Dashboard > uDig > 2008 > May > 15 > IRC Logs May 15th
uDig
IRC Logs May 15th
Added by Jody Garnett , last edited by Jody Garnett on May 15, 2008
Labels: 
(None)

  1. what is up
  2. catalog
  3. wms kvp params
  4. h2
  5. installer

Jesse_Eichar_: but I got the issue figured out ~10 minutes ago so its ok
Jesse_Eichar_: Hi jody
moovida [n=moovida@bic.e.bkom.it] entered the room.
jgarnett: morning
moovida: thanks Jesse_Eichar_
jgarnett: my "commute" was slow today
moovida: Morning!
jgarnett: the streets are starting to fill up for spring.
jgarnett: morning
Jesse_Eichar_: agenda?
jgarnett: sure
jgarnett has changed the topic to: 0) what is up 1) catalog 2) wms kvp params
Jesse_Eichar_: anyone else?
moovida: not more than that for me
Jesse_Eichar_: ok
Jesse_Eichar_: 0)
Jesse_Eichar_: Fixed a Postgis Bug.
Jesse_Eichar_: working on Styling and considering the KVP bug for 2)
Jesse_Eichar_: There is a bug introduced into styling between RC14 and 15 so I want to fix that
jgarnett: Fixed a WMS params bug on trunk (for someone who was going crazy getting uDig to work for a demo), converting the arcsde datastore to a command queue rather than a simple lock
Jesse_Eichar_: I don't want an RC16
moovida: we won as only Italian University a HP contest which will bring us something like 25 tablet pcs and other hardware to do education and beegis development for field surveys
moovida: currently I am working on gps tool, annotation layer (draw on map)
moovida: and postit notes on maps
moovida: I know not much core
moovida: but it brings udig around
moovida: (beegis is yet another udig extention)
jgarnett: sweet!, I liked the idea of your annotation tool. I had though of something similar to the "pens" shown in Viso.
jgarnett: congrats.
moovida:
jgarnett: moovida; were you ever able to look at that tag of trunk?
moovida: o, exactly, can we spend two words on h2 at the end?
jgarnett: The reason I ask is I am hoping to use your feedback to see what else is needed...
moovida: jgarnett: sure
jgarnett has changed the topic to: 0) what is up 1) catalog 2) wms kvp params 3) h2
moovida: I moved all my projects (including udig sdk based) to Ganymede and cleaned up my old eclipse.
jgarnett: what is "ganymede"? Is that Eclipse 3.3 ?
moovida: now I am ready to follow the guidelines with the proper versions
jgarnett: oh; okay got it ...
moovida: I will do that in the weekend though
moovida: ganymede is 3.4
jgarnett: Is it released yet? I did not look ...
moovida: like europa last year
Jesse_Eichar_: no not yet
moovida: it is not fully released but seems to well
Jesse_Eichar_: but feature frozen I believe
jgarnett: we often found that waiting for the first patch was the best plan.
jgarnett: I have a commercial customer on trunk; so I cannot switch us over for a bit yet ...
moovida: I understand that, btw I want to get that work on trunc by next week start
moovida: so you will get the feedback on the docu
moovida: with the right versions
jgarnett: okay cool; now that I know your "timeline" I can stop bugging you
moovida: lol
jgarnett: shall we move on?
jgarnett: oh; did everyone handle the svn switch okay?
Jesse_Eichar_: yeah
moovida: yes
jgarnett: 1) catalog
jgarnett: So I need to smack catalog aroud a little bit; and it is my first chance to review what has gone on since uDig 1.0, read all the new methods and see what their javadocs say they are good for, compare with what the code actually does and so on...
jgarnett: I have updated the javadocs on 1.1.x, and trunk.
jgarnett: But I have a couple things to do:
jgarnett: - (with budget) Add IProcess to the list of things the catalog manages, capturing local and remote "workflow" / "process"/ "operation" kind of things
jgarnett: - (no budget) change over to some kind of H2 backed catalog; rather than XML
jgarnett: I also took a run at this page; http://udig.refractions.net/confluence/display/DEV/2+Catalog
jgarnett: diagrams to follow shortly.
jgarnett: Jesse, the split between "memebers" and "resources" has actually worked out rather well as a lead up to having a service of process.
jgarnett: So is any of this news or surprising?
moovida: everything?
Jesse_Eichar_: not for me.
jgarnett: heh; so that is why we have IRC meetings
moovida: the process support sounds really new to me
Jesse_Eichar_: anything about changing IGeoResource and IService so they have setters and cache the data? is that what you mean?
jgarnett: Here are my notes: http://udig.refractions.net/confluence/display/HACK/WPS
jgarnett: refractions is doing a Web Processing Service "end 2 end" from GeoServer, through client api in geotools, to user interface in uDig.
jgarnett: Jesse, that falls under the (no budget) idea above...
jgarnett: it is an area where we are weak.
moovida: wow,
Jesse_Eichar_: ok fair enough
jgarnett: Justin / GeoServer did up a sensible design last year at our request.
jgarnett: but they have so many of their own needs they do not really want to collaborate (ie share interfaces)
jgarnett: we can however steal the design (ie it is the same design we have in udig, just backed on to H2 using Hibernate)
jgarnett: now I would be really tempted to put this at the geotools level; since we want less code in uDig
jgarnett: but perhaps we should learn from bitter experience on this one?
jgarnett: So one solid question and then we can move on
jgarnett: The ICatalog interface has had a lot of methods added to it that don't belong; ie they are for communicating with the "local" catalog.
jgarnett: So I need to split this into two
jgarnett: So I have a choice between breaking API and keeping the code clear
jgarnett: or introducing a new super class of ICatalog (maybe ISearch) and tryig to explain that to everyone.
jgarnett: Any preference?
Jesse_Eichar_: This question again eh
Jesse_Eichar_: What methods are bothering you right now?
jgarnett: anything to do with temporary resources
jgarnett: or looking up by an id
Jesse_Eichar_: I can see temporary resource. But look up by id? why that?
jgarnett: (we have a lot of subtle methods for looking up by id, and everytime we don't understand them we make another )
Jesse_Eichar_: example?
jgarnett: opening codebase...
jgarnett: getResourceById is really for finding a live resource in the local catalog
jgarnett: not something you want to do from a remote catalog
jgarnett: so perhaps the ISearch interface better communicates my purpose ...
jgarnett: would it be more clear to have:
jgarnett: - ISearch for searching GeoNetwork, Google, Yahoo etc...
jgarnett: - ICatalog for managaging all known local resources, it would implement ISerach implying that local resources are searchable...
Jesse_Eichar_: getResourceById makes sense for catalogs though even remote ones does it not? But I guess it is not necessary for ISearch.
Jesse_Eichar_: So ISearch<-ICatalog? Where ICatalog is essentially ILocalCatalog?
jgarnett: OWS Catalog spec has a some kind of get by id method; so yeah ...
jgarnett: correct; basically ICatalog was turned "local" on me; where as if I was paying better attention I would of made an ILocalCatalog some time a go...
Jesse_Eichar_: do you want to deprecate CreateTemporaryResource and move it to ILocalCatalog?
Jesse_Eichar_: If we do that do we need ISearch?
jgarnett: no we do not
jgarnett: there are two ideas here; trying to sort out which communicates better ...
Jesse_Eichar_: and which is easier to migrate too.
jgarnett: option 1) ILocalCatalog exetends ICatalog; "local" methods are pulled down into ILocalCatalog
jgarnett: option 2) ICatalog extends ISearch; "remote" methods are pulled up into ISearch
Jesse_Eichar_: Option 2 has a better migration path but is potentially less intuitive than Option 1.
Jesse_Eichar_: But I think Option 2 is probably better because deprecated methods aren't very intuitive either.
Jesse_Eichar_: especially in an interface
jgarnett: okay option 2 it is ...
jgarnett: 2) wms kvp params
jgarnett: This is just a small thing; I found a frustated user on #udig yesterday who had "discovered" a "bug"
jgarnett: basically MapServer is not a very good WMS; it tries to put non KVP params into the query part of its URL; and does not play the version negoatiaton game ...
jgarnett: http://jira.codehaus.org/browse/GEOT-1815
jgarnett: I patched our WMS Request code to "do less damage" to the provided GET URL and I can now connect to their service on trunk...
jgarnett: do we care to backport?
Jesse_Eichar_: I think so
Jesse_Eichar_: Mapserver is one of the big guys
Jesse_Eichar_: If we don't properly support it that is pretty embarassing
jgarnett: I will back port to 2.2.x then
jgarnett: you can then pick up the change Jesse?
jgarnett: Or do you want to backport and review my work ...
Jesse_Eichar_: yes
Jesse_Eichar_: I can try but I am having problems getting as much done as I want.
Jesse_Eichar_: I have to port the translation to trunk still.
jgarnett: okay; I will back port to 2.2.x; I will need to check it out again after the move ...
jgarnett: I will mark the bug closed when the patch has been applied in both places.
jgarnett: 3) h2
jgarnett: moovida?
moovida: yup. I'm here
jgarnett: you wanted to talk about h2...
moovida: just wanted to kno
moovida: know
moovida: I followed some mail
moovida: since we are using hsqldb in a non spatial way
moovida: is there already a geotools module?
moovida: I thought that would be born with the soc guy
moovida: instead you are telling me there is something already?
moovida: that takes spatial data also?
moovida: (even if without indexing?)
Jesse_Eichar_: for hsql yes
Jesse_Eichar_: but it was dropped form later versions and h2 has been replacing it
jgarnett: on trunk there is an unsupported/h2
Jesse_Eichar_: h2 is reputed to be faster that hsql
jgarnett: put together by justin; the quality is actually better then most of the other datastores
moovida: which does some basic spatial?
jgarnett: and it is being used as a base for other JDBC datastores; like mysql
Jesse_Eichar_: hsql exists in 2.2 but has some bugs. Although it does work
Jesse_Eichar_: I think both have spatial ops
jgarnett: which does some basic spatial; it is a javadatabase, we "taught" it the JTS functions and get it to store geometry...
moovida: alright, I would migrate to h2 if that is what will be in future
jgarnett: but with out a spatial index some of those functions (like testing relationships) is slow; speeding it up is what the summer of code student is up to
jgarnett: but you can use it now.
moovida: cool, is there some doc page about that somewhere (apart of Justin's head?)
jgarnett: hsql did not offer "Transaction Independence" it was like a JDBC level 0 database; H2 is JDBC level 3 - basically a real database.
jgarnett: It is a datastore; you don't need to do much to it ...
moovida: now that is nice!
jgarnett: after all you are the master of createSchema now right?
moovida: also it should be accessible with postgres odbc in windows
moovida: ha ha
moovida: you are right... one day I will learn those features
moovida: Last question
moovida: throught that datastore I can deal with spatial data, but what if I want to join other tables? I access it the usual way?
jgarnett: that is where geotools falls short
jgarnett: there is no "join"
jgarnett: one hack is to make a view in your database; and geotools sees that as a normal (read-only) table
moovida: ok, got it
jgarnett: geotools is getting better; they now have the data model to represent "relationships" between tables (ie the external keys) as requested by that dude on the devel list...
jgarnett: but we need someone to try it out, RobA from social change was waiting for the goahead - and I let him know "now or never" last week....
jgarnett: so we got the API for interesting content; we just don't have an implementation in the public view yet
jgarnett: (RobA has a couple on branches of geoserver nobody every checks out)
jgarnett: If you had to do something today I would recommend the view approach
jgarnett: or a hibernate + udig combo
moovida: I wanted to have a look at hibernatespatial
moovida: sooner or later... which means... you know...
moovida: btw is the h2 db server is started by the module?
moovida: or do I have to start it on my own?
jgarnett: ?
jgarnett: unsupported/h2 exists in geotools
jgarnett: we need to make a net.refractions.udig.catalog.h2 plug-in to hold it
moovida: ok, got that, but that is done to access a db or also to start an instance of it?
moovida: because with hsql at the moment at udig startup be start the db server
moovida: and play with it
moovida: I don't understand if the geotools module accesses a server or starts one
moovida: btw I will have a look at it, I can't exactly understand what can be done
jgarnett: oh
moovida: don't want to mess the irc time
jgarnett: the JDBC driver for H2
jgarnett: includes the database
jgarnett: so there is no server to start ...
moovida: ok, thanks,that was what I needed
jgarnett: no worries
jgarnett: thanks for the chat guys ... end of meeting?
mohbana left the room (quit: Read error: 110 (Connection timed out)).
moovida: Ok for me
moovida: ups what about scripting? Jesse_Eichar_?
jgarnett: Jesse is there anything you need over the next week?
jgarnett: testing help etc...
jgarnett has changed the topic to: 0) what is up 1) catalog 2) wms kvp params 3) h2 4) scripting
moovida: right, when are you going to release?
moovida: in two weeks I will have a bunch of students working with that release out in the fields, so that would be a good testbed
moovida: as good as frightening
jgarnett: perhaps he went to sleep...
jgarnett: moovida are you talking about a release of a scripting plugin? or udig 1.1.0 release ....
moovida: oh, no, the release was connected to your proposal of testing
moovida: I thought you were talking about the release
moovida: ?
moovida: I wanted to know about scripting from Jesse, since I digged into groovy to get into the game, but Jessse seems more on scala now
Jesse_Eichar_: 1 sec
jgarnett: he is a hard act to follow
moovida:
Jesse_Eichar_: ok I'm back
Jesse_Eichar_: ummm.
Jesse_Eichar_: help eh.
Jesse_Eichar_: always testing
jgarnett: basically I want a days warning; and a relase to try walkthough 1 and walkthough 2 on
jgarnett: I have a training course first week of next month; and don't really want to be stuck between versions in testing hell
Jesse_Eichar_: Of course.
Jesse_Eichar_: Good to know.
jgarnett: note; the last RC15 I downloaded installed; but did not run ...
Jesse_Eichar_: for windows.
jgarnett: for windows
Jesse_Eichar_: yes I found that today as well
Jesse_Eichar_: its was a extra \ at the end so it was trying to write the .metadata to an illegal location.
moovida: in which plugin? I wanted to try that out and am dependent from that version
moovida: can I just export a plugin from the sdk? So I can test
Jesse_Eichar_: moovida. I'm lost... You talking to jody I hope?
moovida: ha ha
moovida: sorry
moovida: where is that error?
moovida: in a plugin of the windows udig version?
moovida: (I think that party yesterday did something to my brain...)
Jesse_Eichar_: Ah no
Jesse_Eichar_: I think I have the wrong nsis script in my workspace.
mohbana [n=mohbana@87-194-191-26.bethere.co.uk] entered the room.
moovida: ahhh, just the installer, you mean
Jesse_Eichar_: so it created the windows start menu item incorrectly
Jesse_Eichar_: yes
moovida: ok, then no problem, sorry
moovida: wanna stop here the meeting/
jgarnett: yep
moovida: we can take scripting to the next, isn't that important
Jesse_Eichar_: yes I got to run
jgarnett: I will post; since you guys want to head home.
moovida: yup
Jesse_Eichar_: thanks ciao all
moovida: thanks
moovida: Ciao

May 2008
Sun Mon Tue Wed Thu Fri Sat
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30 31
             

IRC Logs May 29th
IRC Logs April 24th

Powered by a free Atlassian Confluence Open Source Project License granted to uDig. Evaluate Confluence today.
Powered by Atlassian Confluence 2.7.1, the Enterprise Wiki. Bug/feature request - Atlassian news - Contact administrators
User-friendly Desktop Internet GIS