[Marble-devel] Marble Café - today's log file

Tue Sep 8 19:09:56 UTC 2015

Hi everyone,

attached you'll find today's log file of our first Marble Café session.
As mentioned before we want to establish this as a weekly opportunity for people to stay informed about the latest status of Marble. And we'd like to make it easier for people to get involved.

The next Marble Café session will be on 

Tuesday, September 15th, 19:00-21:00 CEDT, #marble on Freenode

We are looking forward to meet you there! 

Best Regards,

Torsten

[19:01:56] <utku> Hi everyone
[19:02:02] <Earthwings> hi utku
[19:02:03] <tackat> Hi, welcome to the Marble Cafe! :-) Keep your coffee mug around
[19:02:04] <Earthwings> hi tackat
[19:02:17] <tackat> hi utku
[19:02:53] <tackat> Is anybody here specifically to join the Marble Cafe meeting?
[19:02:53] --> apol (hermes at conference/kde-sprint/x-rooszaqstnqojuou) hat #marble betreten
[19:03:31] <utku> Me
[19:03:39] <tackat> yay :)
[19:04:21] <tackat> so at the moment we are working on getting vector tiling ready for production - also for our Android port
[19:04:33] <Earthwings> here's some visual impression of our progress on vector tile rendering during the last days in Randa: http://nienhueser.de/marble/marble-android-vector-tiling-randa-1.mp4
[19:05:17] <tackat> So dennis created 500 Megs of vector tiles using some bigger osm dump 
[19:05:44] <tackat> and then resurrected the work of Ander Piojan  
[19:05:55] <Earthwings> my server is currently busy creating vector tiles as well
[19:05:58] <Earthwings> http://vtiles.nienhueser.de/18/
[19:06:18] <Earthwings> we plan to move that to a kde server soon
[19:06:36] <tackat> does this work with current marble master already?
[19:06:49] <utku> What's the reason for low FPS in the video by the way?
[19:07:18] <-- apol (hermes at conference/kde-sprint/x-rooszaqstnqojuou) hat das Netzwerk verlassen (Ping timeout: 244 seconds)
[19:07:29] <Earthwings> it's faster when i don't record a screencast at the same time
[19:07:44] <Earthwings> it's a darn old Nexus 4
[19:08:05] <utku> :)
[19:08:17] <Earthwings> tackat: you just need the vectorosm dgml theme, no further code changes
[19:09:40] <EgorMatirov> hi all :)
[19:09:46] <EgorMatirov> video looks really cool
[19:10:21] <tackat> hi EgorMatirov
[19:10:47] <Earthwings> hi EgorMatirov
[19:10:54] <-- gscscnd (bc89641d at gateway/web/freenode/ip.188.137.100.29) hat das Netzwerk verlassen (Quit: Page closed)
[19:11:20] <Earthwings> EgorMatirov: we're now looking into vector rendering improvements
[19:11:30] <tackat> Earthwings: did you submit the vectorosm dgml to master yet?
[19:11:35] <Earthwings> extracting and rendering more features, larger icons etc
[19:11:47] <tackat> I'm working on bigger icons at the moment
[19:11:49] <Earthwings> tackat: no, shall i? we didn't agree on the server part yet
[19:12:18] <tackat> Earthwings: what do you mean by "we didn't agree on it yet?" :)
[19:12:43] <Earthwings> tackat: currently my server has data for Bern
[19:13:02] <tackat> I think we can just work with your experimental server of Bern in the Dgml file and just not install it by default
[19:13:02] <Earthwings> tackat: and we could use a kde server from a start in case we get access to one soon
[19:13:17] <tackat> ... and later change to the KDE server (tomorrow or so)
[19:13:28] <tackat> .. and add a cmake option :)
[19:13:33] <Earthwings> my server is not experimental, it's proven debian stable :-P
[19:14:01] <tackat> so is anybody looking for a task or a junior job? :)
[19:14:11] -*- tackat looks around
[19:14:29] <tackat> ok, just wondering
[19:14:40] <tackat> the other nice thing we are looking is usage of Phabricator
[19:14:52] <Earthwings> we're also looking into phabricator at the moment, https://phabricator.kde.org/project/profile/38/
[19:15:27] <tackat> so that will hopefully help us to better organize tasks, bugs and the workflow
[19:15:32] <Earthwings> thanks to sgclark for helping with it :)
[19:16:19] <EgorMatirov> I'm busy with school now, so sorry, can't make much help :(
[19:16:59] <Earthwings> no worries
[19:18:14] <Earthwings> so the nice thing with vector tile rendering is that we have a basic tool chain up and running already now
[19:18:17] <-- notmart (diau at kde/mart) hat das Netzwerk verlassen (Ping timeout: 268 seconds)
[19:19:14] <Earthwings> still lots of things to do, but we can generate tiles from any osm xml/pbf file and have them automatically load in marble, both the desktop version and marble maps (android)
[19:20:01] <EgorMatirov> Earthwings: do you have an apk which you used for that video? I want to try it :) (as far I understand it's not just master version)
[19:20:37] <Earthwings> EgorMatirov: it wouldn't help right now, because it references the internal server here with a local IP
[19:20:44] <utku> Yeah, I'd like to give it a spin as well
[19:20:54] <Earthwings> but i can work on setting it up for my server (nienhueser.de) later
[19:21:11] <Earthwings> one of the problems here in Randa is that the Internet sucks :-(
[19:21:56] <utku> Don't you guys meet at a university or somewhere like that?
[19:22:04] <EgorMatirov> oh, ok, understand, no problem :)
[19:23:04] <Earthwings> utku: no, in a nice house (think hostel) in a very small village in the middle of nowhere
[19:23:25] <rku> as long as internet is available… ;)
[19:23:29] <Earthwings> utku: http://www.hausranda.ch/
[19:23:57] --> KyriosBrastianos (~quassel at ppp046177147232.access.hol.gr) hat #marble betreten
[19:24:00] <Earthwings> rku: well... like... ava
[19:24:13] -*- Earthwings mumbles
[19:24:17] <rku> oh, didnt read the "internet sucks" thing yet :(
[19:25:12] <Earthwings> at the moment it's ok
[19:26:21] <tackat> hi
[19:26:29] -*- tackat just came back from lunch at Randa
[19:26:51] <tackat> .. and Earthwings just went to get some food :)
[19:27:39] <tackat> yes, internet has become better during the last few hours
[19:27:45] <tackat> rku: how is life?
[19:27:46] <tackat> :)
[19:28:06] --> shentey|m (~yaaic at x5ce0ff48.dyn.telefonica.de) hat #marble betreten
[19:28:11] <tackat> hi shentey|m
[19:28:30] <shentey|m> hi tackat
[19:28:35] <rku> fine, you guys have fun at randa?
[19:28:46] <tackat> yes, lots
[19:29:03] <tackat> fewcha is also there
[19:29:10] <tackat> so it's a little Marble meeting
[19:29:34] <tackat> and so far we have focussed on the vector tile support
[19:30:27] <tackat> and we can also build on the work of our awesome GSoC 2015 students
[19:31:18] <tackat> The nice thing about the vector tile stuff is that at this point it should be rather easy to distribute work and define lots of small tasks (also for Google Code-In)
[19:33:09] --> shentey (~shentey at x5ce0ff48.dyn.telefonica.de) hat #marble betreten
[19:33:45] <tackat> In other news: I just updated our routing icons last week: http://devel-home.kde.org/~tackat/preview-0.png
[19:34:06] <abhgangwar> hi *
[19:34:09] <tackat> hopefully that will provide a better look to this feature
[19:34:12] <tackat> hi abhgangwar
[19:34:22] <EgorMatirov> hi abhgangwar
[19:35:03] --> cmex81 (bca24019 at gateway/web/freenode/ip.188.162.64.25) hat #marble betreten
[19:35:07] <tackat> hi cmex81
[19:35:47] -*- tackat takes another cup of coffee
[19:36:14] <abhgangwar> so, I have been away from marble development for some long time :/ . Looking at the Marble updates.
[19:36:21] <utku> By the way I really want to be a mentor during GCI this year if I can find the time
[19:36:33] <tackat> utku: that would be awesome :)
[19:36:45] <tackat> utku: any preference for topics?
[19:37:27] <tackat> utku: we had the idea to focus on further work of the Android port and on OSM Vector Tile support - but of course any other ideas are welcome
[19:38:08] <utku> Well, the last time I wrote C++ was during GSoC 2013, so it'd be a good thing for both mentee and me if I stick to the areas that I know
[19:38:29] <tackat> utku: heh
[19:38:47] <tackat> utku: I'm pretty sure you won't have problems to follow the code :) 
[19:39:19] <-- cmex81 (bca24019 at gateway/web/freenode/ip.188.162.64.25) hat das Netzwerk verlassen (Ping timeout: 246 seconds)
[19:39:25] <tackat> we'd like to create lots of smallish tasks again
[19:39:38] <tackat> hopefully fabricator will help to organize that better :)
[19:39:48] <utku> I can ask both Earthwings and you for help in case that happens :)
[19:39:52] <tackat> sure
[19:39:53] <tackat> :)
[19:41:10] <tackat> but most of the stuff we have in mind should be rather easy to do (at least compared to creating a fully-fledged Map Creation wizard or a Syncing feature ;)
[19:41:22] <-- shentey|m (~yaaic at x5ce0ff48.dyn.telefonica.de) hat das Netzwerk verlassen (Ping timeout: 244 seconds)
[19:41:51] <tackat> are dates set for GCI already? I guess it will start in November ...
[19:42:46] <utku> I can already think of some basic tasks about map creation wizard. As far as I remember its labels have hardcoded font settings
[19:43:07] <tackat> utku: of course that would be cool as well :)
[19:43:23] <tackat> utku: there's still lots of potential for the wizard
[19:44:26] <tackat> abhgangwar: are you on master? :)
[19:44:37] <tackat> abhgangwar: lots of stuff has changed recently
[19:48:33] <tackat> Earthwings thinks that streets in Marble should be slimmer 
[19:49:39] <utku> Also, I have a Macbook Air on my disposal so I think I can help with Mac installer and platform specific issues as well
[19:50:01] <tackat> utku: that would be cool
[19:50:02] <utku> Although I can't promise anything yet since I didn't test if I can juggle both work and school at the same time :)
[19:50:14] <tackat> we had a chat with sgclark yesterday
[19:50:29] <tackat> about using jenkins on Windows and Mac
[19:51:22] <tackat> and we also want to use it this way to create packages for end users
[19:51:48] <tackat> utku: so ideally you could help with ensuring that those packages work for real :)
[19:52:24] <tackat> Currently we are still at Marble 1.5 for Mac, which is a shame
[19:53:45] <tackat> utku: any further thoughts on the Mac port?
[19:54:46] <utku> I have no idea at the moment, I tried compiling Marble once but I was too busy with work so I couldn't finish it
[19:54:54] <-- Nessa05 (~rajdeep at 117.211.86.108) hat #marble verlassen ("Konversation terminated!")
[19:55:02] <utku> Let's see if I can pull that one off first :P
[19:55:06] <tackat> :)
[19:55:15] <tackat> we use Qt5 now btw
[19:55:20] <tackat> primarily
[19:55:33] <tackat> we still support Qt4,  but it's not the primary citizen anymore
[19:57:51] -*- tackat points Earthwings to https://developers.google.com/kml/articles/draworder regarding the street rendering
[19:59:49] <tackat> btw, if you haven't seen latest Marble master check out our beautiful streets: http://1.bp.blogspot.com/-LqSks0bylWw/Vdar6PvgbfI/AAAAAAAACx4/GC7USpvnzWU/s1600/marble32.png
[20:00:17] <tackat> http://4.bp.blogspot.com/-WoNiwtUptGg/Vdar6NQKKMI/AAAAAAAACx8/A8YZD4c25kE/s1600/marble31.png
[20:00:51] <utku> Is there any standard about colours which we follow?
[20:01:13] <tackat> well in terms of implementation we still follow KML semantics
[20:01:24] <tackat> and in terms of color usage
[20:01:37] <tackat> we try to make it look exactly like the map on openstreetmap.org
[20:01:44] <tackat> that's the goal
[20:01:53] <tackat> still lots of work needed to tweak the details
[20:02:30] <tackat> here is a nice video regarding street rendering: https://www.youtube.com/watch?v=7uekJ9pdoCg
[20:11:56] <-- KyriosBrastianos (~quassel at ppp046177147232.access.hol.gr) hat das Netzwerk verlassen (Read error: No route to host)
[20:12:07] --> notmart (diau at kde/mart) hat #marble betreten
[20:12:44] <tackat> rku: btw, ana made some great progress regarding LatLonQuad rendering during ESA SoCiS: https://dl.dropboxusercontent.com/u/85300/Screenshots/Screen%20Shot%202015-08-27%20at%2019.08.40.png
[20:13:07] <tackat> (comparison for Marble vs. Google Earth)
[20:13:20] <tackat> I guess we are the only OSS project which renders this correctly
[20:14:39] --> KyriosBrastianos (~quassel at ppp046177147232.access.hol.gr) hat #marble betreten
[20:14:52] <rku> nice!
[20:15:30] <tackat> she had also been working on the frame around the texture so that it matches -but I don't have a screenshot readily available for that one
[20:15:39] <tackat> rku: you can even edit the Quad
[20:16:10] <tackat> rku: oh btw - we have a lunar eclipse ahead
[20:16:19] <tackat> ;)
[20:16:32] <tackat> hm, I think pgabor had a patch for that one
[20:19:03] <-> paultag_ heißt jetzt paultag
[20:19:39] <shentey> tackat: what are the ideas for vector rendering you mentioned?
[20:19:47] <utku> How was GSoC 2015?
[20:19:54] <tackat> shentey: did you follow recent development? :)
[20:20:03] <shentey> yeah I did
[20:20:22] <tackat> ok, so right now we want to get the actual osm rendering more accurate
[20:20:30] <tackat> so lots of small stuff
[20:20:31] <shentey> plain .osm files are still loaded afaik
[20:20:59] <tackat> shentey: yes, we create plain osm files for the vector tile server atm
[20:21:17] <tackat> the next aim will be to use pbf instead
[20:21:23] <tackat> that will reduce loading overhead
[20:21:42] <shentey> what vector tile server?
[20:21:43] <tackat> and right now we create tiles so that features get duplicated when the reach across tiles
[20:22:04] <tackat> shentey: dennis set up an experimental one today
[20:22:19] <shentey> ic
[20:22:25] <tackat> see the earlier discussion
[20:22:27] <tackat> today
[20:22:41] <shentey> is there a backlog?
[20:23:14] <tackat> shentey: 
[20:23:16] <tackat> [19:04:21] <tackat> so at the moment we are working on getting vector tiling ready for production - also for our Android port
[20:23:16] <tackat> [19:04:33] <Earthwings> here's some visual impression of our progress on vector tile rendering during the last days in Randa: http://nienhueser.de/marble/marble-android-vector-tiling-randa-1.mp4
[20:23:16] <tackat> [19:05:17] <tackat> So dennis created 500 Megs of vector tiles using some bigger osm dump 
[20:23:16] <tackat> [19:05:44] <tackat> and then resurrected the work of Ander Piojan  
[20:23:16] <tackat> [19:05:55] <Earthwings> my server is currently busy creating vector tiles as well
[20:23:16] <tackat> [19:05:58] <Earthwings> http://vtiles.nienhueser.de/18/
[20:23:16] <tackat> [19:06:18] <Earthwings> we plan to move that to a kde server soon
[20:23:16] <tackat> [19:06:36] <tackat> does this work with current marble master already?
[20:23:46] <tackat> shentey: I'll send the full log around after our Marble Cafe session is done (in about 40 mins)
[20:23:52] --> dh64 (~dh64 at 65-102-179-54.tukw.qwest.net) hat #marble betreten
[20:24:04] <shentey> great :)
[20:24:19] <tackat> shentey: so we want to move that data over to a KDE server once I reach bcooksley
[20:24:27] <shentey> so vector tile rendering won't be an offline feature for the time being?
[20:24:56] <tackat> shentey: we want to get it online this week
[20:25:25] <tackat> shentey: and then we also want to improve current tile cutting  so that we have no feature duplication - but that will take a bit longer I'm afraid
[20:26:13] <tackat> shentey: we really want to get that soon into our Android Marble Maps version
[20:28:12] <shentey> tackat: how do you cut the tiles?
[20:28:34] <tackat> shentey: right now we use osmconvert
[20:28:50] <tackat> shentey: but we probably need to develop our own tile cutter later on
[20:29:35] <shentey> tackat: have you looked into existing tile cutters like osmium?
[20:29:36] <tackat> at the moment we just focus on getting the whole cycle established - from cutting over storing, downloading and rendering
[20:30:06] <tackat> shentey: Dennis says that osmconvert is faster 
[20:30:36] <tackat> (and Earthwings just mumbled that osmium doesn't support some  feature)
[20:31:19] <tackat> shentey: any thoughts on your side about it? :)
[20:32:51] <shentey> hmm, I wonder how maps.me create their vector data, which is both very detailed and small
[20:33:10] <shentey> and their vector format is even searchable
[20:33:52] <shentey> and some blog post suggests that they'll open-source (some of?) their code
[20:34:20] -*- tackat looks up maps.me
[20:36:05] <Earthwings> tackat: happy waiting for the interwebs
[20:36:16] <tackat> here it works quite well
[20:36:18] <tackat> odd :)
[20:36:19] <Earthwings> tackat: your request is queued behind mine
[20:36:26] <tackat> ?
[20:36:27] <Earthwings> tackat: i'll talk to mario about that
[20:36:30] <Earthwings> unormal: ^^
[20:37:54] --> rdieter (~rdieter1 at fedora/rdieter) hat #marble betreten
[20:38:02] <tackat> hi rdieter
[20:38:33] -*- rdieter waves hi
[20:41:36] <tackat> ho :)
[20:41:47] <tackat> rdieter: any marble problems in the packaging dept. ?
[20:42:02] <tackat> :)
[20:42:59] <rdieter> all good now
[20:43:06] <tackat> :)
[20:43:12] <rdieter> I guess I still owe a reviewboard submission
[20:44:06] <tackat> that would be great
[20:44:11] <tackat> indeed
[20:44:13] <tackat> :)
[20:47:51] <rdieter> sneak-peak, http://pkgs.fedoraproject.org/cgit/marble.git/tree/marble-15.08.0-qt4.patch
[20:48:00] <rdieter> fixes building in Qt4 mode for me
[20:48:13] <rdieter> (since we have build both qt4 and qt5 libmarblewidget)
[20:48:19] <tackat> yay
[20:48:20] <tackat> :)
[20:48:51] <tackat> we also have some general Qt 4 build issues as regressions after GSoC 2015
[20:49:11] <tackat> but those should be easy to sort out -will continue to look into those later 
[20:54:23] <-- rdieter (~rdieter1 at fedora/rdieter) hat das Netzwerk verlassen (Remote host closed the connection)
[21:01:10] <tackat> ok, so that ends today's Marble Café session for today. :) 