Thanks dave and gabriel for capturing these
groldan Hi Dave
groldan dblasby, are you there?
dblasby yes
groldan hi, may be you can take a look at the latest additions at
http://docs.codehaus.org/display/GEOTOOLS/FeatureType+Survey#FeatureTypeSurvey-Requirementsbreakdown
dblasby okay - in bit
groldan thanks
dblasby how are things going for the sco_complex?
dblasby FYI - we have an anonymous read-only SVN up now. James just
needs to add users to it and we can start committing.
groldan good, a bit of delay, since I had to study more than
expected (gml, schema, etc)
dblasby yes - there is a lot to learn, eh?
groldan large stuff, I'd never been a GML guru
groldan once comfortable all seems easy, though
groldan where is the svn located?
dblasby codehaus
dblasby I dont think you should check it out until we have accounts
– I think the anonymous and user access is at a different URL
groldan ok, no problem
dblasby james said he will do it soon – he has to have some time at
the same time as the codehaus folks.
groldan aha, it could be a nice thing to have svn at codehaus
dblasby I think so - I've only used it a bit, but it looks pretty
good so far. Very fast!
-->| CIA-10 (n=CIA@flapjack.navi.cx) has joined #geotools
groldan isn't it IRC meeting?
groldan (meeting time, I mean)
dblasby in 5 minutes
dblasby everyone is always late...
groldan ah, ok
-->| jmacgill (n=jmacgill@beaver.codehaus.org) has joined #geotools
groldan dblasby, I'm going out for 20 minutes, can you plase add
FeatureType to the agenda for me?
-->| Polio (n=chatzill@209.53.191.206) has joined #geotools
dblasby ok
dblasby Add to agenda - "FeatureType" for about 1/2 an hour into meeting
dblasby anyone else have agenda items
rgould_ not sure where jody is
rgould_ but his laptop is here
rgould_ so I can pretend to be him
dblasby jmacgill Polio Jesse_Eichar?
\route to host))
Polio sure, shoot
Jesse_eichar hi my powersource just died
dblasby topics
dblasby ?
dblasby (1) LiteRenderer vs LiteRenderer2 – one must be deleted
Polio 1) meeting time
dblasby (2) meeting time
jgarnett I am just reading groldan's FeatureType page ... too much to
think about to comment much today
dblasby (3) FeatureType (gabriel)
dblasby Anyothers? gmacgill?
dblasby jmacgill
jgarnett Well an update on the build system may be good
jmacgill hi sorry
dblasby should we start - are we waiting on anyone/anything?
jmacgill a grant proposal was due in today, just sent it
dblasby good luck, james
jmacgill won't be me that works on it even if it comes off
groldan I'm back, thanks Dave
jmacgill ok, lets kick off then
jmacgill I see some adgenda itmes
jmacgill lite vs lite2 ... ?
dblasby there;s a lite and lite2 – they are both the same
dblasby we need to delete one
Jesse_eichar They are. I want to get rid of 2
dblasby I vote for deleting lite since most people are already using lite2
Jesse_eichar are they?
dblasby Geoserver is and so is udig.
Jesse_eichar But why have a name 2 when there is only one.
Jesse_eichar udig isn't
Jesse_eichar its using 1
jmacgill 2 is much cleaner iirc
dblasby you mean the old-one-now-the-new-2
dblasby its way too confusing.
dblasby I'd say 3 is better than 1
jmacgill or has 2 been renamed 1 recently?
Jesse_eichar yes
dblasby it appears so
dblasby its highly confusing
Jesse_eichar they are exactly the same until dave made some changes last week.
Polio << already lost >>
jmacgill which one did you chance dave?
Jesse_eichar I deprecated literenderer2 a little while ago
dblasby the one I always change – lite2
jmacgill ok, time line....
jmacgill hold typing a sec unless I get a step wrong
jmacgill 1) literender is writen
jmacgill 2) literenderer2 is created which shares styling code with j2dr
jmacgill 3) jesse deletes lite and renames lite2 lite
jmacgill 4) dave misses this and continues to edit lite2
jmacgill 5) jesse wants to get rid of lite2 before any more damage is done
jmacgill hows that?
Jesse_eichar That's about it.
Jesse_eichar I sent an email saying I was deprecating Lite2 but I
guess it was missed.
jmacgill So we need to move daves code from lite2 to lite and kill lite2
Jesse_eichar Yes
dblasby and everyone needs to modify their code to now point back to lite.
jmacgill what branch is all this happening on (trunk)
dblasby is anyone still using lite1?
dblasby both
Jesse_eichar Lite1 is compatible with the old Lite1
Jesse_eichar I think so dave.
MatthiasB I use lite1
Jesse_eichar Lite1 is a superset of lite2
dblasby so there's lite, lite1, and lite2?
Jesse_eichar (except for those changes yo added)
jmacgill ok, new proposal
Jesse_eichar no I call Lite=Lite1
MatthiasB Me too.
dblasby whats the old lite1?
jmacgill was it deleted?
Jesse_eichar I moved all the code that wasn't in Lite2 into the Lite
that I was running and made sure all the unit tests continued to work.
jmacgill can lite2's code be deleted and the class turned into a
delegator to lite
Jesse_eichar as soon as dave move his changes from last week.
jmacgill should make it clear that it is dead but won;t break anyones code
Jesse_eichar I made the changes about a week ago.
Jesse_eichar jmacgill: Yes. I don't like having Lite2 around any more
than necessary.
Jesse_eichar dblasbey: can you move the code?
Jesse_eichar Lite is the same as Lite2 but has a few more methods.
Jesse_eichar dave? that sound like a plan?
dblasby wait - lite and lite2 are different?
Jesse_eichar THe old Lite had some utility like methods. Didn't
affect how it rendered.
Jesse_eichar Lite renders the same as Lite2
dblasby Ya - I'd like to see all the crap removed from the lite renderer.
dblasby like render()
Jesse_eichar Hard to do more than deprecate. We're not the only
people who use LiteRenderer
Jesse_eichar but yes I'd like to deprecate some of it.
dblasby Why dont we make lite2 the non-crap lite?
dblasby "LiteLiteRenderer"
jmacgill
Polio LLR?
jmacgill StatelessRenenderer
dblasby or just "UseMeLiteRenderer"
1.0.3/20050414]")
dblasby I must admit; I'm really confused by the 3 lites. I just want
one thats easy to maintain and works.
Jesse_eichar 2 lite
Jesse_eichar there is LiteRenderer and LiteRenderer2 but sometimes I
erroneously refer to a LiteRenderer1 by which I mean LiteRenderer
Jesse_eichar sorry
dblasby Why dont we keep lite1 as it is for people who need the crap
and keep lite2 as the one without the crap?
dblasby we can depricate 1 and remove it evenutally
MatthiasB Just that lite2 is currently deprecated... that would confuse people.
Jesse_eichar How about this:
Jesse_eichar Make a superclass with all the good stuff
Jesse_eichar make LiteRenderer extend it.
Jesse_eichar then udig and geoserver and who ever else likes it can
use the "Good" renderer
dblasby You mean make a "clean" lite (without the crap) and have the
with-crap extend it?
cholmes I like calling it StatelessRenderer, so as to alleviate all
this confusion
dblasby but this will eventually mean that it'll break...
dblasby "StatelessRenderer" +1
Jesse_eichar StatelessRenderer isn't an accurate name
dblasby Just call it something other than lite...
Jesse_eichar Its not stateless. especially with dave's changes
jmacgill ah, lite is becoming statefull
Jesse_eichar fraid so.
jmacgill gettingheavierallthetimerenderer
Jesse_eichar I think it is a fire and forget renderer
Jesse_eichar create use dispose renderer
dblasby okay FnFRenderer
jmacgill jessy, what caracterists seperate it from j2dr?
cholmes I like CUDRenderer. Just because CUD sounds funnier
Jesse_eichar No features are kept in memory.
dblasby cud?
dblasby could call it StreamingRenderer
jmacgill StreamingRenderer (though Marin will point out that j2d can
be made to stream)
MatthiasB I remind you that after renaming you'll have to rewrita all
tutorials and code snippets... maybe lust keep the name?
jmacgill lets keep the junk and depricate methods that need to go.
Jesse_eichar jmacgill I think so too.
jmacgill lite2 becmoes a delegator to lite
MatthiasB I agree.
Jesse_eichar dblasbey we need to fix Lite2 so that is doesn't keep
the state. Remember last week I told you that I couldn't zoom without
creating a new Lite Renderer?
Jesse_eichar We need to fix that. It will break many people who use
LiteRenderer.
dblasby well - the only thing it needs to know is how big the image
to create to do the parallel stream.
Jesse_eichar The image is the Rectangle passed as a parameter to paint
Aug 15 23:20:05 <Jesse_eichar> Only the 2 paint methods should be non-deprecated for rendering.
Aug 15 23:20:11 <Jesse_eichar> (In my opinion)
Aug 15 23:21:15 <dblasby> ah - then it just need to be save in the /** The size of the output area in output units. */
Aug 15 23:21:26 <dblasby> private Rectangle screenSize;
Aug 15 23:21:38 <dblasby> line 171
Aug 15 23:21:46 <dblasby> you need to keep this up-to-date and it should work
Aug 15 23:22:21 <Jesse_eichar> I thought screensize what updated each call to paint?
Aug 15 23:22:41 <CIA-10> lavila * r15338 udig/community/lavila/plugins/org.cgiar.cip.diva.diversityanalysis/ (8 files in 5 dirs):
Aug 15 23:23:03 <CIA-10> lavila * r15339 udig/community/lavila/plugins/org.cgiar.cip.diva.gbox/src/org/cgiar/cip/diva/gbox/internal/analysis/ (Analysis1.java AnotherAnalysis.java):
Aug 15 23:23:20 <dblasby> well - thats what it uses to create the images. You might have to update the paint area as well - the ony change to that is how it writes the image onto the Graphics2d.
Aug 15 23:24:20 <Jesse_eichar> you are right ti isn't always set. So I think that's all the changes that are needed and it should be stateless again.
Aug 15 23:25:03 <Jesse_eichar> If you set the screenSize in paint(Graphics2d, Rectangle, Envelope)
Aug 15 23:25:06 <dblasby> well - that means you have to put more state into it!
Aug 15 23:25:11 <Jesse_eichar> then we should be good to go.
Aug 15 23:25:14 <Jesse_eichar> No
Aug 15 23:25:20 <dblasby> yes - probably only a few lines to change.
Aug 15 23:25:29 <Jesse_eichar> that's not more state
Aug 15 23:25:50 <Jesse_eichar> it is set each time paint is called but reset each time paint is called
Aug 15 23:26:04 <Jesse_eichar> There is not setScreenSize method.
Aug 15 23:26:51 <Jesse_eichar> understand?
Aug 15 23:27:20 <jmacgill> dave?
Aug 15 23:27:26 <dblasby> yes
Aug 15 23:27:40 <dblasby> geoserver is CUD so I dont really have any way of testing it.
Aug 15 23:28:06 <Jesse_eichar> Ok you want me to do it then?
Aug 15 23:28:31 <dblasby> sure - there should only be one line to change
Aug 15 23:28:33 <dblasby> graphics.drawImage(fts_array[t].myImage,0,0,null);
Aug 15 23:28:38 <dblasby> 10073
Aug 15 23:28:41 <dblasby> 1073
Aug 15 23:30:12 <Jesse_eichar> ok
Aug 15 23:30:28 <Jesse_eichar> So we agreed on what to do with Lite?
Aug 15 23:30:32 <Jesse_eichar> Keep lite
Aug 15 23:30:38 <Jesse_eichar> deprecate garbage
Aug 15 23:30:43 <Jesse_eichar> Lite2 delegates to Lite
Aug 15 23:30:47 <Jesse_eichar> ?
Aug 15 23:30:57 <dblasby> I'd like to get rid of the garbage as its impossible to figure out if you're breaking it
Aug 15 23:31:04 <dblasby> and it makes things more difficult that they need be.
Aug 15 23:32:15 <dblasby> its already 1500 lines long!
Aug 15 23:32:43 <Jesse_eichar> how do we do that? other than deprecate it. I can move it all to the end of the file and put a big dashed line across saying all code beyond this point is evil!
Aug 15 23:33:31 <Jesse_eichar> We're supposed to be backward compatible. Aren't we?
Aug 15 23:33:35 <Jesse_eichar> can we do that on trunk?
Aug 15 23:34:16 <Jesse_eichar> Damn I have to go. I have a meeting.
Aug 15 23:34:16 <jmacgill> we need to be backwards compatable for a version
Aug 15 23:34:32 <Jesse_eichar> for a version. So next version we can kill the deprecations?
Aug 15 23:34:34 <dblasby> you mean 2.1 --> 2.2?
Aug 15 23:34:46 <dblasby> or 2.1.1 --> 2.1.2?
Aug 15 23:34:50 <Jesse_eichar> (is deprecations a word?)
Aug 15 23:35:14 <jmacgill> we need to release a vewsion with the code marked as depricated
Aug 15 23:35:20 <jmacgill> then in the next one it can go
Aug 15 23:35:28 <jmacgill> so, 2.2 will need it, 2.3 can kill it
Aug 15 23:35:37 <jmacgill> we plan to move though versions much faster
Aug 15 23:35:56 <CIA-10> lavila * r15340 udig/community/lavila/plugins/org.cgiar.cip.diva.diversityanalysis/src/org/cgiar/cip/diva/diversityanalysis/ (5 files in 3 dirs):
Aug 15 23:35:56 <Jesse_eichar> ok
Aug 15 23:36:06 <Jesse_eichar> +1 for me for deprecating garbage now.
Aug 15 23:36:07 <jmacgill> I jusgest making a class called LiteJunk and delegate calls to it for the deprecated methods
Aug 15 23:36:15 <Jesse_eichar> sure
Aug 15 23:36:30 <jmacgill> should cut the size down a bit
Aug 15 23:36:59 <Jesse_eichar> I'll check the logs out later for the rest. Talk to you guys later. Sorry for having to take off.
Aug 15 23:37:12 <Jesse_eichar> I won't have a house to live in soon if I don't go.
Aug 15 23:37:13 <Jesse_eichar>
Aug 15 23:37:23 <MatthiasB> Bye.
Aug 15 23:39:37 <jgarnett> So we kind of got some closure there ...
Aug 15 23:39:38 <jmacgill> I think that wraps that topic up then
Aug 15 23:39:43 <jgarnett> ... what is next on the agenda?
Aug 15 23:39:46 <jmacgill> Yeah, I think so
Aug 15 23:39:52 <groldan> meeting time
Aug 15 23:40:15 <jmacgill> The proposal was to move it back to the time when people actualy manage to turn up
Aug 15 23:40:26 <jgarnett> When would that be?
Aug 15 23:40:38 <jmacgill> about 30mins after it is supposed to start
Aug 15 23:40:38 <groldan> one or two ours early?
Aug 15 23:41:05 <jmacgill> whatever works for people
Aug 15 23:41:36 <jgarnett> well ealier does not get a Martin, but neither does 30 min later.
Aug 15 23:41:54 <jmacgill> what does pick up Martin?
Aug 15 23:42:06 <jgarnett> I think later picks up martin
Aug 15 23:42:17 <jgarnett> if we have a meeting that goes two long sometimes we pick him up.
Aug 15 23:42:26 <jmacgill> lets put this to the list - ask him what time he can manage
Aug 15 23:42:38 <jmacgill> It may be that I can start to do this from home now...
Aug 15 23:42:42 <jgarnett> We should ask simboss who set up a meeting with him last.
Aug 15 23:43:06 <jgarnett> (speaking of which he asked for a vote on email, if you have not done sone please chime in for him)
Aug 15 23:44:50 <dblasby> I have to go
Aug 15 23:45:03 <cholmes> I'd actually say move it a half hour forward, so that stragglers might actually show up by the time it's supposed to start.
Aug 15 23:45:17 <cholmes> Do people actually have conflicts preventing them from getting on time, or just general forgetting?
Aug 15 23:45:36 <cholmes> If we move it back, can Polio still make it? He can only come on his lunch break.
Aug 15 23:46:02 <cholmes> Have people reviewed Martin's code? If so I'm +1, the ideas behind it are sound.
Aug 15 23:48:19 <groldan> we may just ask the list as james said, since there are too few peoplo today anyway, thuogh I don't know if ultimatelly there are more people than the presents, polio and martin?
Aug 15 23:51:53 <cholmes> ok
Aug 15 23:52:23 <groldan> shuold we move on next tipic? FeatureType stuff
Aug 15 23:54:01 <groldan> (shuold --> should, tipic --> topic. sorry)
Aug 15 23:56:55 <jmacgill> yes lets
Aug 15 23:57:21 <jmacgill> the wiki info on this is excelent btw
Aug 15 23:58:05 <groldan> thanks, so somethings are arising, lets call them problems
Aug 15 23:58:17 <groldan> sorry, brief introduction:
Aug 15 23:58:47 <groldan> the point is to ensure that Feature/FeatureType/AttributeType API is able of expressing a GML(3) schema
Aug 15 23:59:08 <groldan> what makes a GML 3 schema different from a GML 2 is basically
Aug 15 23:59:26 <groldan> - complex features ==> nested attribute types
Aug 15 23:59:40 <groldan> - attributes from different namespaces
Aug 15 23:59:47 <groldan> - Xlinking of nested features
Aug 16 00:00:59 <groldan> initial requirements and design is at: http://docs.codehaus.org/display/GEOTOOLS/FeatureType+Survey#FeatureTypeSurvey-Enhancementrequirements
Aug 16 00:01:45 <groldan> may be we can choose to set up a breakout meeting on the topic and let more people get involved?
Aug 16 00:02:25 <groldan> so the basic requirements from the API point of view:
Aug 16 00:03:12 <groldan> - AttributeType is not able to express the whole range of XML Schema simple types, we rely on AttributeType::getType:Class
Aug 16 00:04:13 <jgarnett> I sent an email with my comments out of band
Aug 16 00:04:34 <groldan> and there are no direct java types for all the schema types (for example, we can manage Integer and Long, but integer in XML land is the infinite set of integer numbers, "int" is a 32 bit integer, but we cannot manage nonPositiveInteger, etc
Aug 16 00:06:12 <jgarnett> actually their are well known mappings due to JAXB
Aug 16 00:06:19 <jgarnett> to things like BigDecmial
Aug 16 00:06:32 <jgarnett> combined with our AttributeType "facets" we should be good to go
Aug 16 00:06:33 <jmacgill> we need to be GML3 like, but I think that upper limits on ints and doubles etc are acceptable. We should provide for BigInt etc, but they are slow so should not be our default
Aug 16 00:07:09 <jgarnett> Would also like to check that we remain a superset of whatever standards are on the table
Aug 16 00:07:34 <jgarnett> (I tried to implement geoapi Filter, but there was no point as it "does" less then what we do already)
Aug 16 00:07:54 <jgarnett> By being a superset at the Object level we can respond to specifications as they come out.
Aug 16 00:08:09 <jgarnett> (did that make any sense james?)
Aug 16 00:08:37 <jmacgill> yes, but I think a breakout is essential
Aug 16 00:08:44 <jmacgill> I've run out of time too
Aug 16 00:08:56 <jgarnett> yep, can we set a time? I am available after wednesday?
Aug 16 00:09:10 <jgarnett> It would be nice if RobA could make it...
Aug 16 00:09:34 <groldan> ok, that would be great. May I ask on the list for Friday, at usual meeting time?
Aug 16 00:10:27 <jgarnett> sounds good, I should head out as well (someone should grab the logs)
Aug 16 00:11:00 <groldan> is that a good time for you, james, chris, dave?
Aug 16 00:11:57 <cholmes> I think dblasby is away. I unfortunately can't make any time the next two weeks... I'm in the states and will be super busy.
Aug 16 00:12:06 <cholmes> I can read the transcripts on email.
Aug 16 00:12:09 <jgarnett> Okay I got another meeting right now, I will catch up on logs later.
Aug 16 00:12:33 <cholmes> oh yeah, getting davidz there would be very good, send him an email
Aug 16 00:14:14 <groldan> jmacgill? are you comfortable with friday usual time? it may change due to others timing anyway
Aug 16 00:14:34 <jmacgill> works for me
Aug 16 00:15:08 <CIA-10> magnasound * r15341 udig/community/lavila/plugins/org.cgiar.cip.diva.diversityanalysis/src/org/cgiar/cip/diva/diversityanalysis/ (3 files in 3 dirs):
Aug 16 00:16:39 <CIA-10> magnasound * r15342 udig/community/lavila/plugins/org.cgiar.cip.diva.gbox/plugin.xml:
Aug 16 00:16:46 <MatthiasB> Bye, all.
Aug 16 00:17:10 <groldan> ok, that's all from me so.
Aug 16 00:17:18 <groldan> I'll send the email right now