Thursday, January 22, 2009

uDig IRC Meeting January 23rd

0) what is up
1) 1.2-M2 timeframe
2) raster bug
3) projection question
4) style editor question

Action items:
- jgarnett: export style extension point for moovida to test
- moovida: turn UDIG-1469 into a table of sample rasters to test
- jgarnett: hunt down imageio-ext project and see if a breakout QA session can be arranged

(7:12:02 AM) moovida: oh, ok, just wanted to hear how your John Doe bug is doing :)
(7:12:13 AM) moovida: Mr. Garnett, welcome!
(7:14:20 AM) moovida: can we begin?
(7:14:31 AM) mauricio: ok
(7:14:44 AM) moovida: do we have an agenda?
(7:15:18 AM) jgarnett: morning
(7:15:23 AM) jgarnett: I am not sure we do yet
(7:15:25 AM) emily_g: I have a projection question that is probably better suited to geotools; but I thought I'd ask here to see if anybody else has come across it
(7:15:26 AM) jgarnett: but we can start one
(7:15:41 AM) moovida: yep
(7:16:21 AM) moovida: jgarnett: you do the moderation?
(7:16:33 AM) jgarnett has changed the topic to: 0) what is up 1) 1.2-M2 timeframe 2) raster bug 3) projection question
(7:16:36 AM) jgarnett: sure
(7:16:42 AM) moovida: great
(7:16:46 AM) mauricio: ok
(7:16:55 AM) moovida: 4) style editor question
(7:16:58 AM) jgarnett: anyone else got a subject for the agenda?
(7:17:09 AM) jgarnett has changed the topic to: 0) what is up 1) 1.2-M2 timeframe 2) raster bug 3) projection question 4) style editor question
(7:17:34 AM) jgarnett: okay let us start
(7:18:05 AM) jgarnett: 0) what is up (type in a line saying what you are working on; the idea is to be social and notice when two people are working on the same thing)
(7:18:19 AM) emily_g: I'm poking away at a wfs feature cache using the geotools unsupported gt-caching module
(7:18:49 AM) jgarnett: jgarnett - I am thinking about geotools involvement; trying to find more work; and looking into how raster support works in 1.2 if I get time or funding
(7:19:14 AM) mauricio: Mauricio: more spatial operation, unit test and learning about udig code base
(7:19:43 AM) moovida: Andrea Ant.: back on udig - postgrid/constellation connection
(7:20:26 AM) silli [n=chatzill@] entered the room.
(7:20:36 AM) jgarnett: fun fun
(7:20:42 AM) jgarnett: 1) 1.2-M2 timeframe
(7:20:56 AM) jgarnett: Just a quick note; I said I would get to this
(7:21:13 AM) jgarnett: and I am being held up due to geotiff files not working in many cases
(7:21:26 AM) jgarnett: I do want to package up the results of our code sprint however :-)
(7:21:55 AM) jgarnett: I was just hoping to have the raster display be a bit more capable; in part because I am getting bug reports and complaints on and off list
(7:22:16 AM) ***moovida in standing ovation
(7:22:20 AM) jgarnett: It does sound like all the fixes are in from the code sprint; Jesse went a few extra days
(7:23:04 AM) moovida: the raster stuff should be really tested with different images in type and size.
(7:23:04 AM) jgarnett: This will be a similar release to the 1.2-M1 release; just download zip archives; no fancy installer (although I may include a JRE)
(7:23:37 AM) jgarnett: that will bring us to the next topic moovida :-)
(7:23:52 AM) jgarnett: any further requests / comments etc.... on making a 1.2-M2 release?
(7:23:55 AM) moovida: sorry, I thought uyou started on this, beg for pardon :)
(7:24:18 AM) moovida: my feeling was that you wanted to put in also some of the raster fixes?
(7:24:29 AM) jgarnett: nope; I was just mentioning it as the reason for my delay.
(7:24:36 AM) jgarnett: yes; you got it
(7:24:46 AM) jgarnett: well let us move on then ....
(7:24:49 AM) jgarnett: 2) raster bug
(7:24:50 AM) moovida: which would also include the raster chat before deciding the timeline
(7:25:00 AM) ***moovida is too slow in typing
(7:25:02 AM) jgarnett: Yeah perhaps I have the agenda in a bad order
(7:25:12 AM) jgarnett: Here is an example of a bug report
(7:25:32 AM) jgarnett: but most of the reports are along the line of "it does not work"; and some truly impressive workarounds
(7:25:55 AM) jgarnett: The bug report above lists a bunch of same data from around the web; and says when they work in uDig 1.1; and if they work or not in uDig 1.2
(7:26:08 AM) jgarnett: So far I have two main ideas on the cause of this trouble
(7:26:35 AM) jgarnett: - RasterSymbolizer default (now that we pay attention to raster symbolizer .. we need to make a good default for each kind of image RGB, Grayscale etc...)
(7:27:07 AM) jgarnett: - Image metadata (in particular coordinate reference system definition). It does not appear as if this is getting picked up all the time
(7:27:41 AM) jgarnett: So the first one seems to be a uDig problem; the second one may have a solution in the geotools codebase
(7:28:42 AM) moovida: Are you sure that it is just that?
(7:29:00 AM) moovida: The errors I incurred in during the sprint were of different nature
(7:29:05 AM) moovida: with nice exceptions
(7:29:16 AM) moovida: worldimages not recognised
(7:29:21 AM) moovida: and so on
(7:29:29 AM) jgarnett: I see
(7:29:42 AM) jgarnett: Can we package up a couple bug reports with those nice exceptions :-)
(7:29:43 AM) moovida: that is what I mean
(7:30:03 AM) jgarnett: I test the clouds.jpg worldimage included in the sameple dataset without problem
(7:30:04 AM) moovida: yes, I put some on the sheet we had
(7:30:20 AM) moovida: clods.jpg is not a serious thing to test
(7:30:29 AM) moovida: we need some ortophoto
(7:30:35 AM) moovida: tecnical cards
(7:30:45 AM) jgarnett: it is however small enough to ask people to download; the bug report above includes some orthophotos
(7:30:45 AM) moovida: maps that use particular compressions
(7:31:01 AM) moovida: yes, the bug report above is great, impressive
(7:31:04 AM) jgarnett: hrm; it would be good to be organized and methodical here
(7:31:12 AM) moovida: exactly
(7:31:18 AM) jgarnett: so we do not dupicate tests; or leave out something obvious
(7:31:25 AM) moovida: we need to setup something easy like for the sprint
(7:31:33 AM) moovida: so that testers can help out better
(7:31:47 AM) jgarnett: I am afraid I do not often work with raster files; and am unsure how to make a nice grid of different options
(7:32:01 AM) moovida: we can be of help here
(7:32:09 AM) jgarnett: well perhaps we could start with that UDIG-1469 bug report; place the files in a google docs spreadsheet
(7:32:30 AM) jgarnett: into a grid based on what kind format they are trying to be (RGB, grayscale, scientific, whatever)
(7:32:31 AM) moovida: Ok, I can take the effort to start that one and add some new with images
(7:32:58 AM) moovida: and then how do we organise?
(7:33:24 AM) silli: hey both, if you need I can try some rasters and give you a feedback
(7:33:47 AM) moovida: :D I bet you can silli
(7:35:13 AM) moovida: jgarnett: what were you thinking of for fixing those bugs?
(7:35:38 AM) jgarnett: well I think (due to it being a bug) we still do not know where the problem is
(7:35:43 AM) jgarnett: or how many problems there are
(7:36:03 AM) jgarnett: I suspect that we will end up fixing stuff in udig as we find it (like generating good default styles)
(7:36:25 AM) jgarnett: and for others we may end up with a series of very kindly worded bug reports for geotools and/or imageio-ext
(7:36:35 AM) jgarnett: I also note that imageio-ext is trying to hit version 1.0
(7:36:44 AM) moovida: yes, true
(7:36:52 AM) jgarnett: so I am worried we may be "too slow" with our testing effort to help them be great rather than only good
(7:37:29 AM) moovida: what about planning a couple of hours of syncronized work?
(7:37:50 AM) moovida: you stay up late and I wake up early or vice versa
(7:38:02 AM) moovida: to get into the things and have a common start point
(7:38:25 AM) moovida: obviously I say "I" and "you" which is everybody :)
(7:38:35 AM) moovida: I'm sure mauricio wouldn't let us alone
(7:38:46 AM) mauricio: :)
(7:38:46 AM) jgarnett: :-)
(7:39:03 AM) jgarnett: monday is a holiday here in Australia; so I could easily do monday evening with you guys
(7:39:13 AM) jgarnett: I think that works out to Sunday for you however :-(
(7:39:39 AM) jgarnett: But yes I am up to doing a short breakout session on this subject; but we should coordinate with the imageio-ext guys
(7:40:03 AM) moovida: yes, too bad they didn't come
(7:40:05 AM) jgarnett: (I did try and invite simone to this meeting)
(7:40:19 AM) jgarnett: okay let us organize a time on email
(7:40:29 AM) jgarnett: and move on w/ our agenda (no sense having long meetings)
(7:40:36 AM) jgarnett: 3) projection question
(7:40:38 AM) jgarnett: emily?
(7:40:59 AM) emily_g: when I load the default states population from geoserver into udig
(7:41:03 AM) emily_g: it displays nothing
(7:41:11 AM) emily_g: untill I reproject the map to 4326
(7:41:24 AM) emily_g: I've boiled it down to there being a difference between these to projection definitions
(7:41:39 AM) moovida: what is the projection it pics up natively?
(7:41:41 AM) emily_g: ...
(7:41:51 AM) emily_g: ...
(7:41:52 AM) emily_g: AXIS["Geodetic latitude", NORTH],
(7:41:53 AM) emily_g: AXIS["Geodetic longitude", EAST],
(7:41:55 AM) emily_g: ..
(7:41:57 AM) emily_g: and
(7:42:09 AM) emily_g: AXIS["Geodetic longitude", EAST],
(7:42:11 AM) emily_g: AXIS["Geodetic latitude", NORTH]]
(7:42:30 AM) emily_g: the native projection of the data is 4326
(7:42:39 AM) jgarnett: in the basicwmsrenderer
(7:42:40 AM) moovida: you mean when you look up the CRS in geoserver you find this stuff?
(7:42:55 AM) jgarnett: there is a bunch of crazy code to see how the coordinate reference system of the map matches the one from the wms
(7:43:06 AM) emily_g: no I was debugging through the code trying to figure out why the renderer was rendering nothing
(7:43:15 AM) emily_g: wfs
(7:43:18 AM) emily_g: not wms
(7:43:25 AM) jgarnett: we have updated to epsg-hsql (and forced lon/lat order) ... any of these changes should require us retesting I guess
(7:43:28 AM) jgarnett: oh
(7:43:33 AM) emily_g: it renders nothing because the x&y values of the bounding box are being switched around
(7:43:38 AM) jgarnett: wfs is the same deal; but even more fun to debug
(7:43:52 AM) jgarnett: if you use the sample requests demo thing in geoserver
(7:44:02 AM) jgarnett: can you determine what order it wants the coordinates in?
(7:44:46 AM) emily_g: maybe I don't think it has much to do with geoserver though; geoserver is feeding up the correct data
(7:45:04 AM) emily_g: geotools is then switching around the long/lat values of the bounding box returned by geoserver
(7:45:10 AM) jgarnett: question
(7:45:18 AM) jgarnett: is this the WFS 1.0 DataStore
(7:45:26 AM) jgarnett: or the WFS 1.1 DataStore that gabriel "just" wrote
(7:45:33 AM) jgarnett: which one has this problem; or both?
(7:45:50 AM) jgarnett: When you do go into the code you should be able to see a ReferencedEnvelope come in
(7:45:50 AM) emily_g: I believe its the 1.1 datastore; I don't know about the 1.0 datastore
(7:45:52 AM) emily_g: I will try it
(7:46:01 AM) jgarnett: and be "adjusted" to match the WFS DataStore expectations
(7:46:06 AM) DrogoNevets left the room (quit: Connection timed out).
(7:46:09 AM) jgarnett: there is a FilterVisitor that does this work
(7:46:24 AM) emily_g: yes; and the only difference between the srs's is the order in which the axis are defined
(7:46:36 AM) jgarnett: I am not too happy with the WFS 1.1. datastore; it is a bit young and does not do editing yet
(7:47:08 AM) jgarnett: in the WFS 1.0 DataStore this was handled by a specific filter visitor (that had to modify the origional request to match the target crs)
(7:47:11 AM) emily_g: I'll ask the geotools list and stop taking up everyones time. I'll poke around with the wfs1.0 datastore too thanks.
(7:47:23 AM) jgarnett: I am not sure if this has been done well/correctly for the WFS 1.1 datastore
(7:47:28 AM) jgarnett: talk to gabriel as well
(7:47:33 AM) emily_g: ok good place to start - thanks
(7:47:37 AM) jgarnett: good question! not easy but good
(7:47:43 AM) jgarnett: 4) style editor question
(7:48:08 AM) moovida: ok, can we make the import and export buttons overridable?
(7:48:24 AM) moovida: it is a problem for non sld styles
(7:48:29 AM) moovida: in the style editor
(7:49:06 AM) jgarnett: thinking
(7:49:23 AM) jgarnett: can we make them only enabled for the SLD style pages?
(7:49:35 AM) jgarnett: or have an extension point which asks the style configurator what data it is editing
(7:49:37 AM) moovida: also perfect
(7:49:45 AM) jgarnett: and looks around for a wizard to be able to export that data?
(7:50:00 AM) jgarnett: so to answer your question; yes we can
(7:50:06 AM) jgarnett: but what is a good way to handle it ...
(7:50:13 AM) jgarnett: even for geotools style objects
(7:50:25 AM) jgarnett: we are going to get two good export formats (SLD 1.0 and SE 1.1)
(7:50:33 AM) jgarnett: so I am a bit puzzled
(7:50:42 AM) jgarnett: do you have a specific bug that is bugging you?
(7:50:49 AM) jgarnett: say exporting the style of a scalebar failing or something?
(7:50:50 AM) moovida: I don't see your point
(7:50:58 AM) moovida: it is not a bug
(7:51:09 AM) moovida: exporting of JGrass raster styles
(7:51:13 AM) moovida: which are not sld
(7:51:23 AM) jgarnett: I was thinking that some styles may have more then one format available; so we may need to think about that when trying to digure out how to export stuff
(7:51:24 AM) moovida: so the buttons are there and user always try to export
(7:51:34 AM) moovida: and get some xml crab file
(7:51:35 AM) jgarnett: okay cool
(7:51:39 AM) jgarnett: you have a use case
(7:51:57 AM) moovida: can't the responsability be given to the style editor?
(7:52:08 AM) jgarnett: I would be willing to make a quick extension point; if you are willing to try and test it
(7:52:10 AM) moovida: it should know the data it threats
(7:52:22 AM) moovida: Jody, I am very glad to do so
(7:52:23 AM) jgarnett: I would feel better if we had 3 implementations rather than just geotools Style and JGrass
(7:52:31 AM) jgarnett: any thing else we can think of?
(7:52:33 AM) moovida: for 1.1 and 1.2
(7:53:10 AM) moovida: I use it also for annotation layers in beegis
(7:53:20 AM) moovida: and that would be another format
(7:53:52 AM) moovida: but what are you thinking about? I'm not sure I get your lightspeed coding thoughts
(7:54:57 AM) jgarnett: hrm; I am trying not to change the API on 1.1 that much :-P
(7:55:00 AM) jgarnett: stable release and all
(7:55:13 AM) jgarnett: but if you have to you can backport and make a 1.1.2 release
(7:55:25 AM) jgarnett: okay sounds like we got some good activities to work on
(7:55:31 AM) jgarnett: thanks for the meeting everyone
(7:55:42 AM) moovida: agree with you. 1.2, not 1.1
(7:56:03 AM) jgarnett: aside: emily ... thanks for helping out John Hudson the other week
(7:56:17 AM) emily_g: no problems
(7:56:23 AM) emily_g: I hope its all working for him now
(7:56:32 AM) jgarnett: I will post the logs
(7:56:58 AM) moovida: thanks
(7:57:12 AM) jgarnett: moovida I will do an initial cut of a style export extension point now
(7:57:16 AM) jgarnett: and ask you to review via email
(7:57:36 AM) jgarnett: do you want to start thinking of a good format for us to work through the raster files in an organized manner
(7:57:45 AM) jgarnett: I could start; but I feel I do not know enough

No comments: