeoTools Release Week

It's been a big week for GeoTools releases, with both 2.6-RC1 and 2.5.8 hitting the streets.
GeoTools 2.5.8 has seen a number of changes including:

  • Improvements Shapefile locking
  • Various fixes to ImageMosaic stability
  • Support for JDBC aggregate functions

To try out the latest release, head on over to the downloads page .

GeoTools 2.6 has finally reached release candidacy, with the resolution of a few key issues, including:

  • Solving a crash when reading ECW files
  • Updated the WCS 1.1 EMP model
  • Resolved a reader exception caused by FastBBox filters
  • The long-awaited return of the Javadoc build

For more information look to the 2.6.x branch page.

Breakout IRC Meeting

1) cleanup feedback
2) geoapi relationship
3) referencing
4) GeoTools site/blog
5) geotoolkit response

Action items:

(5:10:35 AM) aaime: rise and shine jgarnett (smile)
(5:10:43 AM) jgarnett: morning
(5:10:48 AM) jgarnett: a bit too early of a morning
(5:11:03 AM) aaime: damn timezones
(5:11:23 AM) aaime: can't you just live during the night and sleep during the day? (smile)
(5:11:28 AM) jgarnett: yeah daylight savings kicked in a while back so this is at least an hour earlier then expected
(5:11:38 AM) jgarnett: nope the point of moving to a sunny location is to see the sun!
(5:11:51 AM) aaime: aah, ok ok
(5:12:07 AM) aaime: jdeolive should be back, he's gone grab some food
(5:12:29 AM) jgarnett: okay cool
(5:12:52 AM) jgarnett: you will need to keep me from pestering him about jdbc-ng stuff; which is much more fun
(5:13:04 AM) jgarnett: I am going to do svn up and then get to work
(5:13:14 AM) jgarnett: I assume we are only clearning up trunk here?
(5:13:21 AM) aaime: yeah
(5:13:28 AM) aaime: that's what I was thiking to do anyways
(5:13:38 AM) aaime: 2.6.0 should see the light in 1-2 months anyways
(5:13:56 AM) aaime: (or else gs will have to cut it early and depend on some beta version of it)
(5:14:03 AM) jgarnett has changed the topic to: IRC Chat / GeoTools Future / Module Cleanup
(5:14:22 AM) jgarnett: I was thinking we should go with the Java numbering scheme
(5:14:37 AM) jgarnett: call it GeoTools 6 to respect both reality and the age of the project (smile)
(5:15:01 AM) aaime: lol
(5:15:26 AM) jgarnett: hey it is a good idea - you laugh at me :-P I am going to put coffee on
(5:16:27 AM) jgarnett: I am going to poke IanT who I think I see online
(5:16:46 AM) aaime: he is
(5:17:04 AM) ianturton: I'm back
(5:17:20 AM) jgarnett: ah ha!
(5:17:23 AM) jgarnett: good morning
(5:17:44 AM) ianturton: stupid notifier is quiter than the music
(5:18:13 AM) aaime: yeah, mibbit notifications are not good
(5:18:19 AM) jgarnett: stupid question
(5:18:21 AM) aaime: oh well, it seems jdeolive is not showing up
(5:18:29 AM) jgarnett: plugin\wms should really be extension\wms
(5:18:42 AM) jgarnett: except that we are starting to have "multiple" implementations of this sort of thing
(5:18:44 AM) jgarnett: wms-c
(5:19:00 AM) jgarnett: and there is a SOC student on the hook to do the other tile servers this summer.
(5:19:01 AM) ianturton: why are we having multiples?
(5:19:15 AM) aaime: well, as long as they don't share a plugin base stored in library, they are not plugins anyways
(5:19:22 AM) jgarnett: To capture all that stuff we would need a plugin mechanism.
(5:19:39 AM) aaime: but until we don't have it, it's not a plugin
(5:19:39 AM) jgarnett: indeed
(5:19:42 AM) aaime: let's think today
(5:19:47 AM) ianturton: OK
(5:19:49 AM) aaime: there's always time to move that back
(5:19:57 AM) aaime: if a plugin mechanism shows up
(5:20:00 AM) jgarnett: I will continue the move to extensions then
(5:20:08 AM) aaime: sounds good to me
(5:20:30 AM) jgarnett: but it may be wise to think up a plugin system; so we can capture some of these code contributions for the geotools library
(5:20:43 AM) aaime: jgarnett, I'll consider it when it's there (smile)
(5:21:02 AM) aaime: we have our hands full with today stuff without having to call for future/maybe stuff
(5:21:18 AM) jgarnett: some of this will fall to disucssions with you Andrea - with respect to MapContext and MapLayer. I thought I was going to have a free hand there but now both you and michael are active.
(5:21:45 AM) aaime: jgarnett, off topci
(5:21:58 AM) jgarnett: heh; repressed
(5:22:00 AM) aaime: please let's focus on today's issues
(5:22:11 AM) aaime: otherwise we won't get anything done
(5:22:37 AM) aaime: so, let's gather some topics?
(5:23:18 AM) jgarnett: sure; I was not sure if this was a meeting or a work party.
(5:23:30 AM) aaime has changed the topic to: IRC Chat / GeoTools Future / Module Cleanup: 1) cleanup feedback 2) geoapi relationship
(5:23:40 AM) aaime has changed the topic to: IRC Chat / GeoTools Future / Module Cleanup: 1) cleanup feedback 2) geoapi relationship 3) GeoTools site/blog
(5:24:05 AM) aaime: anything else?
(5:24:53 AM) ianturton: looks good to me
(5:25:12 AM) jgarnett: we have a couple other orders of biz; but perhaps they can be saved for email
(5:25:29 AM) aaime: what are they?
(5:25:40 AM) aaime: (can we talk about them here?)
(5:25:47 AM) jgarnett: we have to get a news post up with respect to new PMC members; and I see Cedric has accepted my nomination
(5:25:55 AM) aaime: Cedric?
(5:26:07 AM) jgarnett: wrong one (sad)
(5:26:12 AM) jgarnett: sleepy
(5:26:24 AM) aaime: ah, Christian
(5:26:25 AM) aaime: yeah
(5:26:45 AM) jgarnett: I almost lost his response in my email bucket
(5:27:00 AM) aaime: I don't keep mail folders threaded for that very reason
(5:27:09 AM) aaime: every now and then people answer an old thread
(5:27:26 AM) aaime: and I did loose the mail totally by keeping the folders threaded
(5:28:08 AM) aaime: voted for him
(5:28:17 AM) ianturton: Fortunately Gmail moves the tread up to the top when there's a repley
(5:28:26 AM) aaime: ah, nice one
(5:28:33 AM) aaime: thunderbird does not seem to be as smart
(5:29:10 AM) aaime: let's get started?
(5:29:13 AM) jgarnett: I have a filter set for thunderbird called "less stuff" that I have set to handle such things.
(5:29:16 AM) jgarnett: yep
(5:29:34 AM) aaime: 1) Cleanup feedback
(5:29:50 AM) aaime: Anyone has anything important/urgent about the module cleanup mails that was not addressed by mail?
(5:29:56 AM) ianturton: every thing looked good to me
(5:30:06 AM) eclesia entered the room.
(5:30:20 AM) eclesia left the room.
(5:30:24 AM) aaime: at the end of the cleanup
(5:30:29 AM) aaime: everything should have a maintainer
(5:30:42 AM) aaime: I'm just a little worried about the mif/geomedia/gpx stuff
(5:30:51 AM) aaime: CampToCamp is using those modules
(5:31:01 AM) aaime: and they told me that they are thinking about stepping up as maintainers
(5:31:09 AM) aaime: but I had no confirmation
(5:31:17 AM) aaime: so I was thinking to keep those modules on hold
(5:31:26 AM) aaime: worst case, we can delete them next week no? (wink)
(5:31:44 AM) aaime: Also it seems the javacc-jjtree plugin is hard to kill
(5:31:50 AM) aaime: because of the way the parsers weer setup
(5:31:55 AM) ianturton: I don't think mif ever worked (unless someone finished it for me)
(5:31:59 AM) jgarnett: that sounds fine; I was also worried; as I have had reports of those modules saving peoples but in the past
(5:32:07 AM) aaime: (not in a standard way, and our custom plugin takes care of the weirdness of that)
(5:32:36 AM) aaime: ianturton, yeah, mif was taken over by an italian guy some time ago
(5:32:37 AM) jgarnett: I got a bit lost on that the jjtree plugin; I thought CQL used antlr or something?
(5:33:01 AM) aaime: and spatial data integrator from CampToCamp uses mif datastore successfully
(5:33:11 AM) aaime: jgarnett, nope, antlr is used by mbedward separate project
(5:33:15 AM) ianturton: cool - I don't have a MapInfo license any more
(5:33:46 AM) aaime has changed the topic to: IRC Chat / GeoTools Future / Module Cleanup: 1) cleanup feedback 2) geoapi relationship 3) GeoTools site/blog 4) geotoolkit response
(5:33:54 AM) aaime: jgarnett, anything else?
(5:34:12 AM) aaime: anyways, I'll write a last warning mail before removing the modules
(5:34:21 AM) jgarnett: a couple stupid questions I guess
(5:34:44 AM) aaime: shoot
(5:34:47 AM) jgarnett: can we issue a milestone release when this is done; and can we run mvn site and check that the various developer roles come out correct
(5:34:55 AM) jgarnett: (and we should publish javadocs again)
(5:34:59 AM) aaime: mvn site does not work
(5:35:10 AM) aaime: I tried out, it starts spinning on itself
(5:35:13 AM) aaime: building over and over the same moduels
(5:35:29 AM) aaime: I have no idea why
(5:35:33 AM) jgarnett: can we seperatly version the maven build tools
(5:35:49 AM) jgarnett: (which also prevent us building from a clean checkout when tagged)
(5:36:01 AM) aaime: I agree on that, but wouldn't that require a restructure of the svn
(5:36:08 AM) aaime: so that we give those a separate trunk or something?
(5:36:38 AM) aaime: or you say, we just keep them there and version them separately, working only on the current trunk
(5:36:46 AM) aaime: deploying them each time they are changed?
(5:37:17 AM) jgarnett: yep
(5:37:23 AM) aaime: that could work
(5:37:25 AM) jgarnett: just do not list them as children in the modules section
(5:37:27 AM) jgarnett: yep
(5:37:36 AM) aaime: but can we keep it as a separate step
(5:37:38 AM) jgarnett: we would need to change the developers guide a smidge
(5:37:40 AM) jgarnett: yep
(5:37:42 AM) aaime: after the cleanup is done?
(5:37:53 AM) jgarnett: I was figuring this was part of the clean up
(5:38:00 AM) jgarnett: but perhaps mvn site / javadocs is too much?
(5:38:08 AM) aaime: it is
(5:38:10 AM) jgarnett: (it was going to be my acceptence test to check if the work was done)
(5:38:16 AM) aaime: if you keep on expanding the scope
(5:38:19 AM) jgarnett: okay moving on ...
(5:38:20 AM) aaime: I will give up
(5:38:28 AM) aaime: because it's no more something I can do during the weekend
(5:38:38 AM) aaime: if I don't get chunks of stuff that I can start and end in a weekend
(5:38:45 AM) aaime: I won't be able to do it
(5:39:05 AM) aaime: I'm not saying not to do things
(5:39:17 AM) aaime: I'm saying to partition them into self contained, small steps
(5:39:35 AM) aaime: alternatively, let's find out someone that can spend a week doing this
(5:39:47 AM) jgarnett: understood; agreed; moving on?
(5:39:52 AM) aaime: yep
(5:39:59 AM) aaime: 2) geoapi relationship
(5:40:13 AM) jgarnett: This one is tough for me personally
(5:40:42 AM) aaime: in the end, to me it's simple
(5:40:44 AM) ianturton: me too
(5:40:56 AM) jgarnett: a) I find the geoapi project very worthwhile professionally - it has brought me a lot of great contacts; and has severed as a good venue for discussion and collaboration with the OGC
(5:40:57 AM) aaime: do we have the resources to be involved and discuss for days and days
(5:41:01 AM) aaime: the details of some iso spec?
(5:41:05 AM) ianturton: I feel we should support geoapi but if no o ne else is why bother
(5:41:41 AM) jgarnett: b) I find it a complete failure for its origional goal. James gave up in disgust considering it hijacked by the OGC; and then by Polexis and now I am concerned it is becoming a geomatys pony
(5:41:42 AM) aaime: geomatys people are going to push more and more changes into it
(5:42:02 AM) aaime: it is
(5:42:15 AM) ianturton: but unless Deegree or OpenJump start to use it we gain nothing
(5:42:28 AM) aaime: only someone that has plenty of time to discuss the nitty gritty details of all the ISO specs they are trying to implements
(5:42:32 AM) jgarnett: that was its origional purpose; but as I said that has been a failure.
(5:42:34 AM) aaime: can keep up with geoapi imho
(5:43:02 AM) aaime: I don't even have copies of the specs, nor time to read them
(5:43:16 AM) aaime: so, at least for me, it's impossible to participate in any meaningful way
(5:43:28 AM) ianturton: I have the same problem
(5:43:31 AM) jgarnett: well I certraintly cannot keep up with it on my own; I have talked privately with Martin with repspect to moving it to a milestone release cycle; but belive me I cannot find any easy answers.
(5:43:46 AM) aaime: I saw
(5:43:50 AM) aaime: I'm still subscribed to that ml
(5:43:52 AM) jgarnett: I find people pay me to keep up.
(5:44:34 AM) jgarnett: I also find it a nice quck way to pay attention to the specs; a few developers going over them and producing interfaces. But I have worked on geoapi a lot can can do this quickly.
(5:44:57 AM) aaime: so, it seems we don't have the resources to be involved, and we don't get any significant advantage out of that, no?
(5:45:38 AM) jgarnett: I do not want to be in the position of the only "outside" contributor; I was surprised just now to see a bunch of patches showing up for an upcomming OGC meeting. So I do not feel that involved; much like with geotools I feel like decisions are being made behind closed doors ... and that does not feel good.
(5:45:53 AM) jdeolive n=jdeolive@ entered the room.
(5:45:54 AM) aaime: the same amount of effort could be spent doing code, doing blogs, integrating with other osgeo work
(5:46:41 AM) aaime: jdeolive, log of the discussion so far:
(5:47:15 AM) aaime: in the end I don't think it's a good idea to be spread that thin
(5:47:20 AM) jgarnett: I recognize this andrea; I feel we should talk the other PMC members; in particular I expect Ben's organization to freak out on this one
(5:47:44 AM) aaime: well, we don't have the resources to keep on fighting on that front
(5:47:49 AM) jgarnett: but for me personally I would be in +0 to dropping geoapi; and +1 to dropping geoapi-pending.
(5:48:15 AM) aaime: so if they freak out, they should provide the necessary resources to keep up the work
(5:48:20 AM) jgarnett: Indeed I would want to inform this decision based on our handling of the referencing module (the current scope of the geoapi jar)
(5:48:24 AM) ianturton: jgarnett - that sounds like a plan to me
(5:48:26 AM) aaime: I cannot just find time out of thin air
(5:48:46 AM) aaime has changed the topic to: IRC Chat / GeoTools Future / Module Cleanup: 1) cleanup feedback 2) geoapi relationship 2.5) referencing 3) GeoTools site/blog 4) geotoolkit response
(5:48:49 AM) ianturton: I'm going to the next OGC meeting so I can try to find out what's going on there
(5:49:03 AM) jgarnett: what time frame is that Ian?
(5:50:07 AM) jgarnett: andrea would you be okay with dropping geoapi-pending; and depending on a snapshot of the geoapi jar? So you are not forced to keep up...
(5:50:22 AM) ianturton: June somethime (endish)
(5:50:32 AM) aaime: for the short term
(5:50:43 AM) aaime: I would like to freeze the geoapi we depend on
(5:50:44 AM) jgarnett: how about until hune sometime (endish)
(5:50:49 AM) aaime: and stop changing it
(5:50:55 AM) aaime: for the next versoin of gt2 (2.7?)
(5:51:05 AM) jgarnett: (or 7)
(5:51:08 AM) aaime: I'm ok with dropping parts of geoapi and we can discuss what to do about referencing
(5:51:13 AM) ianturton: that would make sense
(5:51:21 AM) ianturton: - freezing that is
(5:51:38 AM) aaime: ianturton, yeah, that works as long as we need 0 changes on the interfaces
(5:51:55 AM) aaime: unfortunately we made the bad call of using geoapi interfaces everywhere in the code
(5:52:03 AM) ianturton: true
(5:52:04 AM) jgarnett: andrea while I feel this is a sensible response; I do not want to pressure you into anything; if Ben and the other PMC members come back and say to kill it I will not be in the way (see +0 comment above)
(5:52:05 AM) aaime: instead of using gt2 ones that did extend the geoapi ones
(5:52:39 AM) aaime: shall we move to the next one, referencing?
(5:52:56 AM) aaime: (and then come back to this one?)
(5:52:59 AM) jgarnett: let me formally propose so we can vote and move forward.
(5:53:12 AM) aaime: ok
(5:53:52 AM) jgarnett: I propose we drop geoapi-pending and depend on a milestone release of the geoapi jar until Ian reports back from the next OGC meeting.
(5:54:01 AM) jgarnett: +1
(5:54:06 AM) ianturton: +1
(5:54:07 AM) aaime: wait a second, I'm missing soething
(5:54:14 AM) aaime: how do we drop geoapi-pending?
(5:54:32 AM) aaime: by freezing the dependency on it?
(5:54:53 AM) aaime: and we keep on depending on geoapi-core snapshots?
(5:55:03 AM) aaime: Like, we're nearing a release for 2.6
(5:55:09 AM) jgarnett: fair nuff: I was thinking of depending on a milestone release; and then either folding it into geoapi (a license issue) or doing a bzr shadow copy.
(5:55:12 AM) aaime: so we should not do big api changes now no?
(5:55:31 AM) aaime: folding it into geoapi...
(5:55:41 AM) jgarnett: no the geotools api module
(5:55:47 AM) aaime: ah ok
(5:55:55 AM) aaime: so the plan really is
(5:55:58 AM) jgarnett: the geotools api module is designed to hold classes that are not ready for geoapi yet
(5:56:10 AM) jgarnett: so for me that is the same thing as geoapi pending
(5:56:30 AM) aaime: keep on depending on geoapi-core snapshots and depend on a milestone of geoapi-pending
(5:56:45 AM) aaime: mind thought, within 1 or 2 month we'll need something for 2.6.0
(5:56:58 AM) aaime: something that can be called a release, even if we are the ones labelling it
(5:57:35 AM) aaime: jgarnett?
(5:57:47 AM) jgarnett: I am catching up
(5:58:33 AM) jgarnett: so we all think the approach is "sane"; we just need to sort out the details for geoapi-pending; can we make use of a milestone release now; and come back to this another meeting.
(5:58:59 AM) aaime: imho we better start depending on a geoapi-core non snapshot as well
(5:59:05 AM) jgarnett: I intend to have us depend on geoapi milestone releases not snapshots for both geoapi and geoapi-pending
(5:59:15 AM) aaime: but we can go back to that when 2.6.0 is nearing completion
(5:59:18 AM) jgarnett: to be blunt
(5:59:40 AM) jgarnett: I want to draw a line in the sand (ie a milestone release) and then hear back from Ian towards the end of June.
(5:59:54 AM) aaime: that's totally fine by me
(6:00:03 AM) jgarnett: I think that is sufficient to allow us to proceed is it not?
(6:00:07 AM) aaime: let's package the geoapi of today as the one we depend on
(6:00:25 AM) aaime: and let's hear from ianturton in June
(6:00:38 AM) jgarnett: it is -r 1396 for reference
(6:00:41 AM) jgarnett: I will try tagging it now
(6:00:43 AM) aaime: jgarnett, would that work for you?
(6:01:09 AM) aaime: Ok, so the plan is package up geoapi as it is today, and depend on that solid
(6:01:27 AM) aaime: until we hear anything that may make us think otherwise
(6:01:30 AM) aaime: ok?
(6:01:30 AM) jgarnett: look it would work for me; and it is a responsible course of action; I just am generally a grumpy old man finding it hard to admit geoapi has failed in the goals I set for it
(6:01:46 AM) aaime: ok, +1 for me
(6:01:57 AM) ianturton: +1
(6:02:03 AM) jgarnett: we will need votes to come in from other PMC members out of band
(6:02:10 AM) jgarnett: next...
(6:02:17 AM) aaime: 2.5) referencing
(6:02:35 AM) aaime: I have some ideas here
(6:02:42 AM) aaime: one for the short term, one for the mid term
(6:02:51 AM) aaime: for the short term (2.6 timeframe)
(6:02:59 AM) aaime: we simply do nothing, keep the referencing modules as they are
(6:03:19 AM) aaime: it's not like there is anyone in a screaming need of major changes no?=
(6:03:50 AM) aaime: Opinions for this short term (lack of action) plan?
(6:04:04 AM) ianturton: sounds good to me
(6:04:14 AM) jgarnett: thinking
(6:04:50 AM) jgarnett: no that is fine; I will hear the mid term plan first...
(6:05:16 AM) aaime: mid term plan, for the 2.7 series
(6:05:39 AM) aaime: would be to find a way to get all the improvements that went into the geotoolkit referencing modules
(6:05:51 AM) aaime: without giving them a way to strangle us
(6:06:11 AM) aaime: which would be, to use mercurial to get our own copy
(6:06:12 AM) ianturton: how much of a change was there?
(6:06:16 AM) aaime: massive
(6:06:21 AM) aaime: you would not recognize the code anymore
(6:06:34 AM) ianturton: OK - is this a wait to move to java 6 thing?
(6:06:48 AM) aaime: no, I would port back their code to java 5
(6:07:11 AM) aaime: the thing is, I recognize there are improvemetns there
(6:07:19 AM) aaime: but I don't want them to set our pace
(6:07:30 AM) aaime: just like they did with Jody's patches
(6:07:42 AM) ianturton: well we could pull an update at any time we liked then
(6:07:44 AM) aaime: (1.5 year and no review, then total rewrite of it on geotoolkit)
(6:07:49 AM) aaime: that's the idea
(6:07:57 AM) aaime: clone, revert back to java 5
(6:08:12 AM) aaime: be free to use our copy of geoapi
(6:08:12 AM) jgarnett: yep
(6:08:14 AM) aaime: or our interfaces
(6:08:32 AM) aaime: be free to make our own changes without having to wait months
(6:08:46 AM) jgarnett: here is the thing; for planning. how much of an effort is that; do we need to get you a couple weeks free? what does it need to make it happen
(6:08:55 AM) aaime: and discuss for days on what is standard for them and what is reasonable for us
(6:09:12 AM) jgarnett: martin did talk to me about assisting; but I think it was a short term offer; and I have enough to handle right now.
(6:09:16 AM) aaime: right, I don't exactly know about that
(6:09:32 AM) aaime: simboss seemed to be interested in working on that
(6:09:39 AM) aaime: but sure resourcing is a problem
(6:09:59 AM) aaime: as opengeo is moving more and more to build on top of GeoServer instead of workin on the GeoTools foundations
(6:10:07 AM) ianturton: I keep meaning to get my head round refferencing - this might be my chance (smile)
(6:10:16 AM) jgarnett: so I think it is a fine mid term plan; or even short term plan if the amount of effort is low. the trick is figuring the amount of effort.
(6:10:29 AM) aaime: the code in geotoolkit seems less nightmarish than the one in gt2
(6:12:17 AM) aaime: so yeah, the idea of depending on a clone of geotoolkit works if we have resources
(6:12:31 AM) aaime: that needs to be compared with the eventual effort needed to keep the gt2 code alive
(6:12:51 AM) aaime: for very very minor maintenance I can keep up
(6:12:59 AM) aaime: I'm not worried about the math
(6:13:03 AM) aaime: but the class design is a maze
(6:13:18 AM) aaime: very deep inheritance hierarchies
(6:13:26 AM) aaime: silly tricks with SPI
(6:13:27 AM) ianturton: that's why I keep backing out of learning it
(6:13:36 AM) jgarnett: okay so as the PMC our best assistence would be to try and scare up resources in the time or money sense.
(6:13:48 AM) aaime: yep
(6:13:55 AM) jgarnett: aaime I do know the code; and have a plan to remove the silly tricks with SPI
(6:14:06 AM) jgarnett: but they have always fallen on deft ears
(6:14:30 AM) jgarnett: place that under "long term" plan
(6:14:35 AM) aaime: right, I remeber when everybody but Martin wanted to get rid of SPI and use some more modern container
(6:14:51 AM) aaime: (that was end of 2006)
(6:15:35 AM) aaime: (and who remembers the logging wars, where everybody but Martin wanted to use something other than java logging?)
(6:15:55 AM) aaime: but yeah, I agree that is more long term plan
(6:15:59 AM) jgarnett: stay on topic (smile)
(6:16:05 AM) aaime: roger
(6:16:19 AM) aaime: so it seems we have a sensible plan, we only need to scare out resources
(6:16:27 AM) aaime: let's move forward?
(6:16:31 AM) ianturton: +1
(6:16:36 AM) jgarnett: aaime; can we get a better scope on the work; that would help scare out resources.
(6:17:02 AM) jgarnett: 3) geotools site/blog
(6:17:42 AM) vheurteaux entered the room.
(6:17:43 AM) aaime: So here the idea would be to refresh the gt2 site and have a blog people do read
(6:18:02 AM) aaime: which seems a great one, but without any backing in terms of resourcing?
(6:18:09 AM) jgarnett: we could accomplish this by not posting IRC chats to our news feed; and submitting it to
(6:18:14 AM) jgarnett: or setting up a blogger channel.
(6:18:49 AM) ianturton: that might work - announce on SlashGeo too
(6:19:04 AM) jgarnett: I think we can make improvements on the blogging side; I am keen to see a website developed and am on the osgeo marketting committe with this goal in mind (specically a geotools website with osgeo branding etc...)
(6:19:17 AM) jgarnett: however the marketting committee is kind of stalled out; making logos and so on ....
(6:19:48 AM) ianturton: I don't manage to keep my own blog going but I could write to a GT one sometimes
(6:19:56 AM) jgarnett: I think I would recommend blogger; why? because the confluence RRS feed is stupid
(6:20:00 AM) jgarnett: when wiki links update old posts.
(6:20:32 AM) aaime: is blogger the bloggin site manged by google?
(6:20:51 AM) ianturton: I think so
(6:21:08 AM) aaime: I would be ok with anything that is really dead easy
(6:21:23 AM) ianturton: Blogger is easy
(6:22:02 AM) aaime: what would the blog entries contain?
(6:22:09 AM) aaime: news about new features of geotools
(6:22:11 AM) aaime: releases
(6:22:14 AM) jgarnett: we use it for udig here:
(6:22:25 AM) aaime: did not even know it existed
(6:22:40 AM) ianturton: I could tell people about the book chapter on GT I wrote
(6:22:49 AM) aaime: indeed
(6:23:00 AM) aaime: jgarnet, that's on a different site thought
(6:23:05 AM) aaime: blogspot instead of blogger?
(6:23:30 AM) ianturton: same site
(6:23:34 AM) aaime: I see
(6:23:45 AM) jgarnett: okay back on track
(6:23:56 AM) aaime: so yeah, I would be ok with that
(6:23:59 AM) aaime: what about the wiki?
(6:24:00 AM) jgarnett: even a four page "site" that links of to the wiki would be fine
(6:24:09 AM) jgarnett: it has worked well for udig (the balance)
(6:24:21 AM) aaime: would the blog become
(6:24:24 AM) jgarnett: and what is up on codehaus/sourceforge etc right now is pretty scary.
(6:24:33 AM) aaime: with direct pointers to the wiki?
(6:24:35 AM) jgarnett: ah you miss understand me ...
(6:24:46 AM) jgarnett: compare:
(6:24:46 AM) jgarnett: -
(6:24:46 AM) aaime: so ok, it would be separate
(6:24:55 AM) jgarnett: -
(6:24:58 AM) jgarnett: they are two seperate things
(6:25:03 AM) aaime: yeah, ok, family feeling
(6:25:12 AM) aaime: but how do we deal with the site/wiki then?
(6:25:29 AM) aaime: in order to better customize it we'd need to move it away from codehaus no?
(6:25:46 AM) aaime: (that's what we did for, besides a lot of issues with the site being unreachable)
(6:26:04 AM) ianturton: we should probably move to
(6:26:11 AM) jgarnett: we could move the wiki to osgeo hardware
(6:26:22 AM) aaime: but they don't support confluence
(6:26:33 AM) jgarnett: for an example of the new osgeo branding:
(6:26:38 AM) aaime: how expensive would be the migration? It would seem it would take a lot of days?
(6:26:51 AM) jgarnett: aaime it would be us who would install and support confluence
(6:27:09 AM) aaime: look, confluence is a nightmare
(6:27:14 AM) aaime: it eats a lot of resources
(6:27:22 AM) ianturton: very nice branding - if only I had a booklet
(6:27:24 AM) aaime: you sure they would allow us to install it?
(6:27:39 AM) ianturton: I think it
(6:27:43 AM) aaime: with we had to give it a dedicated instance
(6:27:45 AM) ianturton: 's been talked about
(6:27:46 AM) jgarnett: then we can leave the wiki where it is; and seperate the site from it?
(6:27:54 AM) aaime: and remove qiute a bit of plugins to avoid it going OOM every 2 days
(6:28:56 AM) jgarnett: ianturton: you are correct I do have a geotools booklet - for a tutorial for foss4g - the important part is they are starting to have more then logos so we can start to build a site around the material.
(6:28:56 AM) aaime: and keep just the guides over there? may work, not super nice, but may work
(6:29:36 AM) aaime: I'm worried about maintenance
(6:29:37 AM) ianturton: actually I may turn parts of my web mapping course into a book/let
(6:29:46 AM) aaime: I guess we'd have to at least move to osgeo all release pages
(6:30:10 AM) jgarnett: hrm lots of ideas here
(6:30:33 AM) aaime: jgarnett, usual issue: resourcing (or, lack of it)
(6:30:34 AM) jgarnett: andrea as a wikie; confluence is not working due to codehaus anit-vandlism steps.
(6:30:37 AM) jgarnett: yep
(6:30:50 AM) jgarnett: we can take steps to cut down on the amount of work
(6:31:06 AM) aaime: jgarnett, my radical opiniong would be move the main site to osgeo
(6:31:14 AM) aaime: and turn the dev guide and user guide into sphinx format
(6:31:27 AM) jgarnett: we can drop the wiki release pages; and use the sourceforge etc...
(6:31:32 AM) aaime: (there is a converter from xhtml to restructuredText afaik)
(6:31:50 AM) jgarnett: in anycase i would find this less important then
(6:31:57 AM) jgarnett: getting an initial site together
(6:32:08 AM) jgarnett: and that less important then getting blog posts picked up
(6:32:24 AM) aaime: ok for initial site back pionting to the old wiki
(6:32:29 AM) aaime: and blog
(6:32:31 AM) aaime: you do it?
(6:32:32 AM) jgarnett: yep
(6:32:44 AM) jgarnett: not sure when I can get to this
(6:33:01 AM) aaime: (I can stretch to the impossible and do some blog posts... note how many I made on the blog (wink) )
(6:33:04 AM) jgarnett: it would at least kick the osgeo marketting email list along
(6:33:28 AM) jgarnett: aaime you already send "newsy" email to the devel list
(6:33:34 AM) jgarnett: sending those to the blog would be fine
(6:34:11 AM) jgarnett: in anycase can we table this; it is a strong direction; we can take action on the blog side; and will leave the website for another meeting
(6:34:24 AM) jgarnett: 4) geotoolkit response
(6:34:25 AM) ianturton: +1
(6:34:37 AM) pramsey n=pramsey@ entered the room.
(6:35:13 AM) jgarnett: I feel I am in a good position to respond. I will compose an initial letter to the geotools-administration list; and then we can post it.
(6:35:47 AM) jgarnett: I would like to wait until after blog posts are sorted; and after the module clean up.
(6:35:58 AM) ianturton: I'm still not convinced we need to respond
(6:36:13 AM) jgarnett: will it be easier if I make this as a letter from me; as a geotools pmc member; or should I try and gather a bunch of input.
(6:36:31 AM) jgarnett: well I do feel the need to respond to martin leaving
(6:36:59 AM) ianturton: why?
(6:37:11 AM) jgarnett: we should thank him for his hard work if nothing else
(6:37:19 AM) ianturton: OK that
(6:37:31 AM) ianturton: is fair (stupid return key)
(6:38:11 AM) jgarnett: I will compose a letter; you can give me feedback before I post it.
(6:38:23 AM) jgarnett: if you hate it will not post it etc...
(6:39:11 AM) jgarnett: here is a trick question. Where do we "post" the IRC logs? I would like to avoid posting them to the news feed (since that is what got us kicked off
(6:40:06 AM) ianturton: can we just have a wiki page that links to them and point to that in the blog
(6:40:07 AM) aaime: why did that happen?
(6:40:22 AM) aaime: can't we just have the feed report the summary, the topics?
(6:40:55 AM) jgarnett: each aggregator does it differently
(6:41:02 AM) jgarnett: grabs the entire message
(6:41:09 AM) jgarnett: that is how it works....
(6:43:20 AM) jgarnett: okay I think we are out of topics
(6:43:23 AM) jgarnett: can we wrap this up?
(6:43:40 AM) CIA-76: jive * r33077 /trunk/modules/ (5 files in 4 dirs): move wms to extension
(6:43:48 AM) jgarnett: we can post this as a news item on confluence
(6:43:57 AM) ianturton: good plan
(6:44:05 AM) jgarnett: I can set up a seperate blogger thing for geotools
(6:44:11 AM) jgarnett: thanks everyone
(6:44:17 AM) jgarnett: productive meeting; if a little early.
(6:44:31 AM) jgarnett: thanks for the hard work all around

GeoTools 2.5.5 Released

GeoTools is alive and kicking; well kicking out 2.5.5 release anyways. The developers have been very busy lately (organizing due to some staff changes).

Catching up with news posts - on May 5th GeoTools 2.5.5 was released featuring:

  • Improved support for sparse shapefile
  • JDBC-NG improvements
  • Support for case insensitive like comparisons
  • WMS client timeouts
  • New filter functions that can convert types and format numbers
  • New SLD vendor parameters to turn off label conflict
  • resolution and set the goodness of fit while labelling polygons

For more details please check out the Release Notes.

Useful references for GeoTools users

There is now a GeoTools CiteULike library for useful papers, articles and other references relating to geo-spatial applications and computational geometry.

The library is freely accessible to search and browse. Items generally have keywords, abstracts and citation details plus a link to the full text where available. You can comment on an item, for example to point out errors or limitations in an article or to suggest an application.

You can also contribute your own references to the library (please do !). For this you will need to register with CiteULike.

Calling all grid coverage users.

Are you someone who:

  • Needs to create, analyse or manipulate raster data ?
  • Is often working at the image level with Java Advanced Imaging (JAI) ?
  • Would like things to be easier ?

If so, then we need you...

GeoTools and JAI

GeoTools provides a number of classes that wrap JAI operators and allow you to do such things as arithmetic operations, resampling and basic convolution without needing write to JAI code yourself. But more complex operations usually involve working with the grid coverage's backing image data directly, often using JAI image operators.

The JAI-tools project

JAI-tools is an open source project, recently started by members of the GeoTools community. It aims to extend the Java Advanced Imaging library with new image operators and utilities.

JAI-tools is not a geo-spatial project itself. All of the JAI-tools components work in image-space (ie. with band and pixel coordinates) rather than geo-space. But one of the motivations in starting the project was to support GeoTools applications working with grid coverages.

One strand of the JAI-tools project is already being tested in this context: Jiffle, an image scripting language to apply mathematical expressions to image data. Andrea Antonello has bravely embedded Jiffle into the map calculator in his development version of JGrass. See Andrea's page on this work here .

Looking for guinea pigs

Now we are looking for other interested GeoTools users to try out JAI-tools components and give us some feedback. The source code is available from the project site and is set up to compile with Maven. Please be aware that all of the is in a pre-release state at the moment and assume that it contains bugs and other undocumented features.

If you would like to provide specific feedback on the jai-tools code, the best way to do so is to subscribe to the mailing list . General comments and suggestions that you think would be of interest to the wider GeoTools community can be posted to the GeoTools users' list.

  • No labels