GeoTools : 2 December 13 2004

polio heya
cholmes woo hoo! I made it on, downloading mozilla to an internet cafe computer
cholmes also note that when I use a : I mean a question mark, since my computer key board is screwy
jgarnett (smile)
jgarnett Hi Chris - did you get my plea for metadata help?
cholmes Yeah, unfortunately I am kinda fucked at the moment.
cholmes My ac power adapter has thrown a huge wrench in my plans.
cholmes This week I am going to try to put out a release using an internet cafe computer.
cholmes A release which should have easily been out last wednesday
cholmes Have not been able to do any work on it since.
cholmes But if this weeks goes decently I may be able to help out
cholmes was hoping to start working on raster support, but I can maybe put it off a bit.
cholmes I skimmed the email, but I wasnt too sure exactly what you wanted me to do.
jgarnett Raster support would be good 2
cholmes And for the offer I was thinking more design help and review, were you wanting me to write code:
jgarnett Here is my hit list for Jan 11th
jgarnett Raster support makes the list as well
cholmes What is needed for raster support anyways:
cholmes I am hitting it first because it is currently winning in my meager geoserver next feature poll
jgarnett Ah - if not I can get david to send you explicit metadata classes that need help.
cholmes If I could trade raster for metadata I def. would, as GeoServer has no real desire for better metadata
cholmes That could work.
jgarnett Sad. Metadata is really screwing my project over :-P Spent more time on it then rendering and have very little to show for it.
jgarnett But yeah I can make that trade
jgarnett I will really need that metadata api to be reviewed before I come back to geotools with it.
jgarnett I placed our initial learning on metadata and construction issues in the recent batch of geoapi updates.
cholmes Ok, I can definitely review the metadata api
cholmes Ok, I will look into the geoapi email list
jgarnett Lets wait till march - I won't have any volunteer time until then.
cholmes I have a buttload of email at the moment
jgarnett The geoapi proposal goes out today so it is already too late.
cholmes Did adam make any progress on installer:
cholmes I am answer his email right now
jgarnett I think so, thanks for the pointers.
cholmes you are saying march for rolling udig stuff into geotools:
jgarnett Lets say march for moving geotools 2.1 into stable mode as well.
jgarnett Sounds like geotools 2.2 may take off before then.
cholmes Sounds good
jgarnett As long as geotools is not doing metadata we could go stable in January once the CRS api is trusted? What do people think?
jgarnett (Oh and I do have a couple of real agenda items as well)
jgarnett I assume you want to go back to stable development chris (smile)
jgarnett We can talk about it when CRS is back in the landy of stability.
cholmes I love stable development
jgarnett 1) Rueben PMC recomendation
jgarnett 2) Some guy martin wants svn access for (let me check email)
cholmes Dope. CRS will probably be my second 1.3 thing
jgarnett Hi Ian.
IanS hi
rschulz hi, are last weeks irc logs available?
jgarnett not sure ...
rschulz also, did you get the CRSServices proj issue worked out?
rschulz (Jody)
jgarnett which one? I think we did - our wkt was not complete on the shapefile.prg
jgarnett bleck - I mean shapefile.prj
rschulz you had a test using crsservice to re-project geometries. It gave wierd results for the bounding box
rschulz I thought crsservice might be modifying its input geometries
jgarnett ah no I never figured that out - I think I got the "ask after the 18th" when we switch over to the new implementation answer.
rschulz if you have time, run the test I send with one of my replies
jgarnett Yeah I won't have time until after the 18th (sad)
jgarnett So I don't see that many PMC members here
jgarnett I can drag justin as a module maintainer in here for agenda item 2) svn access for someone Martin wants to work with
jgarnett Here is the email:
rschulz what is item 1)?
jgarnett Apparently Proshun did a lot of improvements to the j2d renderer and would like to roles the changes in, he has martin's permission and just needs svn access.
cholmes Oh that dude has done some code before, no:
jgarnett We need more PMC members for 1) - basically recomending someone who showed late to the meeting for PMC
jgarnett (smile)
jgarnett I think so, if not I offered to review his modifications, he has read the developers guide at least.
rschulz item 3) wiki sample code/tutorial organization
jgarnett In anycase it gets a +1 from me, the more help martin gets with j2d the better ...
polio sure, sounds ok here +1
cholmes +1 from me
IanS what the hey +1
jgarnett Okay sending paul the email.
jgarnett How many pmc do we have?
jgarnett I think I see 3, how many do we have in total?
cholmes We should talk about taking people off the PMC
jgarnett Well andrea is taking a break
cholmes LIke I think Rob Hranac is still on, and he should not be, as I dont think hes done geotools in almost 2 years
jgarnett I see 8 project admins on the SF site:
jgarnett Is that the best list we have?
cholmes I think so.
cholmes And where is James at the moment:
cholmes has anyone bugged him on ym:
cholmes (note that for me : == question mark, as my keyboard is not right)
jgarnett Not sure - says we need 75% PMC to do anything.
rschulz there is some info in the maven xml file about PMC status
cholmes We can go to the list if we need to do anything.
jgarnett Just IM james
IanS i was wondering about that (smile)
jgarnett Okay I would like to formally nominate Rueben for PMC ... he has done a lots of great work, user focused work - exactly what being a PMC is already about.
cholmes +1 from me
jgarnett Hats off to your rueben geotools is a lot better for your dedication.
jgarnett +1 from me.
rschulz yes, but I need to take a break and focus on classes in jan!
IanS ill toss my 1 in as well
polio +1
rschulz thanks for the complements
jmacgill hi, sorry tied up with admin stuff at this end
jgarnett You can review the
jmacgill this is a no brainer for me +1
cholmes Thats ok, it can be honarary until you decide that it officially activates (smile)
jgarnett Actually he is already meeting the obligations (smile)
jgarnett Well we are at 50% of the PMC, we need 75%
jgarnett James is the link to our best list of the current PMC?
rschulz don't I have to be a module maintainer to be on pmc?
jmacgill hang on, best list is effected by DNS problems... working on it.
jgarnett It says should, and you are pretty much the wiki maintainer with the amount of work you put in
rschulz james, any progress on dns issues?
jgarnett I count 7 on that list, we have 4 votes
jmacgill actualy, cameron is no longer PMC - he steped down
jgarnett So we are down to 6 ...
jgarnett We just need to bootstrap the process, get rueben to vote himself in...
jmacgill (smile)
jmacgill I was going to call an 'election' in January to refresh the PMC
jgarnett Or get someone to abuse their "Exalted Leader" status and vote +2.
rschulz + 1 (after I finish my masters work)
cholmes Did you count Rob on that list of 7:
jgarnett I think I did...
cholmes He should not be
cholmes (Rob H)
jgarnett hell I am going to edit the Roles and Responsibilities page - who do we have as PMC these days.
jmacgill Well, adrea is on sabatical
cholmes jody, ians, james, me, iant
cholmes seang was, but he stepped down
cholmes artur is no longer, cameron is no longer, robh is no longer
jmacgill martin
jmacgill The most important file to get right is the top level project.xml file
jmacgill that is the 'master' reference and it is very wrong at the moment
jgarnett aaime is on leave as it were?
jmacgill so, currently active, I make it - chris, james, iant, ians, martin, jody
jmacgill has IanT voted?
cholmes So if IanT does a +1 for rueben as pmc then we have 5 out of 6
jgarnett has an updated list
jgarnett I need a few people confluence handle though.
jmacgill excelent
jgarnett chris, rueben
cholmes should be cholmes
jgarnett fixing
cholmes cholmesny is only when I cant get cholmes
cholmes and now should really be cholmesbr
rschulz remember that I will not have much time over the next 3 months
rschulz jody, I am rschulz on confluence
jmacgill How long are you staying in Brazil? when are you off to africa
ianT +1
jgarnett Notice you get your own category, page is now fixed
jgarnett And now I have to kill the category - congrats rueben
rschulz damn (I mean yah...)
rschulz I probably will not be able to make many irc's next term - class on monday
jgarnett Fair enough - Now someone had another agenda item ....
rschulz item ?) wiki organization
jgarnett Nifty - you have the floor. I will point out that I have broken out a separate confluence space for udig user docs. If we need more space we can ask for it.
rschulz ok, a new user (Xavier) has been doing a lot on the wiki
rschulz he started a new page called Snippets
rschulz right now we have code spread over: tutorials, how-to's and snippets
rschulz I was thinking that short how-to's could become snippets
jmacgill I read your proposal in the email, seems sound.
rschulz and longer how-to's should be graduated to tutorials
rschulz thanks James, any other comments
rschulz I propose removeing the how-to's so we only have two places for code examples
jmacgill ok
rschulz if no one objects, I will start on this during the week.
jgarnett Some where there is a clear distinction
rschulz some of the developer how-to's will be moved to the devel guide
jgarnett has a bit - but I thought we had something better
jmacgill Snippet - sample of code for specfic task, does not require fully worked example
jgarnett We had been placing these in the FAQ
rschulz I will also work on "Wiki Documentation" to make this clearer
jmacgill Tutorial - discusion + sample code - code should compile and work stand alone
jgarnett Cool
jgarnett Um James - if I can get admin access to geotools I would love to do the navigation bar along the top
jgarnett the same way the udig and geoserver sites work.
jmacgill consider it done...
jmacgill well, consider it almost done
rschulz so, a well written snippet can form the basis for future tutorials
jgarnett CodeHaus is also moving up to version 1.3 which provides for a navigation pane on the left.
jgarnett Translation I would like to make the geotools wiki useable with out the script.
jmacgill try now
jgarnett Got it
jgarnett thanks - not sure when I will have time. I will send and email or create a news post when it is done.
jgarnett One idea that occured to me - purly selfish. Is that it would be nice to grab the reference section out into a separate wiki.
jgarnett Right now I am maintaining two reference sections (one in udig and one in geotools).
jgarnett I have mostly given up and moved pages to geotools as it is more complete.
jgarnett How would people feel about a shared reference wiki between geoapi, geotools and udig?
rschulz good idea jody, I fear to tread in the references
rschulz (they are big)
jgarnett There is sure lots of info, problem is it is very nice to link directly to stuff when working on geotools docs.
jgarnett You can beat typing OGC WFS in the middle of a sentence.
jgarnett I would worry that Geo:OGC would slow people down and thus not be worth the effort.
jgarnett So their are pros and cons.
rschulz oh, forgot about that
rschulz what syntax is needed to refer to a page in another space?
jgarnett The one I just gave Home
jgarnett or text
jmacgill so it would just be OGC ?
rschulz OGC is not too bad
jgarnett Yeag but it seems the for "OGC spec" the text would appear as "georef:OGC spec"
jgarnett Maybe 1.3 will have the problem licked.
jgarnett One nice thing is you can move pages between projects. I was able to move the udig online help to a new space in a couple of hours.
rschulz or you can include a name in the link
jgarnett We should as for the space "G" (smile)
jgarnett Okay I can make the request next time I IRC to codehaus. We will see what they can give us.
jgarnett Did anyone have the logs from last week? We missed posting them to the website.
jmacgill fraid not, I'm still not IRCing directly
jmacgill I have a quick extra topic
jmacgill environment vairables in SLD
jmacgill The next version of SLD includes the conecpt of environment variables
jgarnett So we allow access to
jmacgill no,
jmacgill it defines things like MapScaleDenominator
jmacgill i.e. named values that are not known withing the context of the SLD itself (or the Features which it referes to )
jmacgill could also include things like Day/Night Screen/Paper
jmacgill this slots into the Expression mechanism which is a bit of a pain.
jgarnett The swing text package has a series of Map subclasses for this kind of thing.
jmacgill I have two workable solutions
jmacgill if we are to have expression.getValue(Feature) to work
jmacgill we can either do:
jmacgill exp.getValue(feature, envVarMap)
jmacgill Lets the EnvVariable evaluate itself properly, but requires a rewrite of the entire fitler/expression system
jmacgill the alternative is
jmacgill resolvedExpression = expressionResolver.resolve(expression, envVarMap)
jmacgill which replaces EnvVariables with Literals
ianT I guess the second is easier to do
jmacgill I favour the latter approach as it is has fewer impacts
IanS seems a little cleaner to adjust the api, imho
jmacgill it is cleaner yes.. thats why I brought this up for discussion. I'm a bit torn
IanS when would the resolver be called?
jmacgill it would be called in the rendering pipeline (only time when map scale is actualy known)
jmacgill it is at the point when SLD is being turned into GraphicsObjects
IanS so it would be called on all Expressions?
jmacgill well, I have a - hasEnvVariable visitor impl
jmacgill that scans for them
jmacgill the solution would actualy probably involve your parser
jmacgill as a test I have the following working.
jmacgill String in = exp.toString();
jmacgill in.replace("sld:MapScaleDenominator", scale);
jmacgill Expression out = parser.parse(foo);
jmacgill where foo = output of the replace call
IanS i see
jmacgill actualy seems to work, and is only called if there actualy IS an environment variable.
IanS the 'rewrite' of the filter/expression system would mostly involve method signatures though, right?
jmacgill yes
jmacgill remember that this is one call per map rendering, not one call per feature
IanS does it actually modifiy the incoming sld 'dom'?
jmacgill no, it can't
jmacgill as the map scale is not known at that point
IanS sorry, i mean our in-memory object model
jmacgill as the same style object could be being used by more than one renderer at once (different zoom windows) then I can't modify the style object
jgarnett Hey guys I have to go - I will post the IRC logs in a bit. good meeting.
jmacgill I have to produce a modified clone
IanS that would be the only problem i can see, but you could just clone the whole thing
jmacgill that what I'm doing effectivly
jmacgill the to string, replace, to expression is just a clone
IanS but where is the resulting expression , out, placed?
rschulz I need to get back to work, good bye
jmacgill I'm doing this inside the rendering pipeline
jmacgill as the sld is converted to the internalStyle object for use by LiteRenderer2 or J2DRenderer
jmacgill be right back, just got to drop a form off
IanS gotcha
jmacgill back
jmacgill the more I think about this, the clone appoach has the advantage of beeing a once per renderin thingas opposed to a once per feature thing
IanS the computed styles are cached, right? What if the environmental variables change?
jmacgill they are cached on a per scale basis (as even now SLD is scale dependnent)
IanS i suppose that would be another argument against literal replacement vs always evaluating..
jmacgill what I'm proposing is a literal replacement at the start of each rendering
IanS okay, sorry again
jmacgill its ok, I need to thrash this out against someone
IanS i tried my best to find holes (wink)
jmacgill because it dosn't feel right to me (the replacement stratergy) but I can't find a god argument against it
jmacgill thanks for trying (smile)
IanS a god argument? (Thou shalt not replace)
jmacgill btw, how hard is it to knock out a new version of that parser when the spec changes?
jmacgill (wink)
ianT cu all
IanS the natural language parser you mean?
jmacgill yeah
jmacgill we actualy use it a lot in house (in geovistastudio) its very handy
IanS it depends really, I just made the syntax up
IanS thanks, Im glad to hear someone is using it
jmacgill are the syntax files in cvs?
jmacgill svn even
IanS looking...
IanS yes, its the expr.jjt file in filter/parser
jmacgill so if I was to add support for env variables thats what I muck around with
IanS this is a jjtree grammar which is used to generate all the other classes
jmacgill what cc did you use?
IanS yeah, youd have to figure out some way to slot that in the grammar
IanS $VARIABLE would probably work and be nice and shelly
IanS javacc
jmacgill just what I was thinking (smile)
IanS the real work is done in the ExpressionBuilder class, however
jmacgill ok. I also need to fix teh toStrings as the brackets are wrong
jmacgill so, I'd ahve to add $varname to the expr.jjt file and then modify ExpressionBuilder, and work out how to run javacc (something I have done once a long time ago)
IanS yeah, the most complicated part would be $varname I think
jmacgill humm, maven has a jjtree plugin. I wonder if I can replace the files and build the generation into the build process.
IanS i thought about it, but right now javacc generates sources which won't compile under 1.5 (enum keyword useage)
jmacgill true, but we are forcing target=1.4 and source=1.4 at the moment
IanS if it works, i won't complain
jmacgill I'll give it a shot
jmacgill life still busy at your end?
IanS yeah, need to release (whatever that entails) our stuff in januarary
IanS finished my shower retiling job which ate the past 3-4 weekends and some other days
IanS its getting better
IanS for the $variable stuff, look at the PrimaryExpression non terminal rule
jmacgill Rebecca = approx 0 free time
jmacgill PrimaryExpression, ok, will do
IanS yeah i can imagine, offspring creation is also on my agenda
jmacgill wow
jmacgill I have to run
jmacgill thanks for your help
IanS all right, mail if you need help
IanS bye
jgarnett Hi James still around?
jgarnett I was going to ask about DNS and/or Geotools 2.0
jgarnett I live in hope...