GeoTools : 2004-09-27 September 27th

ianT_: hi
aaim1: Hi Ian...
aaim1: Jody started to speak about reprojection, then he disappeared...
jive: Hi again
jive: was trying to set up WMS w/ CRS information
jive: So meeting time?
aaim1: Guess so (smile)
jive: 1) JTS 1.5
jive: 2) Reprojection - we need to help Martin with Generic Name?
jive: 3) CRS information for DataStore and GCE
cholmes ~chatzilla@201009075116.user.veloxzone.com.br è entrato nella
stanza.
cholmes: Hi all, sorry I'm a bit late.
cholmes_ ~chatzilla@201009075116.user.veloxzone.com.br è entrato nella
stanza.
jive: We are only just starting: 1) JTS 1.5 2) Reprojection 3) CRS for
DataStore & GCE
cholmes_ ha abbandonato la stanza.
jive: Does anyone have a 4)
cholmes: Did the factory bug I submitted yesterday get resolved? I was
wondering about David's solution, if we were going to use it, ect.
cholmes: (not yesterday, last week I mean)
jive: It is not fixed in the origional
jive: I ran into this problem the otherday as well (and may of fixed it
separately).
jive: We are talking FactoryFinder here are we not?
jive: Only reading one line?
cholmes: yes. Well let's put that as 4)
polio ~anonymous@mail.refractions.net è entrato nella stanza.
polio: 5) webDav
jive: Cool shall we start?
cholmes: sure
jive: GEOT-268 has just been created - aka Use JTS 1.5
jive: Sounds good to me +1
jive: Andrea does this include your CoordinateSequence work?
aaim1: No, it doesn't
jive: Darn.
aaim1: (sorry, I was busy re-creating my gt2 project in Eclipse after the OS
reinstall)
jive: GEOT-268 actually asks to upload JTS 1.5 to iblblio - 5) web-dev offers
an alternative.
aaim1: David said that we are going to work on it for 1.5.1
polio: ? Martin Davis
aaim1: Hopefully out in a couple of months...
aaim1: Uff... yes, sorry... I keep on confusing names...
polio: np
jive: I can' think of much more to say about JTS 1.5 - lets give it a go.
cholmes: I can probably upload to ibiblio during the meeting if you'd like...
polio: maybe wait untill point 5?
jive: Do you want to jump queue david ?
aaim1: It should be faster for generic geometry relation testing, but I did
optimize that path more than a month ago
polio: sure
polio: I set-up a webDav folder on the svn server and pointed maven at it
polio: lists.refractions.net/geotools
polio: everyone has access (developer username/pass)
cholmes: You updated all the modules to point towards it?
polio: nope, just the build.properties file in the head, do i need to do the
others too then?
cholmes: I'm pretty sure build.properties for each module needs to be
updated...
polio: yup, just checked (sad) ... gimme 5 min
jive: It sounds like from you email that it will still use iblblio if
lists.refractions.net goes down?
polio: but yah, do people like this idea? (we need the repo for udig, so just
pointed gt at it too)
aaim1: This way we would solve the problems with the maven repository
mantainance... great (smile)
polio: yah, well look at ibilio when you cannot contact lists ...
aaim1: You need it for udig? Do you mean that udig is built my maven too?
aaim1: my -> with (sorry)
aaim1: (I'm still recovering from fever...)
polio: nope, using ant to load a libs plugin for udig
polio: then spliced it into eclipse
polio: eventually I would like a script for nightly udig builds ... but that a
little bit off yet
IanS ~cholmes@chaos.tentacle.net è entrato nella stanza.
jive: Note to chris we have a nice little antscript in udig now - should allow
geoserver to grab the share this same idea.
jive: That is archive the script not the jars.
jive: Chris offered to upload JTS 1.5 to iblblio - perhaps he can try the
webdev folder instread.
jive: instread - instead.
jive: And we will need to update the developers guide - david?
polio: mmm, should not affect the naive developer
jive: I ment instructions on how to update lists.refractions.net (smile)
aaim1: I'm all for trying the new webdav server.
jive: I will do it - just trying to figure out where.
polio: i guess, type the address into your file browser and drag and drop
jive: That is right major change = vote? +1 from me
aaim1: +1
polio: +1
aaim1: I'd say, let's go on...
jive: Hard to know if this falls under Building / Using Maven or Conventions
(which really is "How the project works")
polio: agreed (but i missed the agenda (sad))
jive: 1) JTS 1.5 5) Web dev ...
jive: 2) Reprojection
jive: This is just a general call to try and help Martin with Reprojection.
jive: Apparently he is stuck on something called Generic Name.
aaim1: Which is?
ianT_: +1
jive: This is the hard part - I am not sure either.
jive: It is: Something they voted on using
jive: I think Martin has just run into the usual question of "How?" and "What
does it mean?"
jive: I assures me this is the onlything holding him up from working on
reprojection - and I need reprojection mid Oct (or devel grinds to a halt).
jive: (well our devel anyways).
aaim1: As far I can see from his e-mail there is something strange in the ISO
19111 spec...
jive: I have a hard time working against specifications I don't have access to

  • not a lot of fun.
    polio: can we just put one of those inefficient levels of indirection here,
    and worry about the naming at a later date?
    jive: I am not sure - I think client codes needs to know this information for
    creation.
    jive: So he is stuck - if he does not know what client code needs to write
    then it is hard for him to proceed.
    jive: (James just emailed me a "huge appologies from me
    jive: jrmacgill: I've had a 'build a demo by yesterday' droped on me from a
    hight
    jive: "
    jive: So I don't think he will be here (smile)
    jive: So for 2) Reprojection - I am not sure how we can all help, I just ask
    that we try.
    polio: rueben said we would in the near future (I'm totally lost when i gets
    to the reprojection code)
    jive: If anyone has access to the ISO information other than Martin they may
    be able to offer him a second opionion (or understand his emails).
    jive: Shall we move on - I don't know how to give this issue closure. Maybe
    set up a breakout IRC with martin or something?
    polio: the build hits the webDav now (smile)
    polio: maven just got quicker (smile) (and we don't have a direct connection)
    jive: 6) Arcsde BBox encoding (http://jira.codehaus.org/browse/GEOT-264)
    jive: For later.
    jive: 3) CRS
    Martin__ ~Martin_De@rocher.ird.nc è entrato nella stanza.
    aaim1: We just need to tell Martin that geoapi snapshot must be published on
    that server too
    jive: General annoucement that CRSServices is available in data.crs - if
    datastore providers and GCE implementors can start hooking up (or providing
    their own CRS information) that would be great.
    jive: Does anyone need any help with that?
    jive: Or should I make a bunch of individual Jira tasks?
    jive: Hi Martin__ we were just talking about GenericName.
    Martin__: Hello (Jody I guess?)
    cholmes ha abbandonato la stanza (quit: Read error: 110 (Connection timed
    out)).
    aaim1: A bunch of jira tasks would be nice, yes.
    jive: Yes (smile) It is jody.
    jive: Okay ... shapefile already has one.
    Martin__: I had a little bit trouble to connect this moorning. What was yours
    though on GenericName? I have a little bit of trouble with this one.
    jive: We know - I have not had much to think on the matter. We all feel a bit
    out of the loop on how to help on this one.
    aaim1: Martin, is the ISO 19111 spec publicly available?
    Martin__: The 2003 version of ISO 19111 is available on the OGC web site
    (public document), but there is a more recent revision in circulation only in
    the member OGC area for now.
    jive: Basically would love to stick to publically available specifications
    (aka OGC), but this time the OGC has told us to use something we do not have
    ready access to.
    Martin__: The 2003 version of ISO 19111 didn't used GenericName. The 2004
    revision of ISO 19111 now use it.
    aaim1: If the trouble is with the new spec... then I don't know how can we
    help you...
    jive: GeoAPI now has a working group status - can these people get you the
    information you need.
    jive: That is we would love to help - but we don't know either.
    aaim1: (I can't even try to interpret the contents of a document that I cannot
    read, of course...)
    Martin__: I have sent a question yesterday. Waiting for an answer.
    Martin__: I think I can send you the document in a private mail if you want,
    Andrea.
    aaim1: That would probably be a good idea, since this problem is a blocker...
    Martin__: Preparing the mail right now...
    jive: Hrm this does not seem to be that sane a way to do development - I
    presume something will be public by the time we are done?
    Martin__: I don't know when it will be public.
    jive: Well that is frustraiting - you are sure you want to do this?
    Martin__: Not quite sure actually.
    ianT_: I made some enquiries at the last OGC meeting as to how GeoTools could
    be come a member
    jive: Pros: You don't have to rewrite later, Cons: you mostly get to develop
    by yourself.
    ianT_: I was even ready to put the cash up from Leeds, but currently only
    Legal Enities can join
    jive: One advantage to not being a member is we serve as a good example of
    what they have made public.
    jive: Well some projects turn themselves into legal entities.
    jive: But it seems like a lot of work - I would rather code (smile)
    jive: So Martin you are the lead on this module - it really is your call. You
    could ask for a draft document; or tease an clarification from them on how to
    use Generic Name.
    ianT_: that was my feeling - I'll continue to pursue it (though I might be
    able to "employ" some people "at" leeds to allow membership)
    jpmeagher ha abbandonato la stanza.
    Martin__: I'm in the process of implementing GenericName the way I understand
    it. But I can't provide any garantee that my understanding is correct. We may
    have to correct it later if I appear to be wrong.
    jive: That is okay - "code like you mean it" - if what you mean changes we can
    always make a point release.
    jive: Besides you are the module maintainer - you are always right.
    Martin__: Well, I will need feedback from users...
    cholmes ~chatzilla@201009075088.user.veloxzone.com.br è entrato nella
    stanza.
    jive: Martin I will be happy to review your GenericName API and make
    suggestions if it will help.
    cholmes: (sorry, my internet connection decided to crap out for a half hour...
    I gotta find a better way)
    jive: It is okay you missed me asking you how you were going to hook
    PostgisDataStore up with CRS information.
    cholmes: Yeah, I've been wondering about that myself... Did you come up with
    an answer in my absence?
    cholmes: I need to dig into the crs code - you should be able to do an srid
    look up in the postgis table, on the spatial_ref_columns or whatever it is
    called, and read the wkt from there...
    jive: It looks like there is an implementation is module/main that does this.
    jive: I have made data.crs.CRSServices to provide this sort of thing for GCE
    and DataStores that don't have "native" support.
    jive: (I was asked to make a bunch of Jira tasks).
    cholmes: cool. I'll take a look at this week.
    jive: Martin__ can we do anything else to help w/ respect to Reprojection? Or
    reviewing GenericName the best we can do?
    aaim1: Sorry Jody... but are we speaking about CS support or reprojection? The
    data store reprojection code is mostly done...
    Martin__: There is not much to do right now I think. I'm working on
    org.geotools.referencing now, and too many peoples on the same class may be
    confusing.
    Martin__: For sure, understanding GenericName would greatly help.
    Martin__: I would also like to understand why in latest ISO 19111 revision,
    the CRS name is an Identifier type.
    Martin__: (i.e. an object of type 'Identifier')
    jive: I was thinking of Martin's support for reprojection based on his new
    CoordinateReferenceSystem API. I think our existing implementation uses the
    org.geotools.cs classes ...
    aaim1: It does
    Martin__: One thing that confuse me a little bit in latest ISO 19111 is that
    CRS (or any other IdentifiedObject) has 3 attributes:
    Martin__: 1) Code: An object of type 'Identifier' (seems raisonable to me)
    Martin__: 2) Alias: an array of object of type 'GenericName' (again seems
    raisonable to me)
    Martin__: 3) 'name': the primary name, an object of type 'Identifier' (here
    I'm confused: why the name is not of the same kind than alias?)
    jive: From my understanding it seems to be a namespace/name - that is a name
    that is valid only in a provided namespace.
    aaim1: Maybe because alias are not supposed to be unique?
    Martin__: Also is it correct to said that 'Nort American Datum 1983' is an
    'Identifier' rather than a GenericName?
    Martin__: GenericName provides exactly that: a name that is valid only in a
    given namespace.
    Martin__: This is true that GenericName said nothing about uniqueness.
    aaim1: I really have no idea...
    Martin__: Identifier has a code attribute, and an 'authority' attribute. What
    is the authorithy for 'Noth American Datum 1983' name?
    Martin__: I asked this question to the mailing list yesterday. Hoping an
    answer for an OGC guy...
    Martin__: (type: an answer from...)
    Martin__: (typo: type --> typo)
    aaim1: It seems we can only wait for that answer anyway...
    jive: I would assume Posgis if this is WKT from Postgis? EPSG if this is WKT
    from an EPSG service?
    jive: Hmm my assumption is bad - some other body probably defined what
    American Datum 1983 ...
    jive: Lets finish with the last couple of agenda items: 4) Factory File fix?
    cholmes: Yeah, David said he had a fix, and you have one? Could we see the
    fix in the main FactoryFinder class?
    polio: james was commiting stuff the day you asked ... I copied and hacked a
    fix for SchemaFactory about a month ago
    jive: I was running into this problem with CRSService on friday so I don't
    think it is fixed yet.
    cholmes: On trunk I assume? I'll try to roll it back to 2.0.x to test it
    out...
    polio: you wanted multiple entries per file?
    jive ha abbandonato la stanza.
    jive ~jgarnett@mail.refractions.net è entrato nella stanza.
    polio: yup, he was on trunk
    jive: Darn - new client (someone else better have the logs)
    polio: my version is hidden inside SchemaFactory.loadSchemas()
    polio: but is fairly optimized ... tha main change is that I open the file and
    parse it line by line rather than just grabbing the firststring
    aaim1: I can send you the log for the part so far... but I won't be able to
    stay until the end of the meeting
    polio: that'll work, i'm just missing the first 15min or so
    polio: (he'll hit me when he wants the other half)
    jive: I actually think we are done the meeting
    jive: I should bug chris about a bug
    aaim1: Ok, I'm sending you the first part by mail