Agenda Items:

1. What is up
2. jaxb proposal
3. default values
4. svn cut off

Action items:

• waiting on votes for from jdeolive and ianT
• Svn access will be cut off at the end of march for anyone that has not signed a GeoTools Contribution Agreement

desruisseaux: Agenda topic: JAXB annotations on metadata - vote?
jgarnett has changed the topic to: 0) what is up 1) jaxb proposal .... x) svn cut off
Daniele [n=chatzill@host23-197-dynamic.37-79-r.retail.telecomitalia.it] entered the room.
Daniele left the room (quit: Client Quit).
simone [n=chatzill@host23-197-dynamic.37-79-r.retail.telecomitalia.it] entered the room.
aaime [n=aaime@host212-40-dynamic.1-87-r.retail.telecomitalia.it] entered the room.
desruisseaux: We have 2 agenda topic. Anyone else has other ones?
aaime: default values and validation?
Eclesia [n=Administ@ACaen-157-1-114-235.w90-17.abo.wanadoo.fr] entered the room.
jgarnett has changed the topic to: 0) what is up 1) jaxb proposal 2) default values .... x) svn cut off
desruisseaux: Are we ready to begin?
jgarnett: yes; Martin I am fielding some udig questions can you run the meeting today please?
desruisseaux: Will try (I'm not as good as you - don't even know how to change the topic!)
desruisseaux: Si 0) whats up
desruisseaux: On my side: martin) Looks like that I finally got a first version of ImageMosaicReader working. Performances seem good.
desruisseaux: (tried on Nasa's BlueMarble)
aaime: Nice. Is that working against postgrid, seagis?
aaime: I mean, is that something we can try out somehow?
desruisseaux: It is used for postgrid, but it is independant of it.
desruisseaux: (typo: used by)
Eclesia: Johann sorel : i found how to code widget so that they can be inserted in netbeans gui editor, most of the widget are now ready for that
desruisseaux: Yes, it is just a javax.imageio.ImageReader implementation.
groldan [n=groldan@217.130.79.209] entered the room.
jgarnett: jgarnett - udig version hell
***aaime fighting functional testing against live data directories in geosever
simone: how do you store the tile index martin?
jgarnett: jgarnett - should be positive, intergrating some greate german translations for udig
simone: simone: doing non-geotools work
desruisseaux: Not ImageMosaicReader's business. This is user responsability to create a set of Tile objects. On my case I do that on Postgrid side. You can do that using a Shapefile if you wish.
desruisseaux: A Tile object provides the informations needed by MosaicImageReader.
***groldan doing 80-20 ArcSDE work (80% the time trying to get an instance to connect to, 20% working)
jgarnett: heh; you guys should grab an agenda item!
simone: jgarnett: weren't you answering udig questions
You are now known as repressed
simone:
desruisseaux: Can we move to agenda topic 1?
desruisseaux: I assume that the answer is yes...
desruisseaux: Proposal: http://docs.codehaus.org/display/GEOTOOLS/JAXB+annotations
desruisseaux: Reminder: no new dependencies
desruisseaux: Only drawback I could see: would increase the size of metadata JAR.
jgarnett: only part missing is who does the tasks.
desruisseaux: Mostly Cédric
You are now known as jgarnett
desruisseaux: He have done almost everything
jgarnett: Specifically I am interested in who does the documentation tasks; something that has been a sticking point on the last several proposals.
desruisseaux: Vincent (vheurteaux), can we give this task to Cédric too?
desruisseaux: I assume that we just need indication about how to parse and format XML from a Metadata object?
vheurteaux: yep!
jgarnett: correct; you have the tasks already on the page - I just wanted to make sure a body was associated with the work.
desruisseaux: (actually I believe that Cédric already started some documentation draft)
desruisseaux: Well - Cédric everywhere.
jgarnett: (and it is not like they need to be long; just enough of a code example that users can start asking real questions on the mailing list)
jgarnett: what is his confluence id?
desruisseaux: (looking...)
desruisseaux: Seems to be cedricbr
vheurteaux: cedricbr
desruisseaux: So can we call for a vote?
jgarnett: okay with that accounted for I can vote +1
desruisseaux: Thanks
simone: +ò
simone: ops
desruisseaux: +1 on my side too of course
simone: +0
jgarnett: we have not managed to get a vote out of IanT for a while; perhaps via email.
jgarnett: aaime ping?
jgarnett: jdeolive ping?
aaime: Sorry
simone: I have a question though
desruisseaux: Yes?
simone: how this work compare to using hibernate or something like it
desruisseaux: Similar idea
simone: xmlbeans
simone: I mean, does it preclude usage of an alternate framework?
jgarnett: no it does not
simone: most people don't use JAXB
simone: at least afaik
desruisseaux: I can't be sure that I'm understanding right because JAXB is a new technology for me and I don't master it. But from what I have understood, I have the feeling that JAXB is like JDBC : as set of standard interfaces (actually standards annotations) allowing different vendors to plugin their own implementations.
pramsey [n=pramsey@S01060014515fec41.gv.shawcable.net] entered the room.
desruisseaux: Java 6 is bundled with an implementation, but if I'm understanding right we are not forced to use that implementation.
simone: jgarnett: say we would want to use hibernate, we would have to use xml fescriptors
jgarnett: simone my experience is mixed; a lot of people use jaxb on the "intranet" side of things; especially for SOAP/WSDL side of things. They just treat it as part of java and hope the xml stuff never has to be looked at.
simone: we could not use annotations, right?
jgarnett: simone you could use annotations; the annotations do not collide or anything (they are only metadata)
simone: k
simone: just curios...
aaime: anyways +1 for me
desruisseaux: Thanks
groldan: is there a iso19139 document made out of a Metadata object somewhere to have a look at it?
groldan: out of a MetadataImpl I mean
desruisseaux: Cédric have some. Do you want me to ask him to post it on the mailing list before vote?
jgarnett: I am hoping to see that as part of a test case / code example.
groldan: an attachment to the proposal may be?
groldan: and no, I'm not saying I want to see to beleave (before voting)
desruisseaux: No problem I woudl have considered that as something perfectly normal and legitimate anyway.
desruisseaux: I will ask him tomorrow to post his examples as attachment to the proposal page.
groldan: not sure, may be like requiring the job to be complete beforehand
groldan: I'm just curious to see the product of it
groldan: and if the jaxb tech plays well with namespaces and prefixes and the like
desruisseaux: Actually the job is already mostly completed - we wanted to make sure that it was doable before to make this proposal.
groldan: yeah that's smart too
desruisseaux: I know that he have namespace - I can't said if they are all right since I'm not a XML specialist, but to a novice like me they looks like honest namespaces.
groldan: question: what do you do regarding InternationalString and Text elements?
desruisseaux: Cédric is working on it right now (I means today - he will continue tomorrow)
groldan: I mean, is there a way to encode FreeText elements in more than one locale?
desruisseaux: Yes
desruisseaux: Since today
groldan: wow, cool
groldan: +1 vote here, community support
vheurteaux:
desruisseaux: He showed me a working example. He is now tracking a bug in unmarshalling of FreeText with more than one local.
groldan: you never had the feeling InternationalString needed a getLocales():Set<Locale> method?
desruisseaux: Yes
jgarnett: martin did you do the proper ISO thing for InternationalString? As I recall their was a general solution that could be applied to GetCapabilities documenets and the like. Declare the default langauge in the header; and use some kind of tag in the middle of the free text sections for each langauage.
desruisseaux: The problem is that Set<Local> is hard to implements on top of java.util.ResourceBundle.
groldan: yup
groldan: that's why I had to make my own InternationalString implementation a while ago, working with Hibernate
simboss_away [n=chatzill@host23-197-dynamic.37-79-r.retail.telecomitalia.it] entered the room.
desruisseaux: Jody - I'm not familiar with that. But we will look at it - peoples here are pretty sensitive to localization, so I guess that this topic will get attention.
groldan: sorry for the disruption, continue with the meeting
jgarnett: we can talk about it after the meeting
desruisseaux: One a related topic, Cédric will need a SVN write access in order to commit his work.
desruisseaux: He would do that very progressively, begining with only small bit in to let time for peoples to review if they wish (rather than big bunch of commits)
jgarnett: martin you will need to nominate him like normal; and review his work.
jgarnett: (as usual this is mostly a test to see if the developers guide has been read)
desruisseaux: All right
desruisseaux: Thanks for the vote on the metadata proposal. I'm done.
jgarnett: 2) default values
jgarnett: aaime you have the fllor
jgarnett: floor.
aaime: Ah, this is just to summarize my mails about default values and validatio of last week
aaime: since I got no answers to the last one
aaime: To sum up, forcing default values into non nullable fields is a behaviour we have in 2.4.x too
jdeolive left the room (quit: Read error: 110 (Connection timed out)).
aaime: and removing it would break some modules
aaime: in particular MIF
aaime: yet validation can be removed easily and will cause no damage
aaime: (besides one test that needs fixing)
aaime: I'm curious of one thing tought
aaime: all the information needed for validation is stored into Feature and Property
aaime: so why do we use an external utility class to make validation?
aaime: Wouldn't it make sense to have an isValid() method in both Property and Feature?
simboss_away is now known as simboss
simone left the room (quit: "ChatZilla 0.9.81 [Firefox 2.0.0.12/2008020121]").
aaime: hmmm... any reaction?
groldan: thinking...
groldan: I guess it would make sense, and also would make sense isValid delegates to the helper class
jgarnett: thinking ...
aaime: sleeping...
groldan: like to alleviate the task for different implementations..
jgarnett: I would like to remove validation; unless the user asks for it. The default value is available; so if a DataStore wants to make use of it when null is not an option than that is fine. It should probably issue a warning when substing in the default value?
jgarnett: You could add an isValid() method; we have done something similar in the past.
jgarnett: you have a trade off between making methods on the interfaces
jgarnett: (and having to write down the contract for them in javadocs so everyone gets it right)
jgarnett: or making methods as static utility functions; that just depend on the interfaces
jgarnett: so there is no chance of implementors screwing it up.
jgarnett: For this first cut
jgarnett: (ie 2.5)
jgarnett: I would like to keep the interfaces as small as possible
aaime: ah, now I get it, thanks for explainig
jgarnett: after we have some experience on the ground we can consider dragging some of the more populat methods into the mix.
jgarnett: (and when we do the javadocs will say what static utility function is called - ie a very strict definition; but still allowing for optimization - the one reason to place methods on a class api after all)
jgarnett: jgarnett: jdeolive mode off
jgarnett: aaime was that the discussion you needed? if so we really should move on ...
aaime: more or less
aaime: I mean, no one spoke their minds about default values
aaime: but I'm not forcing anyone to do so
aaime: we can go on
jgarnett: hrm; my mind was already spoke
jgarnett: 3) svn cut off
jgarnett: how does the end of the month sounds?
aaime: sounds good, what about snail mail issues?
jgarnett: ie everyone who has not sent in their paper work is shut out
aaime: you cut people that did not give you confirmation by mail right?
jgarnett: we can take peoples word that they have sent in the mail
jgarnett: at least for a few weeks...
aaime: (i.e., there is no guarantee that my mail will get there in time, I waited some packages from Amazon for over 2 months)
jgarnett: we don't need to be mean.
jgarnett: we just need to keep moving.
jgarnett: - http://docs.codehaus.org/display/GEOTOOLS/Graduate+from+OSGeo
desruisseaux: I'm fine with end of March cutoff.
jgarnett: the list is going okay; we have 2 rejections on hand, and david adler is talking to IBM
aaime: 2 rejections?
jgarnett: both rejections allow us access via LGPL so it is not a stress.
aaime: should we be worried?
jgarnett: Byrce always had to reject; his work is in the public domain.
desruisseaux: Which part of the code is affected by the rejections?
jgarnett: David Zwiers has always been clear about not signing (c) away.
jgarnett: we can do an audit and see
jgarnett: audit tools:
jgarnett: - http://cia.vc/stats/project/geotools
jgarnett: - http://www.ohloh.net/projects/3405
jgarnett: we also got a few "huh?
jgarnett: messages from the likes of FrankW who cannot remember what they contributed.
aaime: if it's just an ant build file
aaime: it's not there anymore anyways
jgarnett: audit of david zwiers: http://www.ohloh.net/projects/3405/contributors/14626511142519
aaime: ok, can you explain me that lgpl thing?
aaime: since he's done tons of commits
jgarnett: lets use bryce as an example
desruisseaux: The license stay LGPL, but the copyright is not assigned to OSGEO for the code wrote by someone who rejected the assignment.
jgarnett: bryce releases work in the public domain; we can make use of that in our project and do anything with it - including making it available under the LGPL.
aaime: jgarnett, that case is clear
aaime: but I'm not sure about David's one
groldan: btw, anyone you know on the most active java developers list: http://www.ohloh.net/languages/5
jgarnett: for david zwiers the work he did while at refractions is covered by the refractions signing the OSGeo document
jgarnett: it looks like he has done 6 commits since then.
groldan: go wolf go!
aaime: groldan, no, nobody
aaime: jgarnett, ah, good
jgarnett: yawn; I am pretty tired of this osgeo grind; suppose it had to be done regardless - and the problems were ours beforehand.
jgarnett: so turnning off the svn tap at the end of the month is good
groldan: yup
aaime: sure
jgarnett: after that we can sit down and update the headers / aka providence review part 2
jgarnett: okay thanks for the meeting everyone - and happy hacking.
desruisseaux: Thanks for driving the meeting Jody.
jgarnett: heh; tanks for getting us out of the driveway.
groldan: and share some kudos
jgarnett: (doh; I am full of type mumbles today)
jgarnett: I will post the logs.