* Now talking on #mapbuilder
* Topic for #mapbuilder is: Community Mapbuilder channel | files: mapbuilder.sf.net | docs: docs.codehaus.org/display/MAP | dev-meeting: every tuesday 1800UTC
* Topic for #mapbuilder set by stvn at Thu Dec 29 22:07:53 2005
* #mapbuilder :[freenode-info] if you need to send private messages, please register: http://freenode.net/faq.shtml#privmsg
CameronShorter hello
revenazb Hi
madair hi
Cappelaere hey
CameronShorter Is Pat awake yet?
CameronShorter ok
Cappelaere but it is nap time here!
CameronShorter Have you started yet, or shall we start now?
madair not started yet
CameronShorter What do people want on to discuss?
madair 1.0 final schedule?
Cappelaere 1.1 schedule
Cappelaere 
CameronShorter Bertrand Reveraz, I assume you would like to discus WFS transation design issues?
CameronShorter Mike and Pat, I'd like to talk about the interoperability demo with you afterwards if we have time.
Cappelaere k
revenazb Sure, you guys can use Bert for short
madair you mean the Kentucky thing?
CameronShorter Mike, yes.
madair I have a meeting at 2:30, and then off winter camping until Sunday
Cappelaere what's your temp?
madair -8C
CameronShorter ok, lets, talk 1.0, 1.1, Ketucky thing, then WFS design.
madair the main thing to fix for 1.0 is IE jumpiness and the loading indicator strecthed initially
CameronShorter Bert, we probably won't get to the wfs for 45mins or so. You are welcome to sit in, or you can do someting else for a while.
madair the other 2-3 issues can wait for 1.1 I suggest
revenazb This is my first involvment with open source dev so I'll sit in
CameronShorter Mike, do you know how we can fix these?
CameronShorter bert, ok.
madair not necessarily, the loading thing we can get rid of, but IE jumpiness is trickier
madair personally I can live with that behaviour, but others find it disconcerting
CameronShorter Maybe we could move the loading icon outside of the MapPane, then it wouldn't stretch.
madair yes it can be removed completely, replaced with Loading2 widget
madair that would affect all apps using MapPane however
madair quick and dirty solution is to set width/height when it is loaded
CameronShorter How does Loading2 fix the loading issue?
madair by giving an alternate loading indicator
madair in that case, if you didn't using the Loading2 widget, you would have no loading indicator at all
CameronShorter Why is Loading2 not stretching like the original Loading icon?
madair it's in it's own DIV, the MapPane laoding indicator is in the MapPane div, which gets stretched to the Map size
CameronShorter ok
CameronShorter Mike, that seems like a reasonable fix to me. All we need to do is replace all the demos with Loading2 widgets.
madair yes
madair and advise people with apps that they have to do the same
CameronShorter yes.
madair Actually we cna probably deprecate the original loading widget in favor of Loading2,
CameronShorter ok
CameronShorter For the IE jumpiness.
CameronShorter We didn't have this problem in the early releases did we?
madair yes we did. It's pretty much always been there
CameronShorter What do you mean by Jumpiness?
madair where you can see the old layers while the new ones come in, and the whole map shifts position temporarily
CameronShorter ok
CameronShorter Maybe we need to set the old layers to NULL before we start the re-drawing cycling.
CameronShorter I mean set the HTML <div> for the layers to NULL before drawing.
CameronShorter Would that work?
madair maybe, maybe not. IE is very bizarre
CameronShorter yes,
CameronShorter Are you able to look into these issues?
CameronShorter Matt also note 2 issues. A PNG/Gif incompatability issue, and Memory problem with IE.
madair I'll check into them, but like I said in my email last night, if someone beats me to it, please go ahead
CameronShorter I'm happy to say the PNG/Gif issue is something that has a work-around (use Gif instead), but the Memory problem seems much bigger. It might mean a significant change to the way we do inheritance (and a rewrite).
madair Pat: do you see the memopry leak with the trunk (1.1)?
CameronShorter ok, thanks Mike.
Cappelaere I have not looked for it
Cappelaere do not know
CameronShorter Pat, what would the US Coastguard reaction be to a memory leak?
Cappelaere we are not operational yet
Cappelaere so np
Cappelaere so far
CameronShorter The workarounds would be: 1. Use Firefox, 2. Restart IE after a while using Mapbuilder.
madair we could put that on a list of known problems if we can't solve it
CameronShorter I think we will have to. We should aim to scope the problem. Ie - we expect the problem to occur in IE after X mins of continous use on a normal machine.
CameronShorter Matt might be able to do the research on that.
madair the other major thing for 1.0 (or sooner) is to redirect mapbuilder.sourceforge.net to the codehaus page. It's on the OSGeo home page, and will be ina press release about OSGeo
CameronShorter I suggest we ask the list for help with the outstanding 1.0 bugs. State what needs to be done to fix them and we just want a coder to put in the leg work to change them. That will give Mike time to get onto bigger issues.
CameronShorter Yes.
CameronShorter Do we want to completely redirect everything now?
CameronShorter Or are we still moving content across?
madair not sure
CameronShorter I'll ask Peter if he can take the lead on this since Steven is OS.
madair k
CameronShorter Mike do you have a suggested date for the next 1.0 release?
Cappelaere brb
madair before 1.1 
madair perhaps another week
CameronShorter ok.
CameronShorter Then a week or 2 to build the release.
CameronShorter Do we want to call the next release 1.0rc3 or 1.0?
madair I vote for 1.0, unless there is something major that comes up
CameronShorter If we call it 1.0, then we don't expect to do any more work on the 1.0 thread (except maybe a major bug fix or 2)
madair yes
CameronShorter ok, 1.0 it is. I wonder if Matt has recovered from the last build process yet and is ready to do it again?
CameronShorter I'll ask him.
CameronShorter Anything else for 1.0?
madair not from me
CameronShorter Better wait for Pat before we talk 1.1. You back Pat?
CameronShorter Bert, while we are waiting, let's start talking about your WFS-T.
CameronShorter you with us Bert?
revenazb yep
CameronShorter So you want to move some of your logic from JS to PHP?
CameronShorter That means moving state info from client to server, and I'm not sure how that will work cleanly.
Cappelaere back
revenazb I am primarily a PHP and Database person.
CameronShorter ok, let's flip back to 1.1 - I'll get back to you Bert in 20 mins or so.
madair yes the mapbuilder design is to keep the data in the client; it takes a shift in thinking
CameronShorter What are the plans for 1.1. Pat, I think you will be the driver for this build.
Cappelaere wow
Cappelaere scary thoght
CameronShorter Pat, you have already written enough code for a 1.1 release if you can get it working.
Cappelaere very very close
CameronShorter Do you have a release date in mind.
CameronShorter ?
Cappelaere nope
Cappelaere soon
Cappelaere mid March
Cappelaere end of Mar
Cappelaere is fine too
Cappelaere I would like to have georss and google maps
Cappelaere may be geotracing
Cappelaere seems straightforward
CameronShorter History shows us taking 2 weeks to create a build. So is it mid/end March for code freeze or for a release?
madair I've had a quick look, I like the graphics factory design
Cappelaere there are some good live feeds to use
Cappelaere cool
Cappelaere have you looked at the SLD?
madair 2 things though:
Cappelaere I think a lot of it is reuseable throughout
Cappelaere k
Cappelaere ?
madair in the GeoRss, I know it was a quick hack for the flicker demo (which is cool) but there is some Model code in there
madair it would be better to define a "Flickr" model and keep that out of the widget
Cappelaere ?
Cappelaere I thought it is in a model and not widget
madair The other suggestion is that perhaps you define a MapPane2 widget, so that you develop new stuff in a new widget, that way you can leave MapPane as is
Cappelaere the feed entries are added to the OWC
madair and we coudl depreacte later
madair I forget where I saw that, but it seemed to be ina georss widget
Cappelaere ?
Cappelaere MapPane has been trimmed down
madair I'll have to have a look on Sunday for where I saw that
Cappelaere functionality was moved to the MapLayerMgr
madair maybe you've moved it already
Cappelaere so we can expand there
Cappelaere I am still working on the original georss demo
Cappelaere to make it compatible with the new functionality
Cappelaere almost done
madair yes this is good, but can;'t the original MapPane stay as is, just use MapPane2 for the widget in config files
* greenwoodmap (n=chatzill@63.164.104.51) has joined #mapbuilder
CameronShorter I was thinking that a UML diagram would be useful for describing all the new classes.
Cappelaere sure but why would you?
madair sorry I have to go now
CameronShorter ok, thanks Mike. Same time next week?
Cappelaere we need to clean up a lot of code that is getting deprecated...
Cappelaere we need to tqalk more....
Cappelaere cya
madair no, next week I'm out of the office Thurs/Fri
CameronShorter Actually, can we meet 30 mins later for the meeting?
CameronShorter That is 06:00 for me.
madair fine with me, I think it's Steven who will have trouble
Cappelaere fine with me
madair bye
* madair has quit ("Leaving")
CameronShorter What do you thing steven?
CameronShorter no comment, he must agree.
CameronShorter We can change the time back when he is back from holidays.
CameronShorter Pat, I find UML diagrams useful for quickly understanding the code.
Cappelaere agree
CameronShorter It is a good way to get new programmers up to speed faster.
Cappelaere agree
CameronShorter We had an old uml diagram is mapbuilder/design/mapbuilder-lib.xmi (from memory)
Cappelaere I remember seeing something like that at one time
CameronShorter It would be useful for describing the new Factory patterns.
CameronShorter Do you have any design requirements from the US Coast codes (or coding standards) or similar?
Cappelaere no
CameronShorter lucky.
Cappelaere not requirement for UML either
Cappelaere yeap
Cappelaere it is an experimental site
CameronShorter ok
CameronShorter For your release date, it would be good to have a target release date so we can get everyone on the same page.
Cappelaere I suspect that we will have an interim demo at end of March
Cappelaere nothing formal
CameronShorter Is your project setting a release date that you need to fit into, or are you calling the shots?
Cappelaere but I woul dnot find getting it integrated back
Cappelaere I am calling the shots
Cappelaere trying to do monthly sprints
CameronShorter ok.
Cappelaere we have some big exercises to support this summer
Cappelaere for disater relief (hurricane season starts)
Cappelaere we wnat to be ready this time
CameronShorter What quality does the software need to be at for the 1.1 release. Is alpha ok (this is quicker)
Cappelaere yes
Cappelaere proof of concept
Cappelaere if they like it, they will have to spend some $$
CameronShorter ok, I think 1.0 will be coming out first week of March, how about we aim for a freeze date of 3rd week of March for 1.1.
CameronShorter 1st week of April for 1.1 release. (2 week cycle).
Cappelaere sounds good. I can even use my own SVN repository if I have to.
Cappelaere I just think we can move a little faster...
Cappelaere there seems to be a lot of itnerest recently
Cappelaere moving objects will be cool too.
CameronShorter I'd prefer it if we don't have to use your SVN. If need be, you can develop in a sandbox on mapbuidler servers.
Cappelaere agree
CameronShorter I want to stagger releases so that we don't overload Matt and the release cycle.
Cappelaere he is doing a great job
Cappelaere no rush
CameronShorter yes - building releases has a lot of small jobs.
Cappelaere I think that we should have some monthly prizes to award....
CameronShorter Matt will win - every month.
Cappelaere then we will need more prizes more often
CameronShorter We are increasing Matt's work load - we will have to make sure he has the time to support us.
CameronShorter Otherwise we will need to find another release manager.
CameronShorter I have some GoogleMap layer code to add to the trunk. When would be a good time to start work on that?
Cappelaere we need to talk about this offline
CameronShorter ok
CameronShorter Note, google maps uses fixed zoomLevels, so we need to address our zooming tools at the same times.
Cappelaere I realize that ;(
CameronShorter Unfortunately, the tools are a bit messy.
CameronShorter Oh - another problem:
CameronShorter Google Maps uses a projection that doesn't have an EPSG code, which means that you can't ask for a WMS overlay in the same projection.
Cappelaere that's probably ok. Don;t they use WGS84?
Cappelaere whatever...
CameronShorter For local areas you can use Lat,Long and it is pretty close, but it is totally out for a Map of the world.
Cappelaere that's fine
CameronShorter no. Mercator (from memory)
Cappelaere I think that's what it is
CameronShorter I've been meaning to flag this with Google/WMS communities to get it addressed.
Cappelaere can;t wait to see what transpires of Google joining OGC
CameronShorter yes.
CameronShorter How long do you think it will take to get the trunk stable and ready for others to test/work in again?
Cappelaere it looks fiarly good right now
CameronShorter ok, I have not synced for a day or 2.
CameronShorter By the way, I've just finished applying for a grant to develop a Bicycle Mapbuilder website. Will start in June if I win it.
CameronShorter About .75 man years work over 2 years.
Cappelaere cool
greenwoodmap And how much of that 0.75 MH is in the saddle?
CameronShorter ?
greenwoodmap riding one's bike.
CameronShorter Almost all of it.
greenwoodmap way cool!
CameronShorter oh - almost all of it at my computer desk.
greenwoodmap retract the "way cool!"
CameronShorter writing code and supporting user community.
CameronShorter
Cappelaere computer is powered by the bike
CameronShorter lol
CameronShorter Anything else to discuss re 1.1?
Cappelaere I have some ideas on geotracing...
CameronShorter Shoot.
Cappelaere shoud be trivial now...
Cappelaere we would treat the data as georss feeds
CameronShorter What is geotracing? Is that Feature Entry?
Cappelaere all entries end up in the OWC model
Cappelaere moving objects
Cappelaere ships
Cappelaere cars
Cappelaere bikes
Cappelaere with GPS
Cappelaere check put APRS sites
Cappelaere and geotracing sites
CameronShorter Are they collecting data into a database, or only publishing it?
CameronShorter Ie, are we putting data into a WFS-T?
Cappelaere so we add the entries in the OWC context (so it can be used in layers...we need to have grouping capabilities too..)
Cappelaere nope
Cappelaere the data is already out there in servers
Cappelaere I just need to write a little client to format the data properly
CameronShorter OWC is an extension to the WMC isn't it?
Cappelaere might be able to do it in javascript
Cappelaere yes
Cappelaere as layers
Cappelaere rss layers
Cappelaere annotations... down the road
Cappelaere in our case, movingobject entries from rss/atom feeds
CameronShorter The the OWS store a list of points, or does it contain a URL to an XML feed?
Cappelaere wthey are identified
Cappelaere OWS store an entry
Cappelaere a feed has many entries
Cappelaere one entry per layer right now
Cappelaere until we have grouping
CameronShorter Do you have timestamping of points?
Cappelaere so we can group all moving objects, fixed objects, emergency vehicles...
Cappelaere yes
Cappelaere time and location and id
CameronShorter ok
Cappelaere we refresh on a regular basis
Cappelaere reacquir ethe points and/or add new ones
Cappelaere and adjust the location
CameronShorter So you have one layer for each track (or id)?
Cappelaere yes
Cappelaere so we can keep a history or recent position
CameronShorter ok
CameronShorter Or create a movie loop.
Cappelaere http://www.aprs-is.net/javalive.htm
Cappelaere brb
Cappelaere back
Cappelaere I am trying to localize the code to parse the data...
CameronShorter ok
CameronShorter I can see I need to catch up with you and check out your code.
Cappelaere woul dmake a cool demo and I need the technology for emrgency response vehicles...
CameronShorter ok
CameronShorter Is this going to be for 1.1 or later?
Cappelaere may be 1.1. I do not think it will be too hard to do once I figur eout the parsing
CameronShorter ok
CameronShorter Is this part of the OWS spec or are you extending it?
Cappelaere I have to extend it for RSS layers anyway
Cappelaere so might as well see how far...
CameronShorter OWS allows you to extend it or does you extensions make your doc invalid?
Cappelaere invalid from what standpoint?
Cappelaere there is no schema
CameronShorter ah.
CameronShorter I meant to you break the schema.
CameronShorter Some OGC docs provide an extention tag where you can put your own extensions.
Cappelaere I think these extensions ought to be standard
Cappelaere annotations is another extension I want to push
CameronShorter I tend to agree with you - but it would be good if we were in discussion with OGC owners of this doc and get their input on what we are doing.
Cappelaere they have just release GML/JPEG2000 which has annotation
Cappelaere gotta join OGC
CameronShorter yes - talk later about that.
Cappelaere I am planning to write a paper for GeoWeb 2006
CameronShorter I'd like to give Richard and Bert a run - anything else for you Pat re 1.1 etc?
Cappelaere nope
CameronShorter Richard, do you want to give us a quick run down on Proj and where you are at.
greenwoodmap I sent a draft RFC to Mike and CC'd you. Haven't heard back from Mike.
Cappelaere you might win the next prize
Cappelaere great job!
CameronShorter ah, I don't remember that email.
Cappelaere I still need your picture
Cappelaere and geolocation
CameronShorter I been a bit swamped for the last 2 weeks - maybe I missed it.
greenwoodmap I'll send it again. I beleive I sent it Sunday.
greenwoodmap And I'll get a pic together soon, thanks.
CameronShorter ok - I suggest sending it to the developer list. Usually best to be open with these things.
CameronShorter You get more feedback and people don't feel like they are being excluded.
CameronShorter I'll look out for your RFC and we can work it from there.
greenwoodmap yeah, I agree, but Proj was Mike's baby, so I wanted to give him a chance to comment before I threw it out there. I don't want to be stealing it away or anything.
CameronShorter Yes, sounds good and I agree.
CameronShorter You can run it past Mike first if you like.
CameronShorter Anything you want to discuss about Proj now?
greenwoodmap Not really, I'm re-sending the RFC
CameronShorter ok,
CameronShorter I'm getting bugged to have breakfast with the family.
CameronShorter Bert, I'd like to talk with you - are you free later today?
CameronShorter What country/timezone are you in Bert?
Cappelaere send us your geolocation and pic
Cappelaere pleeeeaaaaaase
CameronShorter Pat is setting up a demo with pics of all the developers.
greenwoodmap yep, will do.
CameronShorter Probably would be a better demo if you geolocated all the "Miss World" contestants instead of us ugly programmers.
CameronShorter You with us Bert?
revenazb Sorry guys stepped away from a sec, I am in Montreal Canada, timezone is eastern US GMT-5
CameronShorter I was about to ask to meet later today - but my wife said I can keep going now.
revenazb I have to go in about 10min but that should be enough time
CameronShorter Bert, I'd like to work out a design for you.
CameronShorter I suspect that PHP would not be appropriate.
revenazb Just to give you a little background, I work for the world bank and we are trying to give developping countries cheap easy to use GIS tools to manage road infrastructure
CameronShorter Couldn't be more worthy a cause. I love it.
CameronShorter Pat, give this man a medal.
CameronShorter Have you looked at the Feature Entry demo?
revenazb The main issue we are running into (besides lack of reliable data) is permissions and history tracking
revenazb Yeah I looked at it, our feature entry needs are minimal, points and editing attributes of existing features
Cappelaere on its way
CameronShorter Permissions reading data, or permission writing?
revenazb we do not need the ability to edit the geometry of features expect for the creation of ne point features
revenazb reading data permission will be done at the application level, the writing is where it gets tricky
CameronShorter How many users will you have?
CameronShorter (with write access)
revenazb we need multiple editor ACL, to limit editors to certain features, we also need to track changes and changes need to be subsequently approved by an auditor before being published
CameronShorter Is it small enough to handle with apache security. (using .htaccess files)
Cappelaere This is actually another topic of mine... security and identity 2.0
Cappelaere a killer
revenazb an example might be an editor in a given province responsible for bridges is only allowed to update attributes for bridges in that province
Cappelaere way beyond MapBuilder-lib
revenazb And any other WMS/WFS-T client, proprieatary or otherwise
revenazb I think it is even even beyond the WFS-T spec
CameronShorter The security needs to be applied at WFS-T but Mapbuilder would need to support the security.
revenazb hence the idea of switching to PHP and doing directly on the database
revenazb I say PHP simply cause I am familiar with it
revenazb I was thinking of hook into mapbuilder that would return the coordinates of AOI or a Point in a way the is accessible to PHP
CameronShorter PHP might be the way to go, but I'd first see what you can do with WFS-T. If we can stick to standards, then we stay a long way ahead.
Cappelaere PHP is on the server side so it is orthogonal to MapBuilder
revenazb I agree standards are definetly better and infinetly more reusable
Cappelaere what is important is to use the right standards for security
Cappelaere problem is that it gets complicated fairly quickly
Cappelaere you need to authenticate the user
Cappelaere and assign a role
CameronShorter I heard a while back that the OGC were doing a pilot with WFS and security. It would be worth researching this and seeing where they got to.
Cappelaere then assign policies for that role
Cappelaere you can use xacml for that
Cappelaere some of the interactions might involve SAML 2.0
Cappelaere and that gets dicey
Cappelaere but doable
Cappelaere we are working in that area right now but not for the map
Cappelaere so you need identity 2.0
Cappelaere XACML 2.0 + SAML 2.0
CameronShorter I think that you can use standard Apache security to restrict access to a WFS-T.
Cappelaere then we need to agree on the interfaces
Cappelaere not enough
revenazb and change tracking and auditor approval
CameronShorter What might work is to set up a number of WFS-Ts for each group of users.
Cappelaere depends on use case I guess.... sorry
Cappelaere then you have a sync issue
revenazb I guess maybe what we are trying to do is more of a database application that has a map interface to select features then a full mapping application
Cappelaere yes
CameronShorter Chris Holmes looked into history with WFS-T a while back - from memory, he was suggesting that history be maintained as an attribute for a feature.
Cappelaere yu need a solid backend
Cappelaere with clearly defined policies
Cappelaere on the client side
Cappelaere we need to properly identify the user/roles when we attempt the transactions
Cappelaere the WFS-T is a web service that needs to follow the policies
Cappelaere there ought to be a PEP (Policy Enforcement Point) to go to and say Y/N
Cappelaere a fairly minor change to the WFS-T for someone who knows how to do it
CameronShorter Another option: The OpenStreetMap project have created a Feature entry Java applet to enter features into a database. They don't use any standards, but we could write a widget which interfaces to their database.
Cappelaere How do they address user-based access control?
CameronShorter I'm not sure.
Cappelaere actually role-based ACL?
revenazb Sorry to cut this short but I have to leave
Cappelaere I have been looking at this for my project
Cappelaere k
Cappelaere we have data we cannot make available to all
revenazb I'll make sure to read the transcript of the conversation
CameronShorter I've been wanting to integrate with the OpenStreetMap project for a while, but they haven't answered any emails.
Cappelaere so it is filtered before it makes it to the map by hte PEP/search engine
CameronShorter I also want this for my BicycleMabuilder project.
Cappelaere
CameronShorter I'm sure there are others who would want it too.
Cappelaere sure
Cappelaere but this is intense server side
Cappelaere light on client side
Cappelaere we might have to use the new SOAP interfaces defined by OGC
CameronShorter It is possible we could write a PHP wrapper on top of the WFS-T but I know the WFS-T guys also want history - so we should be able to get a proper solution.
Cappelaere SOAP supports security
CameronShorter Pat, are you free to talk again in 22 hours?
Cappelaere I would write a service that do the filtering so you do not change the WFS-T
Cappelaere kinda of a proxy with security
CameronShorter ok
CameronShorter Could be useful to have someone with strong PHP skills to help with this.
CameronShorter Bert, I don't suppose you know JSP too do you?
Cappelaere I probably woul dnot do it in PHP but Java
Cappelaere I do
CameronShorter Geoserver is java based.
Cappelaere You could leverage the SUN PEP/PDP framework...
Cappelaere http://sunxacml.sourceforge.net/faq.html
CameronShorter ok
Cappelaere you still need to do a lot of work with policies... LDAP...
Cappelaere may be
CameronShorter I'm going to have to go too.
Cappelaere again, depends on use-case
Cappelaere k
Cappelaere cya
Page Operations
Browse Space