Skip to end of metadata
Go to start of metadata


04 <jmacgill> ignore myulast email, I actualy quite like Jody's new suggestion of a plugin per broken server
05 <rgould> yeah?
06 <rgould> the WMS section is divided up into version
06 <rgould> versions
06 <rgould> currently there are four of them
06 <rgould> our idea for mis-behaving servers was to implement a subclass of a version for that specific server
06 <rgould> so splitting it up into separate plugins is not a bad idea
06 <jmacgill> makes getting rid of it when the server is fixed eaiser
06 <rgould> an interesting thing about the WMS capabilities are the contact elements
07 <rgould> Yes, that is true
07 <jgarnett> Hi James is it meeting time?
07 <jmacgill> humm, so the warning could say 'broken sever, contact ...'
07 <rgould> Often a name and an email address is listed in the contact. We could even get a UI widget going that would enable the user to email the system administartor informing them of their standards deviance
07 <jgarnett> It is monday right? Hard to tell after last weekend.
07 <rgould> and even provide an error log
07 <jmacgill> I think the meeting time is 3:30
08 <jmacgill> i.e. in about 20 mins
08 <rgould> BTW, I am logging this
08 <jmacgill> cool
08 <jmacgill> I think this new client can, but not worked out how yet
08 <rgould> which one are you using?
09 <jgarnett> I can just imagine the response to that - "The 'UDIG Demo WMS' does not support WMS Post! hit 'yes' to email 'Paul Ramsey'"
09 <jmacgill> irssi
09 <rgould> hahahhaa
09 <jmacgill> (smile)
10 <rgould> Part of the difficulty with WMS is the number of very very very subtle differences between versions
10 <jgarnett> 'UDIG Demo WMS' appears to be using geoserver, Hit 'Yes Please' to email Chris Holmes.
11 <jgarnett> You guys are both right, what did you think of the idea of plugins?
12 <rgould> Do you want one plugin per deviant server?
12 <jgarnett> I know richard already has plugins for different versions.
12 <rgould> they are not plugins
12 <rgould> they are classes
12 <jgarnett> We could place a ext/wmsdevient together for all the bad ones.
12 <rgould> or maybe I'm getting the terms mixed up
12 <jgarnett> But you can switch them over to to use a FactoryFinder and allow others to implement new WMS versions (or WMS devients) for you.
13 <jgarnett> Then geotools proper can work right, geotools+evil extention can work more often"
13 <jgarnett> James?
13 <jmacgill> hi sorry, back
13 <jmacgill> yeah, one plugin for all broken ones makes sense
13 <rgould> ok
14 <rgould> in ext?
14 <jmacgill> but that would mean that we would have to have a way to order our plugins
14 <jmacgill> as both would claim to be able to talk to a WMS
15 <jgarnett> We have the same conflict right now with datastores
15 <jgarnett> nobody notices because they all talk to one directly.
15 <rgould> This might take me a while to implement
15 <jgarnett> Basically we would ask the ext to try override the default (based on URL patterns), often say geoserver is has "geoserver/wms in its path
16 <jgarnett> mapserver has a similar deal, but james is right. No sense in confusing your WMS_1_0 class with hacks for different versions.
16 <jgarnett> make the hacks explicit with there own class.
16 <rgould> I was thinking it could be done from a clients point of view? if they want to be flexible, they would use the flexible WMS, otherwise they would use the standard one?
16 <jgarnett> Too bad he did not give us the pep talk 6 months ago (smile)
17 <rgould> Yea
17 <jgarnett> But choosing flexable is not enough, you will never make one with all the possible hacks that will be stable.
17 <jgarnett> See jame's origional point.
18 <jgarnett> Richard is there anything in the get capabilities doc or url we can use to figure out which kind of server we are talking to?
18 <jgarnett> And how many hacks are we talking about anyways?
18 <rgould> Sometimes. There is nothing in the DTD/Schema that states that
18 <rgould> but MapServer and CubeWerx often include comments identifying themselves
19 <rgould> I'm not sure. There are quite a few.
19 <rgould> MOst of the servers out there are not compliant
19 <rgould> CubeWerx is the only one, I believe.
19 <jmacgill> a lot this debate comes from the cubewerx client, it was THE one that people tested against (including me)
20 <jmacgill> and I thought that when my server was working with the cubewerx client I was sorted
20 <jmacgill> sadly they just had work arounds for the most common mistakes
20 <rgould> right
21 <jmacgill> 'MOst of the servers out there are not compliant' is something that I think uDIG can help resolve if the error dialog is helpful and anoying enough
21 <rgould> right
22 <rgould> This sounds like a good challenge (wink)
23 <rgould> I think I will try and break out some design later today
23 <jmacgill> We may need a fallback system for datastores
23 <jmacgill> i.e try A but at the first sign of trouble try B
23 <rgould> Yeah
23 <rgould> that's sort of what I was thinking
25 <jmacgill> with the NY meeting it may be hard for people to join us today
27 <jmacgill> the UN had a tight firewall last year
29 <jmacgill> ok, assuming we are on our own...
29 <jmacgill> lets chat a little about 2.1M2 and perhaps a way for fallback datastore implementations
30 <jmacgill> Jody, is there anything specific you want accomplished before 2.1M2?
31 <jgarnett> Nothing I want to make it today (smile)
31 <jgarnett> To match the shinny new 0.7 we are working on.
31 <jmacgill> could you run the header replacer first?
32 <jgarnett> It is an Mx repelase so I don't care much for quality - just archiving and communicating that this project is very much active.
32 <jgarnett> ie run your script?
32 <jmacgill> yeah, but as you say, less important for an Mx
32 <jgarnett> I can try and run your script - but my focus is on capturing what we have right now.
32 <jmacgill> actualy, don't
33 <jmacgill> it will confuse the svn logs for what has changed
33 <jmacgill> do it imidiadly after making the release instead
33 <jmacgill> or soon after anyway (smile)
34 <jmacgill> so 0.7 is going gold today?
39 <jgarnett> Darn well hope so -
39 <jgarnett> just catching up on some Martin Davis/ Andrea email reading.
40 <jgarnett> Looking forward to some of there speed improvements.
40 <jmacgill> what list is that debate going on?
41 <jgarnett> Email (smile) We have not quite gotten JTS to have an open development process.
41 <jmacgill> ah, never mind.
42 <jgarnett> One question james
42 <jmacgill> shoot
42 <jgarnett> you made lots of improvements to the 2.0 release process, have these been pushed forward into 2.1?
42 <jmacgill> some have, but not all
43 <jmacgill> xjavadoc should be there, and I have cleaned up the build process
43 <jgarnett> Other than that UDIG is going to move over to the new CRS implementation this week, when those bugs are settled we can start winding down 2.1 development for a final release.
43 <jgarnett> Say Feb/March ?
44 <jmacgill> I did not manage to track down why the %optional% list tag was not being populated so on the 2.0.0 branch it is hard coded into the readme
44 <jmacgill> you may have to do the same
44 <jmacgill> I want to do a public API diff between 2.0 and 2.1 v.soon I want 2.0 code to run against 2.1 even if it has deprication warnings
45 <jgarnett> Hey one more idea for the meeting - I am going to try and grab Sam.
45 <jgarnett> He liked you icon suggestions.
45 <jmacgill> cool
45 <jgarnett> Nope he is out to lunch right now.
46 <jgarnett> Did you have any comments with respect to colors? wannt keep the "Geo""Tools" color scheme? As seen on http://geotools.codehaus.org/
46 <jgarnett> or go with out - as sceen on http://docs.codehaus.org/display/GEOTOOLS/Home
46 <jmacgill> yes if possible, again for continuity of brand
47 <jgarnett> Did we have that origionally?
47 <jgarnett> I suppose we did - http://modules.geotools.org/projects-overview.html
48 <jgarnett> Okay I was late to this party so richard will ahve to post the logs.
48 <jgarnett> Nice to talk with you, you have given us something to think about.
49 <jmacgill> v.old logo at:
49 <jmacgill> http://www.ccg.leeds.ac.uk/geotools/
50 <jmacgill> I think that was the last one with a compass big enough to know what it actualy was
50 <jmacgill> looks like I italicised the Geo but had it as one color
50 <jgarnett> Hense the north arrow being a good idea.
50 <jgarnett> Okay so green is good.
51 <jgarnett> I thought we may have latched onto it to match codehaus wiki.
53 * Perig (~chatzilla@ca-rennes-6-62.w80-8.abo.wanadoo.fr) has joined #geotools

  • No labels