View source for Kexi/Junior Jobs/Map Browser Form Widget/Log-2011-04-08

<pre>                                        
[12:53:54] <rockford_> Hi
[12:54:08] <ObKo> Hello
[13:00:29] <tackat> ObKo: hi
[13:01:13] <tackat> Earthwings: jstaniek: rockford_: ObKo: ok, let's start :)
[13:01:25] <tackat> idis and dmarth won't be able to join
[13:01:39] <tackat> so I'll send the log of the meeting around by mail
[13:02:01] <tackat> rockford_: ObKo: Thanks for joining :)
[13:04:49] <tackat> Obko: rockford_ will look into creation of Kexi integration for Marble during GSoC 2011
[13:05:11] <tackat> ObKo: for that a vector based map would be great to have ;-)
[13:05:57] <rockford_> at lest for some parts of it
[13:06:03] <tackat> rockford_: Obko will look into  OpenStreetMap vector rendering as part of his GSoC 2011 activity
[13:06:22] <tackat> rockford_: maybe that could be of help for you at a later point :)
[13:06:52] <tackat> rockford_: have you made yourself familiar with the (Marble) code already?
[13:07:31] <rockford_> a little bit
[13:07:49] <tackat> cool
[13:08:12] <rockford_> i've already done added marble to froms
[13:08:13] <tackat> rockford_: do you need some links for introduction? And do you have a git account already?
[13:08:18] <tackat> ah ok
[13:08:35] <rockford_> i've got access to calligra
[13:08:48] <tackat> I guess you then have access to Marble as well ;)
[13:09:38] <rockford_> at last for read
[13:09:38] <tackat> rockford_: my guess is that you maybe don't get in touch too much with the Marble code itself (compared to the Calligra code).
[13:10:33] <rockford_> yes this would be good to not change marble to much
[13:10:37] <tackat> rockford_: if you need write access don't hesitate to let us know ;)
[13:10:58] <tackat> rockford_: well if it's necessary  or if it would benefit Marble then it would be great :)
[13:10:58] <rockford_> ok, thanks
[13:11:50] <tackat> rockford_: if you have any questions or any doubts don't hesitate to come here and ask. Usually there's somebody there who has an answer.
[13:11:52] <rockford_> ok i'll remember
[13:12:21] <tackat> (like Earthwings, shentey, me, or idis)
[13:13:07] <tackat> rockford_: At least if you let the user change the map in some way then you'll need some info how to do that :)
[13:13:29] <tackat> The other place where you can ask is marble-devel@kde.org
[13:13:57] <tackat> rockford_: some technical write-ups are available at : http://techbase.kde.org/Projects/Marble
[13:14:06] <rockford_> i've already have it in my addrssbook :-)
[13:14:21] <tackat> cool
[13:14:23] <tackat> :)
[13:14:30] <rockford_> i can share with you thinks i've already done
[13:14:39] <tackat> that would be great
[13:14:43] <tackat> ObKo: ping?
[13:14:49] <ObKo> pong
[13:14:57] <rockford_> please look at:  http://rockfordsone.blogspot.com/
[13:15:25] <tackat> rockford_: cool
[13:16:02] <rockford_> i've started blog on which i'll put my progress and any interesting I find
[13:16:14] <tackat> rockford_: ObKo: It would be great if you guys could write down your plans regarding Marble on techbase
[13:16:57] <tackat> rockford_: ObKo: something like a plan/schedule for the upcoming 2-4 weeks
[13:17:44] <tackat> rockford_: well I'm not your mentor but if you have changes in mind regarding Marble that would help to keep everyone in Sync
[13:17:56] <ObKo> ok
[13:18:31] <tackat> Obko: rockford_:  So something like http://techbase.kde.org/Projects/Marble/GSoC2011 would be a good place 
[13:19:34] <tackat> You don't need to write books there, just a few bullet points that cover the next upcoming steps would be great already.
[13:19:55] <bholst> tackat: yes, it is :D
[13:20:00] <bholst> hi tackat
[13:20:09] <tackat> ObKo: rockford_: and then blogging. rockford_ already mentioned it :)
[13:20:40] <tackat> ObKo: rockford_: basically it would be cool if you'd create blog entries about every 2 weeks 
[13:21:47] <tackat> ObKo: rockford_: I guess the first step would be to ensure that planetkde.org publishes your blogs
[13:21:59] <tackat> (if you haven't done already)
[13:22:25] <tackat> ObKo: rockford_: then the next step would be to introduce yourself via a blog entry to the KDE community. :)
[13:22:27] <Earthwings> back
[13:22:34] <tackat> Earthwings: wb Earthwings
[13:22:42] <tackat> bholst: and hi bholst
[13:22:49] <rockford_> Yes, i know i didn't add my blog to plante yet, i'll try to add it by the end of today
[13:23:22] <tackat> ObKo: rockford_: and maybe also cover your plans in your first GSoC 2011 related blog entry already
[13:23:57] <tackat> dmarth: hey I know you're absent but of course the same applies to you ;-)
[13:24:35] <tackat> ObKo: do you think this is a good idea?
[13:24:37] <tackat> :)
[13:24:56] <ObKo> of course :)
[13:25:00] <tackat> cool :)
[13:25:42] <tackat> ideally with screenshots if there are some available already since everybody loves screenshots ;)
[13:26:05] <rockford_> or even films ;-)
[13:26:13] <tackat> that would rock even more
[13:26:14] <tackat> :)
[13:26:19] <ObKo> in 3D?
[13:27:01] <rockford_> my work is rathre 2d so i'll stay with 2D (for now ;-)
[13:27:14] <tackat> ObKo: rockford_: btw. there's also a facebook page for Marble available at   http://www.facebook.com/marbleglobe where you can publish stuff
[13:27:45] <tackat> rockford_: I think ObKo's is currently also still 2D but I'm sure shentey will turn it into 3D at one point
[13:27:54] <tackat> shentey is working on the OpenGL view
[13:28:04] <Earthwings> we have an opendesktop group as well, http://opendesktop.org/groups/?id=439
[13:28:08] <tackat> yep
[13:28:18] <Earthwings> in contrast to facebook, opendesktop does not eat your soul or sell your data
[13:28:24] <tackat> yep ;)
[13:28:46] <tackat> hey we should ask karli for a link like opendesktop.org/marbleglobe
[13:29:05] <tackat> I already asked him for a facebook-style "like" button 
[13:29:35] <Earthwings> hehe
[13:30:25] <tackat> ObKo: rockford_: anyways communication is really important to us since it avoids that you'll do duplicated work or in a more complicated way than necessary
[13:30:50] <tackat> and of course in terms of blogging it brings attention to your and our work :)
[13:31:24] <tackat> jstaniek: are you there by chance?
[13:32:02] <tackat> doesn't look like
[13:32:22] <tackat> I hope everything we bring up here is in jstaniek's interest as well
[13:32:40] <tackat> rockford_: do you have anything that we can help you with? 
[13:33:04] <tackat> rockford_: and do you have any plans in the upcoming days that would affect the Marble code possibly? :)
[13:33:30] <rockford_> no, not in nerest future
[13:33:34] <tackat> ok
[13:34:12] <rockford_> if i'll have something, that i would like to change ill give you a notice a littel bit erlier.
[13:34:34] <tackat> sounds good :)
[13:34:51] -*- jstaniek back
[13:35:04] <tackat> ObKo: dmarth: rockford_: so in the future we'll organize a weekly IRC meeting
[13:35:19] <jstaniek> tackat: the home for all info is here http://community.kde.org/Kexi/Junior_Jobs/Map_Browser_Form_Widget
[13:35:30] <tackat> jstaniek: ah ok
[13:35:55] <Earthwings> can we agree on a day and time for the weekly meeting already?
[13:36:57] <jstaniek> tackat: please tell me if the core ideas presented there are not clear; it would be also nice if you see not-hard-to-deliver features that we missed
[13:37:20] <tackat> jstaniek: what is brought up there looks pretty good and straight forward to me
[13:37:54] <tackat> jstaniek: it mostly covers basically the interface with the MarbleWidget itself. 
[13:38:11] <tackat> (and doesn't go into the inner workings of Marble)
[13:38:35] <tackat> so I don't see any problems there. 
[13:38:39] <jstaniek> yes, its' all possible with just public API
[13:39:12] <tackat> jstaniek: if you intend to interoperate with map contents at a later point then things might become more tricky
[13:39:16] <jstaniek> tackat: it's not a problem if you flood us with ideas so we can pick two or three and prioritize them
[13:39:22] <jstaniek> yes
[13:39:23] <tackat> ;-)
[13:40:22] <jstaniek> we plan to have a gui that selectes two columns from db of type float and assigns them to map widget (longitude/latitude)
[13:40:24] <tackat> jstaniek: rockford_: I don't know whether it makes much sense for you to participate in our meeting but of course it would be great if you two would be present every once in a while ;)
[13:40:57] <tackat> for ObKo and dmarth however the meeting is basically mandatory
[13:41:10] <jstaniek> well, I'd prefer to have access to minutes and communicate via email
[13:41:23] <tackat> jstaniek: sounds ok to me
[13:41:41] <rockford_> i would be nice to get log (if i'm absen)
[13:42:04] <tackat> rockford_: jstaniek: I can send you the logs, no problem
[13:42:32] <jstaniek> cool, since I am here, one question: can I draw using scalable graphics using public marble API? open street maps come to mind
[13:43:56] <tackat> jstaniek: well
[13:44:11] <tackat> jstaniek: how do you want that graphic put onto the map?
[13:44:32] <tackat> jstaniek: should it be placed (undistorted) on the map 
[13:44:46] <tackat> jstaniek: using a geo coordinate as a reference?
[13:44:48] <tackat> Or
[13:45:35] <tackat> jstaniek: do you want your graphic to be projected in map coordinates directly onto the globe/map (so that it gets skewed in a way that it follows the ground/projection ?
[13:46:22] <rockford_> jstaniek is thinking if maps can be drown as scalable graphicsm, widget is raster, is it possible to draw map as vector?
[13:46:51] <tackat> well 
[13:47:00] <tackat> yes to some degree
[13:47:11] <tackat> we can import KML
[13:47:30] <tackat> which can display linestrings, polygons, etc.
[13:48:00] <tackat> and ObKo is also working on OpenStreetMap support
[13:48:36] <tackat> both provide a kind of vector map
[13:49:12] <tackat> rockford_: jstaniek: but I don't exactly understand what problem you want to solve - maybe there's a better answer :)
[13:50:21] <jstaniek> tackat: well, we need vectors to have quality printing
[13:50:32] <tackat> rockford_:  jstaniek: ah ok
[13:50:40] <jstaniek> so we'd like to ask 'marble engine' for given DPI
[13:50:50] <tackat> jstaniek: ok
[13:51:42] <tackat> jstaniek: rockford_: "Improved printing" would be a topic for itself
[13:51:58] <jstaniek> also, we woudl like to assert there's native resolution used without bluring, if we want this, then we'll be able to employ rescaling using our means, e.g. in ODF renderer or kexi reports
[13:52:25] <tackat> jstaniek: rockford_: There are several possible ways to get improved printing working:
[13:52:41] <tackat> jstaniek: rockford_: On one hand there are raster maps
[13:53:03] <tackat> jstaniek: rockford_: * some with labels (like raster OSM) 
[13:53:39] <tackat> jstaniek: rockford_: * some where the labels are not part of the bitmap map data  (like Satellite View, Atlas, etc)
[13:53:46] <jstaniek> yeah
[13:54:11] <tackat> jstaniek: for the latter you could just repaint map with a bigger viewport size
[13:54:18] <jstaniek> yes
[13:54:23] <tackat> (like 3000x2000)
[13:54:37] <tackat> and send the resulting page to the printer
[13:54:46] <jstaniek> btw, can I paint without need to allocate scrollarea, or even qwidget?
[13:54:51] <tackat> ideally you'd make this change available to Marble as part of the Marble API ;)
[13:55:08] <jstaniek> yeah sure
[13:55:11] <jstaniek> why not
[13:55:17] <tackat> good
[13:55:18] <tackat> :)
[13:55:48] <tackat> for some maps (like Atlas, SatelliteView) we put labels on top of the map ourselves
[13:56:14] <tackat> if you print those maps at "high dpi" it would be necessary to increase the font size
[13:56:30] <tackat> same goes for coordinate lines
[13:56:54] <tackat> so if you want to handle improved printing quality then the tasks to do there would be:
[13:57:32] <ObKo> Does placemarks have some kind of layer (z-value)?
[13:57:39] <tackat> * make it possible to render the map offscreen onto a large pixmap reflecting a dpi setting (which we could add to marble)
[13:58:34] <tackat> ObKo: I think it's currently handled automatically via the visual category / priority
[13:58:42] <tackat> ObKo: but better ask idis on that topic
[13:58:47] <jstaniek> the extra apis would render to abstract QPaintDevice (or are there such apis already?)
[13:59:46] <tackat> jstaniek: possibly. I'm not quite sure. Best would be to discuss this topic with shentey
[14:00:28] <tackat> jstaniek: shentey is working on OpenGL and to do the "offscreen" printing properly we need to take the future OpenGL case into account as well
[14:01:02] <tackat> jstaniek: but I expect that it is about as easy as that
[14:01:29] <tackat> jstaniek: and then we need to change the Marble code so that it follows the dpi settings for the lines and selfdrawn labels.
[14:01:56] <tackat> jstaniek: I think those are the most important things how raster maps can be printed at better quality
[14:02:02] <tackat> jstaniek: regarding vector maps:
[14:02:27] <tackat> jstaniek: so far we only have one "real" vector map: that's the "Plain Map" 
[14:02:53] <rockford_> :-)
[14:03:10] <tackat> jstaniek: rockford_: but that one of course is quite limited and coarse ;)
[14:03:48] <tackat> but we have two "real" vector maps on our TODO: the one is the one ObKo is working on (OSM vector rendering)
[14:03:58] <tackat> and then there is the NaturalEarth one 
[14:04:22] <tackat> see http://techbase.kde.org/Projects/Marble/NaturalEarth for more details
[14:04:22] <jstaniek> hmm, OSM vector could be a candidate for playing in our reports...
[14:04:46] <jstaniek> rockford_: let's save these notes and bookmark the links...
[14:04:49] <tackat> jstaniek: first goal with regard to OSM is to just display vector OSM files correctly
[14:05:24] <tackat> jstaniek: and then it's still a few steps to get a full blown map that gets automatically loaded like the other full worldmaps are 
[14:05:46] <tackat> jstaniek: so I think both are rather longterm
[14:06:19] <tackat> jstaniek: rockford_: but if you manage to improve printing for raster then that would automatically work well for the vector maps in the future as well
[14:06:28] <tackat> (which would then have the best quality possible)
[14:06:52] <tackat> jstaniek: rockford_: hope that helped a bit? :)
[14:06:57] <jstaniek> yeah
[14:07:02] <jstaniek> OK, nice. What does your ABI mangement work?
[14:07:23] <tackat> you mean in terms of keeping BC? 
[14:07:24] <jstaniek> I am asking because we may look at the API changes
[14:07:25] <tackat> or what?
[14:07:27] <jstaniek> yes
[14:07:29] <tackat> well
[14:07:29] <jstaniek> BC
[14:07:48] <tackat> we do d-pointers, shared d-pointers etc. for all classes
[14:07:49] <tackat> but
[14:08:09] <tackat> we don't promise BC across KDE minor releases.
[14:08:21] <tackat> we just keep BC for bugfix releases.
[14:08:58] <tackat> so if you want to change add  APIs there shouldn't be too much problems there.
[14:09:21] <tackat> for the main APIs like MarbleWidget we try to keep those API changes small though.
[14:10:25] <tackat> Does that help ? :)
[14:11:38] <jstaniek> yeah
[14:12:16] <jstaniek> we can develop facade on top of some APIs if needed
[14:12:26] <jstaniek> as a BC helper
[14:12:32] <tackat> In general we stick to the advices regarding KDE library policies - but we don't keep BC from e.g. version 1.1 to 1.2 or from 1.2 to 1.3. We so far just enforce it for bugfix releases (like 1.1.0 -> 1.1.1) - or for Marble releases which are not in sync with KDE (like Marble 1.1 was)
[14:13:17] <tackat> jstaniek: If you work in trunk you can just submit your suggested API changes to reviewboard
[14:13:33] <tackat> jstaniek: we'll scream if something isn't right ;)
[14:14:08] <tackat> jstaniek: more questions? :)
[14:14:28] <jstaniek> that's more than enough
[14:14:29] <jstaniek> thanks
[14:14:40] <jstaniek> rockford_: let's note: reviewboard.git.kde.org
[14:14:46] <tackat> yep
[14:14:58] <jstaniek> tackat: thx a lot
[14:15:03] <tackat> ObKo: btw: idis and me looked at your recent patches
[14:15:42] <tackat> ObKo: they look pretty good. You took the comments we discussed into account, ie using geodata 
[14:15:42] <tackat> parser code and use geodata classes.
[14:16:11] <tackat> ObKo: idis is missing test classes though ;) (geodata is the place we have most, so there are 
[14:16:11] <tackat> some examples to reuse)
[14:17:24] <tackat> ObKo: he also thinks that your work could end up turning the naive geometrylayer into a 
[14:17:24] <tackat> geographicsscene, but he's not sure you know this already. (or even that's it 
[14:17:24] <tackat> is the work he envisionned). Well that's a point he should discuss with you directly.
[14:18:02] <tackat> ObKo: still there? 
[14:18:08] <tackat> (sorry it got a bit long)
[14:20:17] <tackat> ObKo: So I think idis and you should discuss this in the future.
[14:21:11] <tackat> Ok, I'll close the log at this point.
[14:21:22] <ObKo> roger that
[14:21:26] <tackat> ObKo: ah
[14:21:32] <tackat> ObKo: hehe
[14:21:47] <tackat> ObKo: do you have any further issues? :)
[14:21:52] <tackat> or problems?
[14:22:06] <tackat> ObKo: did you manage to get around the style problems? 
[14:22:52] <tackat> (the one with the line width in meters)
</pre>