Posting the IRC logs to the email list, will review and respond in a moment.
[12:13] <brent_> ok, lets start with GISP 3
[12:13] <brent_> here is the link to it: http://docs.codehaus.org/display/GEOS/GSIP+3+-+Issues+List+Module
[12:13] <sigq> Title: GSIP 3 - Issues List Module - GeoServer - Confluence (at docs.codehaus.org)
[12:13] <brent_> richard, corey?
[12:13] <chorner> hi
[12:14] <brent_> take it away, tell us about it
[12:14] <chorner> so, we're proposing an issues list module... it's a lot like what jesse implemented a few weeks ago
[12:14] <chorner> (the preference module)
[12:15] <jdeolive> did you read my comment?
[12:15] <chorner> it provides a client service so application like uDig can store information ("issues") in it
[12:15] <chorner> yes
[12:15] <jdeolive> any rebuttle?
[12:15] <rgould> i don't think it would work like that
[12:15] <chorner> each issue could potentially be associated with several features, or no features at all...
[12:16] <chorner> ... so an issue as feature makes my brain hurt
[12:16] <jdeolive> ha
[12:16] <jdeolive> fair enough
[12:16] <chorner> can anyone think of a way to make WFS-T work in that context?
[12:17] <jdeolive> well you can have features that dont have geometries
[12:17] <jdeolive> and then you could just have an attribute that is a list of fids
[12:17] <jdeolive> however i think you might complex-features to pull that off, groldan?
[12:18] <groldan> thinking...
[12:18] <brent_> is this issue list a protocol on its own that you guys will be making?
[12:18] <jdeolive> does that make sense chorner?
[12:18] <chorner> hmm
[12:19] <jdeolive> i beleive that is what the proposal specifies, we are discussing wether just using straight wfs is a better way to go
[12:19] <chorner> it makes sense, just not sure if it would be flexible enough
[12:19] <groldan> you could do that with complex-features indeed, I gues that'll be a list of feature references or some such
[12:19] <jdeolive> yeah, i am not sure you can pull that off with geoserver as it is today, not with out some hacking
[12:19] <chorner> btw, everyone meet Pablo Casado (benguela)
[12:19] <benguela> hi
[12:19] <jdeolive> HI pablo
[12:20] <brent_> hi
[12:20] <appetkov> Hello Pablo
[12:20] <rgould> he will be working on this with us (I think)
[12:20] <groldan> Hola Pablo
[12:20] <brent_> how will a client interact with this issue list?
[12:20] <brent_> like udig
[12:20] <jdeolive> i am guessing a custom http protocol, putIssue, getIssue, etc?
[12:20] <chorner> something like that
[12:20] <rgould> jdeolive, yes, we have yet tto define it
[12:21] <jdeolive> how will issues be stored on the server
[12:21] <chorner> hibernate --> db or flat file, i think
[12:21] <brent_> is it going to tie into validation (ie. will a validation error automatically create an 'issue')?
[12:22] <chorner> it could, but at the moment we're sticking to the client being responsible for populating the issues list
[12:22] <groldan> will the issues plugin catch up on other services (aka, wrap WFS), or other services would make use of it explicitly?
[12:23] <chorner> it would work independently
[12:23] <appetkov> so basically, people who have deployed geoserver can submit issues?
[12:24] <chorner> no, this is a service which clients can submit to, if a geoserver instance has this module installed
[12:24] <brent_> how do you want the 'issues' to be tied in with existing features and services (WMS, WFS)?
[12:25] <brent_> would you want to query the list and get a map back of the issues?
[12:25] <bnordgren> "issue" = nonspatial tabular data associated with features by reference?
[12:25] <chorner> it would reference existing features
[12:25] <chorner> bryce is correct
[12:26] <appetkov> oh OK
[12:26] <bnordgren> This is a model of a generic database.
[12:27] <bnordgren> One can currently define geospatial features which are related by association, correct?
[12:27] <chorner> that is an fm branch construct, isn't it?
[12:27] <jdeolive> in gml yes, in geoserver trunk no, on complex-features yes
[12:27] <jdeolive> this wont be on trunk until fm comes into geotools yes
[12:28] <jdeolive> so i think you are probably right in that you dont have the faccility to do this with straight gml and wfs
[12:28] * asantokhee has joined GeoServer
[12:28] <bnordgren> Ok, given that it's coming...how is an association with a nonspatial feature a special case?
[12:28] <asantokhee> hello, sorry for being late!
[12:28] <bnordgren> oops. Not coming?
[12:28] <jdeolive> so a custom service in my opinion is your best bet, however i would like a bit more about protocol and backend storage before i would vote in favor
[12:29] <appetkov> hello adit
[12:29] <chorner> fair enough
[12:29] <brent_> ok so this is basically just leveraging geoserver as an existing server to store 'issues' that don't have a direct relationshop to the features (indirect in the sense that the issue holds an ID of the feature, but is not bound by rules)
[12:29] <jdeolive> it doesn';t have to be much, not trying to get you to jump through hoops, but the GSIP is pretty vague as is
[12:30] <chorner> brent_: correct
[12:31] <jdeolive> on the other hand, this may not even be something that needs a GSIP
[12:31] <bnordgren> might I encourage whomever takes this up to allow the association to be between spatial features too?
[12:31] <jdeolive> i mean if you are not going to be using any WFS or GML stuff thats there, this will be stricly a plugin
[12:31] <jdeolive> i dont beleive a GSIP is necessary for plugins, if they are something is wrong with the process
[12:32] <chorner> 
[12:32] <brent_> yeh you guys can probably just implement this as a plugin and not worry about the GSIP process
[12:32] <appetkov> Cory do you mind giving a more concrete example of how uDig will be using it?
[12:33] <appetkov> more specific example I mean...
[12:33] <brent_> I can see this being very useful (and a tremendous amount of work) if the "issues" tie in live with the data and validation, with full problem description and resolution prcesses
[12:34] <chorner> hmm...
[12:34] <chorner> the current use in uDig is simply validation
[12:34] <bnordgren> I can see a lot of applications for this. Think of a sensor network (e.g. weather stations.)
[12:34] <bnordgren> The spatial feature contains their locations.
[12:34] <bnordgren> A separate table contains readings and an instrument number.
[12:35] <brent_> data-cleanup too
[12:35] <chorner> did that help at all?
[12:36] <bnordgren> Name it something other than issues and it helped.
[12:36] <brent_> I have a better idea of it now
[12:36] <chorner> so: it is a useful idea, but we're not sure if we need to get this approved or not
[12:36] <jdeolive> so I think it is your call cory, you can either provide more about implementation details in your GSIP, or not go GSIP and just develop your plugin indepenendly
[12:37] <jdeolive> do others agree?
[12:37] <brent_> I would agree
[12:37] <chorner> we are hashing out the details this week... so we will expand nonetheless
[12:38] <chorner> ok... i think that's it then
[12:38] <brent_> next agenda item?
[12:38] <brent_> 2) Multi-dim WCS
[12:38] <brent_> who is taking this one?
[12:38] <bnordgren> Adit: how's that prototype?
[12:38] <asantokhee> yeah its fine
[12:39] <appetkov> is it the one here?
[12:39] <asantokhee> im trying to deply it on a public server
[12:39] <asantokhee> for everyone to see
[12:39] <appetkov> http://lovejoy.nerc-essc.ac.uk:8080/Godiva2/
[12:39] <sigq> Title: GODIVA2 Data Access and Analysis demo page (at lovejoy.nerc-essc.ac.uk:8080)
[12:39] <asantokhee> no thats different
[12:39] <appetkov> not that one?
[12:39] <asantokhee> no
[12:39] <asantokhee> internally irt calls GADS though
[12:40] <bnordgren> Can we have the url when it's up (for wiki page?)
[12:40] <asantokhee> yeah sure
[12:40] <asantokhee> im actually getting few problems withe tomcat server
[12:41] <asantokhee> but it runs fine on my PC
[12:41] <asantokhee> hopefully this will b resolved by tomorrow
[12:41] <appetkov> sounds good
[12:41] <asantokhee> i had another query abt changing welcome page
[12:41] <bnordgren> How are you allowing time and elevation queries?
[12:42] <asantokhee> am following the example of thredds wcs
[12:42] <bnordgren> is that part of galeon?
[12:42] <asantokhee> http://motherlode.ucar.edu:8080/thredds/wcs/galeon/testdata/striped.nc?request=DescribeCoverage&version=1.0.0&service=WCS&coverage=ta
[12:42] <asantokhee> yeah, i think so
[12:43] <asantokhee> although they r not strictly following wcs spec
[12:43] <asantokhee> if we look at this example
[12:43] <asantokhee> or a getcoverage request
[12:44] <asantokhee> ive got another query :~))
[12:44] <asantokhee> suppose i have a dataset which contains several variables: temperature, salinity etc
[12:45] <asantokhee> so if i follow present wcs spec there is no way of representing this dataset as a coverage
[12:45] <asantokhee> i will need to have a separate temperature coverage, 1 for salinity etc
[12:46] <appetkov> that's what my understanding is too... Bryce?
[12:46] <bnordgren> As to the current WCS spec, I don't know.
[12:47] <bnordgren> It can be a coverage, however.
[12:47] <bnordgren> (in terms of the conceptual model)
[12:47] <asantokhee> its fine but sometimes we need to get several variables in a single dataset
[12:47] <appetkov> (currently the way we have it too...)
[12:47] <bnordgren> Let me lookup "rangeSet" in WCS spec.
[12:47] <bnordgren> Certainly this is allowed in the new WMS
[12:48] <bnordgren> spec as <Dimensions> attached to layers.
[12:48] <bnordgren> We may need to graft this onto WCS.
[12:49] <bnordgren> I think using current WCS, RangeSet is your answer.
[12:50] <bnordgren> You can include multiple axisDescriptions for each
[12:50] <bnordgren> coverage.
[12:50] <bnordgren> (an axis is an indepentent variable besides space and time...)
[12:51] <bnordgren> (oops. you want a way to have a vector of dependent variables...)
[12:51] <bnordgren> (nevermind).
[12:52] <bnordgren> However, let me draw your attention to the fact that
[12:52] <asantokhee> then how do we make query to request a specific variable frm the coverage
[12:52] <bnordgren> if Salinity is a float and temperature is an int, you can't
[12:52] <bnordgren> store them in the same variable in a NetCDF file either.
[12:53] <asantokhee> but most of our netcdf files contain multiple variables :~)
[12:54] <asantokhee> i think what i'll do is up,load a sample netcdf file that we use
[12:54] <bnordgren> That's ok. You don't need to map 1 file=one coverage.
[12:54] <brent_> (only 5 minutes left)
[12:54] <brent_> (we can continue after item #3)
[12:54] <asantokhee> ok
[12:54] <bnordgren> ok.
[12:54] <appetkov> ours too, that is why we are working towards nD coverage implementation
[12:55] <brent_> 3) WCS to trunk timeline
[12:55] <brent_> jdeolive wanted to know about this
[12:55] * simboss has joined GeoServer
[12:56] <appetkov> hi simone, just in time for 3) WCS to trunk timeline
[12:56] <simboss> sorry
[12:56] <simboss> guys
[12:56] <jdeolive> hi
[12:56] <appetkov> for me--the sooner the better
[12:57] <jdeolive> so with 1.4.0 about to go out, i am hoping we can branch 1.4.x and start on 1.5.x, which means bringing wcs home
[12:57] <appetkov> excellent
[12:58] <jdeolive> from what i can tell there are no blockers, geoserver can move to geotools trunk no problem and allessio has dont most of the work porting to the new architecture
[12:58] <jdeolive> what are your thoughts simboss?
[12:58] <simboss> afat about moving geoserver to trunk
[12:59] <simboss> we already fixed a lot of issues
[12:59] <simboss> I still have some concerns with the WFS
[12:59] <simboss> but if we talk about just switching to geotools trunk
[12:59] <simboss> the best thing to do would probably be what follows
[12:59] <simboss> :
[13:00] <simboss> 1>align the raster branch with the latest changes on the geoserver trunk
simboss 2>start using raster branch as the new trunk
simboss this way
simboss we should reduce glitches to a minimum
simboss with the new modular architecture
simboss this would proabably be the best path
simboss (imho of course)
jdeolive i agree, just doing a switch like we did for 1.4.x to trunk is the easiest route
jdeolive your branch has been much busier then trunk
appetkov timeline estimate?
jdeolive so I think its probably warranted
simboss timeline...
simboss we are working as hell this month
simboss to get a new release out
simboss we are targeting foss
simboss that is end
simboss sorry, 1st half of september
jdeolive ok, and we need to get 1.4.0 out first
simboss after that we could do the switch
simboss alessio could do that in a couple of days
jdeolive in the meantime, we should whip up a GSIP for this
jdeolive so we can officially vote on it
jdeolive question
jdeolive have you guys run your branch against WCS cite tests?
simboss nope
simboss and we won't do that for the moment
simboss the work in the last months has been focused mostly
simboss on WMS
simboss and speed
jdeolive fair enough
simboss after this release
simboss the main effort will be
simboss real nd WCS
simboss and hopefully
simboss time and elevation for wms
appetkov real nD
simboss real== no tricks
jdeolive ok, well i would like to time this for after foss
simboss me too !
simboss
jdeolive i dont want to slip you guys up before your deadline
simboss even because
jdeolive but i am also weary of all the development that is going on in parallel
simboss we have another important deadline
simboss on 21st sept
jdeolive so i might volunteer up my time to get this done after foss
simboss this==?
jdeolive your branch to trunk
simboss as it happened for the raster branch on geotools
simboss I can guarantte
simboss guarantee
simboss that in a couple of days
simboss we coudl easily merge back to trunk
simboss with not so many glitches
simboss anyway
simboss it is worth to spend sometime on your side to investigate
simboss possible issues.
jdeolive ok, i will
jdeolive and I will probably do up a GSIP in the meantime
jdeolive any objections to that?
brent_ none here
bnordgren none here
brent_ ok I'm going to call the meeting to a close, feel free to continue discussing item #2
brent_ thanks everyone for coming
bnordgren adit: been looking stuff up.
jdeolive thanks for running the meeting brent_
bnordgren thx brent!
simboss guys
brent_ I will post the logs
bnordgren I think this is a good rule of thumb:
simboss could you tell me a bit about what you have already discussed on coverages?
appetkov mainly issues adit is running into...
bnordgren 1] If you want to select by something, put a dimension on it. (it's an independant variable)
bnordgren 2] If yuou want more than one value with different units, forget it.
appetkov as well as conceptual problems, like the one Bryce cited
bnordgren 
bnordgren Section 8.3.3.2.3 of OGC03-065-r3 says different units means different coverages.
asantokhee i see
simboss k
simboss I have a couple of questions/concerns though
bnordgren Is everyone more familiar with how to grab elevation/time/parameters out of a WMS layer?
bnordgren shoot simone
bnordgren shoot <comma> simone!
simboss today I spent some time
simboss comparing
simboss WMS 1.3 and WCS 1.0
simboss as well as reading
simboss http://docs.codehaus.org/display/GEOS/Multidimensional+WCS
sigq Title: Multidimensional WCS - GeoServer - Confluence (at docs.codehaus.org)
appetkov good job sigq
simboss (I really would like to know what this sigq is...)
simboss and I have this question floating around
<--| benguela has left GeoServer
simboss aren't we making confusion between WMS and WCS
simboss ?
bnordgren my thot
bnordgren was to investigate different means of investigating
bnordgren (interrogating)
bnordgren multidimensional data sets..
bnordgren I just noticed that WMS 1.3.0 had an annex devoted to the topic.
simboss I think your idea was great Bryce
simboss because for sure
simboss WMS is far more pragmatic and clear than WCS 1.0
bnordgren We're going to have to figure how to make WCS play nice, though.
bnordgren I haven't read WCS 1.0 too much.
bnordgren is there much difference?
simboss but from he document I cited above it seems that we aim to extend/improve/integrate WCS with something from WMS 1.3
simboss I am afraid this approach could bring us some problems
simboss because WMS is WMS and WCS is WCS
simboss My interpretation would be
simboss let's look closely at both WCS and WMS
simboss to understand the way they handle multdimensional data
simboss which by the way similar
simboss and then
|<-- groldan has left freenode ("Leaving")
simboss let's try to bend ISO 19123 a little bit
simboss to adhere achieve the desired behaviours
simboss but
simboss afat
simboss about services, interfaces and operations
simboss both for WCS and WMS
simboss we should just follow the standard
simboss otherwise
simboss we will end up creating a new one
simboss if I had to change something
simboss I would rather do that at the toolkit level
simboss i.e. ISO 19123
simboss I mean
simboss GDAL docet
simboss it does not implement any standard
simboss but still
simboss everybody uses it
bnordgren i agree.
simboss but afat about service
simboss I would stick as much as possible
simboss with the current standards
simboss we could probably
simboss propose some enhancement
bnordgren I started looking at wms because I thought wcs didn't do multi dim.
simboss this is true
simboss they are pretty similar
bnordgren now I start seeing how same they are.
simboss the main different is
bnordgren Nothing like chasing your tail.
simboss WMS==rendering
simboss WCS==raw data
simboss WCS is about managing real multidimensional gridded data
simboss WMS is more about rendering heterogeneous sources of data
simboss both from local data source
simboss coverages or features
simboss as well as from remote service
simboss WFS and WMS
simboss using styled layer descriptor files
simboss WMS MUST be similar to WCS
simboss because
simboss in the end
simboss you could use
simboss a WMS to rendere data extracted from a WCS
bnordgren main problem I see with WCS:
simboss using the remote ows option provided byt the WMS styles
bnordgren one CRS for 3D.
bnordgren the WMS allows the vertical axis to have own CRS.
simboss you are absolutely right
bnordgren BBOX is 2D(horizontal) in WMS 3D (with unspecified z units) in WCS
bnordgren other than that, axisDescription similar to <Dimension>
bnordgren and one can request any parameter one declares.
bnordgren in either case, one must match up degrees of freedom in output format
bnordgren with the number of parameters which have ranges.
bnordgren (e.g. can make mpeg with TIME range but not TIME and ELEVATION)
bnordgren Actually, I think that ELEVATION (a special dimension for WMS)
bnordgren can be handled with WMS 1.0 as a parameter. ??
simboss one thing is not correct here
simboss in WCs
simboss WCS
simboss z is not unspecified in units
simboss or better
simboss I think WMS is better because it separates vertical from horizontal CRS
simboss while WCS
simboss does not allow this
simboss WCS crs
simboss can be a compound crs
simboss (see ISo 19111)
simboss horizontal+vetical
simboss as of EPSG:4327
simboss hence the uom for the z in WCS bbox
simboss can be inferred from the WCS
simboss however
simboss most services
simboss do not implement 3d compound crs
simboss (actually geotools does it, thx MArtin!)
simboss therefore
simboss with WCS
simboss they do as follows:
simboss 1>they specify a 2d CRS like EPSG:4326
simboss 2>they live the z without uom
simboss but this is not legal, strictly speaking
simboss is common practice tackling an outstanding issue of WCS 1.0
bnordgren main problem: CRS is not defined, it is referred to by identifier.
simboss true
bnordgren agreed on CRS issue.
simboss and this is one of the problems I wanted to throw on the table
simboss it is about time
bnordgren I do not think it is illegal under WCS to specify a 2D CRS,
simboss to implement a stub authority that is able to produce CRS
bnordgren define a "special" parameter called elevation
simboss as required by the new specs for WMS ans WCS
bnordgren and just neglect the "z" in BBOX.
simboss I have seen services doing this
asantokhee r we looking at something like this:
bnordgren I think Martin generates "CRS" authority CRSes.
asantokhee http://motherlode.ucar.edu:8080/thredds/wcs/galeon/testdata/striped.nc?request=DescribeCoverage&version=1.0.0&service=WCS&coverage=ta
simboss exactly adit
simboss I found this example too a couple of weeks ago
appetkov sigq choked on this one...
simboss if you check this coverage offering
simboss you will see that it is not completely legal
simboss at least
simboss imho
simboss the envelope is 3d
bnordgren adit: yes, but add a refSys to your axisDescription, "somehow"
simboss as well as the rectified grid
simboss but then they uses the range to describe elevation
simboss at the end
bnordgren quite right, simone! gotta use 2D envelope and bbox
simboss you have the same dimension describe both as part of the bbox as well as part of the range
simboss that means more or less
simboss that you are mixing domain and codomain
simboss what do you think guys?
bnordgren so...the axisDescription should include a refSys to a vertical CRS
bnordgren and an interval
bnordgren or maybe no interval since we have many singleValues...
simboss here the main error imho is
bnordgren also need a default value unless you require that all requests specify pressure
simboss here
asantokhee interesting, hw abt for multiple timesteps
simboss <gml:Envelope srsName="WGS84(DD)">
simboss the srsname
simboss is uncorrect
simboss because
simboss both the envelope
simboss and the rectified grid
simboss are 3d
simboss so we need a 3d crs
simboss like epsg:4327
simboss or something
bnordgren I agree with your first tactic:
bnordgren keep 2D CRS
bnordgren eliminate third param from BBOX/Envelope
bnordgren add vertical info to existing axisDimension.
simboss correct
simboss with this approach we could
simboss describe
simboss datacubes
simboss with irregular zlevels
asantokhee great!
bnordgren as to time steps:
simboss I was referring to this when I said I could come up with something more complex adit
bnordgren I think these are covered by "temporalDomain" p. 22
asantokhee exactly
bnordgren WCS defines the same special parameter "TIME" as WMS.
bnordgren for requests
simboss time is retty similar
asantokhee so in case time is also irregular we can use similar method as for elevation ??
simboss for both wcs and wms
simboss nope
simboss we need to make a distinction for wcs
simboss between the usage of get and post request
simboss get is simpler
simboss but far less powerful
simboss afat about time
simboss it does no allow directly
simboss ops
simboss hold on a sec
bnordgren I think temporalDomain is brain dead.
bnordgren must be regularly sampled.
bnordgren maybe we shouldn't use it
bnordgren use an axisDimension called time instead.
bnordgren is that illegal?
bnordgren that would take time out of the "domainSet" and put it
simboss back again
bnordgren in the "rangeSet". just like elevation.
simboss TIME allows
simboss to specify irregular instants
simboss both for get requests as well as for post request
simboss (talking about WCS)
bnordgren does this apply to temporalDomain?
simboss WCS?
bnordgren WCS uses TemporalDomain in DomainSet of DescribeCoverage
simboss yep
simboss let me provide a link
simboss to the right xml schema
bnordgren it looks like temporalDomain has min/max, optional timeResolution...
simboss http://schemas.stylusstudio.com/opengis/n9b25a865/element_temporalDomain.html
sigq Title: OpenGIS element temporalDomain (at schemas.stylusstudio.com)
bnordgren oh.... many "timePositions..."
bnordgren let me go look now...
simboss that's it
simboss
simboss you can mix them
simboss you can mix
simboss single time positions
simboss with continuos periods
simboss with discrete periods
simboss that is
simboss you can pretty much do whatever you want
simboss
bnordgren Ok, you win
bnordgren let's use temporalDomain
bnordgren
bnordgren So to recap:
bnordgren 1] Use the right spec (WCS whenever possible)
bnordgren 2] Adopt convention of using "elevation" axis with own CRS for 3D spatial data.
bnordgren 3] Adopt convention of using 2D CRS, BBOX, ENVELOPE parameters.
bnordgren 4] Define axis for anything you want to search for/select by/use as independant variable.
bnordgren right?
asantokhee 4th one
bnordgren 5] Use temporalDomain for time extent.
bnordgren go ahead...
asantokhee a variable must depend on a grid. the grid will consist of different axes
asantokhee so for each variable ,we"ll have different grids
bnordgren ...meaning a different combination of axes is returned for different variable?
asantokhee yeah it could be
simboss 1not sure I understood sorry...
bnordgren oh forgot:
bnordgren 6] You can return multiple values for one point in domain only if they have same units.
bnordgren e.g. f(x,y,z,t) = temp is legal.
bnordgren f(x,y,z,t) = {temp1, temp2} is legal
bnordgren f(x,y,z,t) = {temp, salinity} is illegal/error
bnordgren use two coverages.
simboss this is not completely correct bryce
bnordgren I got it from WCS 1.0 8.3.3.2.1
bnordgren Did I mis-read?
simboss let me explain
simboss In general
simboss what really counts
simboss is not the UoM
simboss sorry
simboss let me rephrase
simboss (I am tired )
simboss in general
simboss we might have a 5d coverage
simboss the domain is lon,lat,elevation,time as an instance
simboss and given a point in domain
simboss you obtain a tensor (5th dimension)
simboss of values
simboss which ALL depend
simboss on the 5th indipendent variable
simboss wavelength as an instance
simboss each single value of the tensor
simboss could have in general a different uom
simboss while still all of them depend on the same independent variable
simboss that is
simboss the wavelength
simboss giving another example
simboss let's look at a discrete coverage
simboss that retrieve various values like income, expectation of lile, etc.. by income
simboss the domain
simboss can be x,y,zt
simboss the independent variable can be the race
simboss (attention no UoM)
simboss and the tensor is made by a certain number of values
simboss with no UoM
simboss this is a single multidimensional covearage
simboss ok, it is not really what we aim to manage
bnordgren we may be getting off the beaten track here.
simboss but it helps with understanding how to handle these things
simboss sure
simboss the second exmample
bnordgren WCS doesn't manage anything but gridded coverages.
simboss is really at the borders
simboss (1.0 but 1.1 in on its way )
bnordgren In terms of the first example. I can grant that units may in general be different.
bnordgren However, WCS explicitly forbids measures of different quantities to be returned from
bnordgren same coverage.
bnordgren (e.g. tensor of wavelengths, but all must be observations of radiance or
bnordgren all must be observations of reflectance, can't mix.)
bnordgren (p. 24.)
simboss fair enough guys
bnordgren in any case, different units would beg the question of
bnordgren different data types
bnordgren being returned as different bands of the same image/grid
bnordgren YUCK!
simboss
bnordgren one coverage, one quantity is a handy convention
bnordgren to keep us out of trouble.
bnordgren Can you live with one coverage per variable, adit?
simboss I am fine with that
simboss
asantokhee thats a bit difficult
bnordgren remember: you don't need one FILE per variable!
asantokhee ah ok
bnordgren alternatively, can you suggest a
asantokhee then its fine
bnordgren way that multiple inhomogeneous values
asantokhee i thought it would mean storing 1 variable in 1 file
bnordgren be returned from a coverage
bnordgren ok nevermind then!
bnordgren are you guys ever going to fill in your bits of the wiki?
bnordgren and adit, are you good with the suggestions we made for your CoverageOffering?
simboss starting thursday, I am officially on vacation and I had to fight enough with gf to make it for this meeting
asantokhee yeah, its fine
simboss how is it going adit?
asantokhee i have managed to develop a prototype
asantokhee for the demo
simboss cool
asantokhee which is working on my pc
asantokhee am trying to install it on a public server
asantokhee i will put the link on wiki
asantokhee i had a few issues like modifying the welcome page
asantokhee i tried recompiling but it didnt change
asantokhee maybe thats a query for the mailing list :~)
appetkov i can probably help you out--probably tomorrow though...
asantokhee most welcome appetkov
asantokhee so my demo is for thursday
simboss tomorrow alessio can give a hand too
simboss (today is vacation here in italy)
appetkov excellent--bryce won't mind
bnordgren yeah I like vacation
asantokhee hopefully im hoping to have geoserver up and running tomorrow on the publi server
appetkov sounds good
asantokhee i can do simple requests for getting data
asantokhee it extracts data from files and then sends a url for downloading the data
asantokhee which shows integration of geoserver with gads
asantokhee at a very simple level!
appetkov sounds very cool
bnordgren a very NASA-esque approach.
simboss well done man!
simboss I have one simple questions guys which does relate directly to what we said so far
bnordgren ok
simboss it sounds scaring
simboss but it is not
bnordgren (ok under desk now!)
simboss should we start thinking about implementing
simboss a crs authorithy in geotools
simboss to handle
simboss crs
simboss specified as URN
simboss (remember bryce, the long term solution to the crs issue)
simboss if we manage to involve martin
bnordgren (can't see. under desk)
simboss I am pretty sure it would not be THAT difficult
bnordgren I vote that martin does it.
bnordgren Mainly, I think we just need a URN parser...right?
simboss well
bnordgren URNs specify authority and id, just like now.
simboss more or less
simboss yeah
bnordgren also has version, if I recall..
simboss I just downloaded the docs
simboss I am going to look over them during the week
bnordgren I don't think you can describe a CRS from scratch tho...
simboss first
simboss "how to advertise dimensions"
simboss
bnordgren yes, your vacation homework!
bnordgren
bnordgren just kidding
asantokhee ok have a nice vacation simone
asantokhee see you all!
appetkov thanks for stopping by simone
simboss alex, I saw you had some problems with the raster branch
appetkov all fixed for now, thanks
simboss did you commit the fix?
bnordgren have a good vacation and don't work too hard.
|<-- bnordgren has left freenode ("ircII EPIC5-0.0.5 -- Are we there yet?")
appetkov well it was an issue with some relic files I had frmo my previous SVN checkout
appetkov cleaning out the web subdir and doing svn update eliminated my problem
simboss cool
simboss anyway tomorrow alessio will be at work
simboss He will check everything out
appetkov will keep playing with it and will keep Alessio posted
appetkov since you are going on vacation
simboss my vacation
simboss ends
simboss tomorrow
simboss (short vacation
simboss )
appetkov excellent will keep bugging YOU then
appetkov just kidding
simboss from thursday
simboss tomorroe alessio!
simboss no really man
simboss bug us as much as possible
simboss we are close to the realease now
appetkov I want to try the latest code changes
simboss and we need people rasing issues
appetkov it will benefit us to deploy it for production
simboss you will love the mosaic
simboss I just added support for embedded overviews
appetkov yep, gotta get that into production
simboss it is soooooooooo fast
simboss when it works 
appetkov it feels fast, even now when I play with it
simboss kidding
appetkov OK, I will keep you posted
simboss k
simboss by man
appetkov see ya
appetkov bye adit, will get in touch with you tomorrow
