IRC 2006 01 27

19:28 madair: hello all
19:28 mattdiez: greetings
19:28 camerons: Hi
19:29 camerons: ehud, would you like to introduce yourself?
19:29 ehud: Sure...
19:29 ehud: I'm working on a project to create some free maps in Israel.
19:30 camerons: CTweedie is a Mapbuilder user is Perth, Australia. I'd be surprised if he is awake.
19:30 ehud: For the purpose of the project I am looking into mapbuilder and already asked a couple of questions
19:30 ehud: in the mailing list.
19:31 ehud: I had some design questions which you already gave some answers to.
19:31 camerons: Yes, I haven't answered your last questios, (Ive just woken up).
19:31 madair: great, once there is something for us to look at, let us know
19:31 ehud: I already got the functionality I wanted up and running, however, it's very specific for my needs.
19:32 camerons: ehud, it sounds like you are proposing to do a bit of development for your application?
19:32 camerons: Are you a programmer?
19:33 ehud: Yes, I'm a programmer. I have some experience in C++, Java but not much in web programming.
19:33 camerons: Patrice and Steven are not coming. So lets start.
19:33 camerons: Anyone want to add topics to the agenda mailed out?
19:33 camerons: Or suggest a topic to cover first?
19:34 madair: where to patch things
19:34 ehud: I'm interested in the 1.0 status.
19:34 camerons: Yes, let's talk subversion.
19:36 * mattdiez is digging through Cameron's old propositions for branching, tagging, and such.
19:36 madair: is branch/1.0 created yet?
19:36 camerons: Ehud, what is your interest in 1.0? You want to help us get there, or want to know an expected date.
19:36 camerons: Ok, first topic: Subversion.
19:37 mattdiez: mike, no branches appear created, so far.
19:37 camerons: I appologise for not putting more effort into setting this up.
19:37 camerons: I tried to create a branches/b1_0 directory, but it doesn't seem to have worked.
19:37 madair: I think an easier strategy is to create tags off the trunk
19:37 camerons: And I haven't debugged to find out why.
19:38 mattdiez: I'm with mike on this.
19:38 madair: and no commits to trunk except 1.0 bug fixes
19:38 madair: develop on a branch or just don't commit iuntil 1.0 is out
19:39 camerons: This seems to be Linda's comment too.
19:39 madair: it's a pain to make the fix in the 2 branches
19:39 camerons: Has anyone been adding features to the trunk - or is it only bug fixes?
19:39 madair: (even if I did it on the wrong one)
19:41 madair: I think it's just bug fixes from me
19:41 mattdiez: (Perhaps a) Stupid Question: Are we serious about branching? Are we gutting/refactoring some key code here?
19:41 camerons: Mike, once we have Subversion worked out, we should have core development on the trunk, and only bug fixes in the branch, but I can see that we probably should make an exception for the case of rc2.
19:42 madair: but it's a real pain that way
19:42 camerons: Matt, Patrice is changing the way rendering is done be refactoring the Layers.
19:43 madair: I think we can ask developers to not commit unstable code from the first rc to the final release
19:43 mattdiez: cameron: Alright then
19:43 madair: and Pat can continue to do that, just not commit
19:43 camerons: I don't think Patrice has committed his code to the trunk yet, so we could ask him to develop in the sandbox directory for the moment.
19:45 mattdiez: Cameron: Still, we'd have to keep a list of bugfixes (+ features?) that were modified so that Patrice (or one of us) would merge those in the sandbox.
19:45 camerons: I acknowledge that it is a lot more work maintaining 2 branches. Consequently, we should only do major bug fixes on the branch to make it stable.
19:45 madair: yes only ones identified as such in the JIRA
19:46 camerons: Matt, yes.
19:47 mattdiez: Can we perhaps predict what might be broken when Pat makes the rendering changes? We run a risk of a whole slew of new bugs arriving as well.
19:47 mattdiez: And, I'd hate for us to then monitor 2 branches of bugs.
19:47 camerons: Right. That is why Pat's code should not be added to b1.0.
19:48 mattdiez: I'm okay with the proposal of only JIRA listed bugfixes in the trunk, and pat developing in the sandbox.
19:49 mattdiez: As long as we can get buyin from pat (or somebody) on merging those bugfixes.
19:49 camerons: The key question is do we want to let devellopers continue to add features while we are making a release stable or not.
19:49 camerons: If so, we need to support branches.
19:50 madair: no I don't think new features should be added, just legitimate bug fixes
19:50 camerons: And if we do support branches, we should use standard practice and develop in a branch directory.
19:50 mattdiez: in this case, i'm with mike. at least until mapbuilder gets substantially more complex.
19:50 camerons: By the way, I'm talking in general now (for 1.1 release and beyond).
19:50 mattdiez: (which is any arbitrary point)
19:52 camerons: For this release, let's make sure no new functionality has been added, then strike a 1.0rc2 from the trunk.
19:52 mattdiez: Absolutely.
19:52 madair: ok with me
19:52 mattdiez: Do we have an estimate on completion from pat?
19:53 madair: there's only 5 left for 1.0 and I think Steven was working on some
19:53 camerons: From RC2 I'd like to switch to mantaining 2 branches again. With the 1.0 branch starting from branches/1_0
19:53 camerons: Mat, I haven't heard how Pat is going. Steven has fixed half but not committed.
19:54 mattdiez: cameron: would you like me to start a practice of building both branches, or just trunk?
19:54 camerons: I'd like to build a 1.0 branch and the trunk.
19:55 camerons: Mind you, I think that we shouldn't do much work on the 1.0 branch at all.
19:55 mattdiez: can do.
19:55 camerons: We should only go back and fix the occasional MAJOR bug we may find down the track.
19:56 madair: but the release is creaated from the branch1.0 no?
19:56 camerons: This should reduce the overhead of maintaining 2 branches.
19:56 camerons: Mat, yes:
19:57 camerons: We will create branch1.0 from the trunk, then create tags/1.0rc2 from branch1.0
19:57 camerons: Make sence?
19:58 madair: yes
19:58 camerons: Matt, are you interested in working out how to do this from Subversion? I thought I knew how to do it, but something didn't work for me.
19:59 mattdiez: Cameron: I'd be glad to do so.
19:59 camerons: It seems the branches/1_0 directory I created didn't get copied into the repository or something.
19:59 camerons: Because others can't check it out.
20:00 mattdiez: Cameron: I'll see what I can come up with.
20:00 camerons: Excellent, thanks.
20:01 mattdiez: Give me about 2 days to get started.
20:01 mattdiez: Schedule's a little hectic (home remodeling!)
20:01 mattdiez: (then again, it always is)
20:01 camerons: Anything else to discuss regarding Subversion?
20:01 camerons: Matt, no problem.
20:02 madair: no
20:02 mattdiez: i'm good.
20:02 camerons: ehud, are you able to wait 30 mins or so before we discuss your design issues?
20:03 camerons: Next topic: 1.0 status.
20:03 camerons: Anything to report here?
20:04 madair: does any one know which issues steven is working on?
20:04 camerons: Matt, I think the technical role you have taken on is "Release Manager".
20:04 mattdiez: For vanity's sake, I prefer the term "Build Engineer"
20:04 camerons: Mike, I think Steven assigned the bugs to himself that he was working.
20:05 camerons: Mat, that is fine.
20:05 mattdiez: But, any nomenclature is fine with me.
20:05 madair: ok makes sense. I'll have a look at the other one
20:06 madair: one other task for 1.0 is to redirect sourceforge home page, or clean that one up
20:07 * linda just joining late ... reading through session
20:07 mattdiez: Doesn't steven take care of those things?
20:07 camerons: Matt, as Build Manager, would you like to monitor what bugs have been fixed, what is outstanding, make sure that someone is assigned to each of the bugs assigned for 1.0?
20:07 madair: that home page might confuse people if the home page being maintaned is the codehaus one
20:07 camerons: Welcome Linda.
20:08 camerons: We have agreed to develop 1.0 on the trunk until rc2 (as I think you suggested via email).
20:09 camerons: Linda, do you know if Patt has added his new Layer redesign and RSS work to the trunk yet?
20:09 mattdiez: Cameron: I'll try to keep up with that.
20:09 camerons: Or the status of Pat's bug fixes for 1.0?
20:09 ! linda [n=lderezin@pcp0010489598pcs.essex01.md.comcast.net] has left #mapbuilder []
20:10 camerons: Matt, thanks. It is quite a bit of work, but very important if you want to keep a build on track.
20:10 ! linda [n=lderezin@pcp0010489598pcs.essex01.md.comcast.net] has joined #mapbuilder
20:11 ! linda [n=lderezin@pcp0010489598pcs.essex01.md.comcast.net] has left #mapbuilder []
20:11 mattdiez: Well, linda's acting strangely
20:12 camerons: Mike, yes it would be good to clean up the SF home page. Unless someone has some free time, I suggest we focus on the documentation after we have frozen the code for RC2.
20:12 camerons: We usually have a week or more between freezing the code and getting the build together.
20:13 madair: ok, I was thinking about before a 1.0 announcement that would attract attention to it
20:14 camerons: yes.
20:15 camerons: Anything else to discuss on RC2. Do we want to walk though the bugs, or wait until next week?
20:15 madair: i picked up MAP-44 Steven has the others so I think we're ok on that
20:16 camerons: ok
20:16 mattdiez: I need to go input the timeseries bug. I wonder if patrice's code will make that easier.
20:16 madair: matt: do you mena MAP-44?
20:17 camerons: Matt, would a 1.1 timeframe for Timeseries bugs be acceptable for you?
20:18 mattdiez: mike: yeah. forgot that was already in.
20:19 madair: ok, I'll let you close it then
20:19 mattdiez: Mike: It's fixed?
20:20 mattdiez: I see it's related to Map-6
20:20 mattdiez: but, both are still open.
20:20 madair: no it's not fixed, it's a layerId thing
20:21 madair: I'll have a look at that one this weekend
20:21 camerons: Looking at the 1.0 release in JIRA, we only have 5 bugs left to fix in this build. Is that about right?
20:22 mattdiez: Cameron & Mike: Hey, I'm delighted anybody smarter than me is willing to work on it.
20:23 madair: sorry, if you are working on it, please go ahead.
20:23 madair: I'm just anxious to get the release out
20:23 mattdiez: mike: I'd prefer if you took it.
20:23 madair: ok
20:24 madair: steven has the other 4
20:25 camerons: Anything else to discuss for RC2?
20:26 camerons: Next topic: OpenGeospatial Foundation
20:26 mattdiez: To all: I'd rather duck out of the politics and organization part of this discussion.
20:27 camerons: No problem. anything else you would like to discuss before you leave?
20:28 camerons: Meeting time next week: Same time same place ok with everyone?
20:28 mattdiez: Nope. Havin' a good time. I'll be working on branching it in the next 3 or so days.
20:28 mattdiez: And, of course, marking bugs and such. I'll email if there are probs.
20:28 camerons: Great. I think you have enough to keep your hands full.
20:29 camerons: Mike, one issue brought up last week is "Who should own the Mapbuilder code"?
20:30 madair: i think the main issue with the foundation is copyright assignment
20:30 camerons: Right.
20:30 camerons: What is your take?
20:30 ! mattdiez [n=mattdiez@65.125.90.210] has left #mapbuilder []
20:30 madair: it's helpful to have that lawyer explain things
20:30 madair: can there not be joint copyright?
20:31 madair: the original author plus the foundation?
20:31 camerons: Pat and Steven didn't seem keen to give the code to the Foundation and allow the Foundation to relicence our code for commercial gain.
20:32 madair: I tend to agree with tht
20:32 camerons: And I tend to agree with them too.
20:33 camerons: I would be prepared to assign copywrite to our PMC but I don't think that is an option as the PMC is not a legal entity.
20:33 madair: LGPL say we should have copyright statements
20:34 camerons: You mean in the code?
20:34 madair: not sure, but in at lesat one place
20:35 camerons: We do have a reference to the LGPL - but not the full text as the LGPL recommends.
20:35 madair: something to look into
20:35 camerons: We probably should have a link to the LGPL from mapbuilder/index.html too.
20:36 camerons: The other issue I have is over control.
20:36 madair: so we agree that our position is to not hand over copyright?
20:36 camerons: I think so, we should run it past Steven and Pat first.
20:36 madair: ok
20:37 camerons: And let the foundation know where we stand and give them a chance to try and convince us otherwise or propose an alternative solution.
20:38 camerons: The Foundation seems keen to assign a Foundation member to the Mapbuilder PMC.
20:39 madair: the control of the foundation is an issue but it's a safe bet that most folks involved have a good viewpoint on open source software so it's not such a big deal to me
20:39 camerons: I think that the PMC should decide who is on it, not the Foundation.
20:39 madair: I have to step out for 10-15 minutes but I'll be back
20:40 camerons: Flipped the other way - if our PMC gets to put a PMC member into the Foundation board, then I'm happier with that.
20:41 madair: me too
20:41 madair: brb
20:43 ehud: camerons, are you still here?
20:45 camerons: yes
20:45 camerons: Just emailed an aswer to your first question.
20:46 ehud: I thought you were sending me private messges along the irc discussion and I tried to reply, wasn't that the case?
20:47 camerons: I've only got a couple of private messages from you. I haven't sent any technical responses.
20:48 camerons: reading your second email ...
20:51 camerons: Can you populate your HTML Form with only the contents of the FeatureCollection, or are there extra fields in the FORM being populated from elsewhere?
20:51 ehud: The form is populated only from the FeatureCollection.
20:52 camerons: ok, so you can use a Model/View/Controller pattern then, with the FeatureCollection being the Model.
20:52 ehud: Yes, that is what I did.
20:53 camerons: Is that an acceptable solution?
20:53 ehud: As I wrote in the email, it seems a bit wrong that my FeatureList widget sends data to the server.
20:54 camerons: You mean it is not a "pure" MVC design?
20:54 ehud: Yes.
20:55 ehud: My solution works fine, but I'm trying to figure out the right solution, so I can contribute the code back to mapbuilder.
20:55 camerons: If you wanted to be pure, you can emulate the current "InsertFeature" code.
20:55 ehud: Another issue, is that if you change the form, it does not update the model (as it sends the data by itself).
20:56 madair: it does seem to me that the feature button widgets have too much Model properties in them
20:56 camerons: The "InsertFeature" will use XSL to build a InsertFeature transaction, then send it to the server. But it is a lot of extra work with minimal (or no) gain.
20:57 camerons: So I suggest you have an appropiate solution.
20:57 madair: I would move webServiceUrl property to the targetResponseModel as defaultModelUrl
20:58 camerons: ehud, the current FeatureList will update the FeatureCollection when it is updated (using a bit of black magic).
20:58 madair: leaving model handling code in the models
20:59 camerons: The FeatureLists notices when a field changes, then calls the FeatureCollection to update the field.
20:59 ehud: The problem is that my FeatureLists can't update the FeatureCollection.
21:00 camerons: Edud, why?
21:00 ehud: My FeatureLists actually displays a single form which is a unity of all the Featurecollection items. If it changed the FeatureCollection, it would override all the items.
21:01 ehud: This would work if you change some data in the form and submit, however, if you change some data and then add some more features to the FeatureCollection,
21:02 camerons: Mike, I'm still thinking about your comments. I might take this up with you later.
21:03 ehud: Thanks for helping but I'm afraid I must leave. Can we continue this through the mailing list?
21:03 madair: we've had the argument before.
21:03 madair: I just think there is too much model code mixed in the widgets
21:03 camerons: Ehud, yes, no problem.
21:04 camerons: Ehud, the key to updating the FeatureCollection is that your Form has a unique id that it can use to reference the FeatureColleciton Attribute.
21:04 camerons: Maybe that is broken.
21:06 camerons: Mike, I'm concerned that if we start discussing this design now we will never finish.
21:06 camerons: Can we put it off until another time?
21:07 madair: sure
21:08 madair: anything else?
21:08 camerons: I'm done.
21:09 camerons: shall we wrap up then?
21:09 madair: sure. Just fixed MAP-44
21:09 camerons: well done.
21:10 camerons: ok, thanks for all the work you have been doing on RC1.
21:10 camerons: bye.
21:10 madair: no problem, bye
21:11 madair: exit