--- Log opened Wed May 27 00:00:19 2015
00:05 -!- KaZeR [~KaZeR@64.201.252.132] has quit [Remote host closed the connection]
00:11 -!- KaZeR [~KaZeR@172.56.38.35] has joined #navit
00:11 -!- mode/#navit [+o KaZeR] by ChanServ
00:18 -!- xenos1984 [~xenos1984@127-166-191-90.dyn.estpak.ee] has quit [Quit: Leaving.]
00:55 -!- KaZeR [~KaZeR@172.56.38.35] has quit [Remote host closed the connection]
00:55 -!- KaZeR [~KaZeR@172.56.38.35] has joined #navit
00:55 -!- mode/#navit [+o KaZeR] by ChanServ
00:57 -!- KaZeR [~KaZeR@172.56.38.35] has quit [Read error: No route to host]
01:15 #navit: < Navit> See compile results history at http://download.navit-project.org/logs/navit/stats.html
02:52 -!- KaZeR [~KaZeR@172.56.42.243] has joined #navit
02:52 -!- mode/#navit [+o KaZeR] by ChanServ
02:53 -!- KaZeR [~KaZeR@172.56.42.243] has quit [Remote host closed the connection]
02:54 -!- KaZeR [~KaZeR@172.56.42.243] has joined #navit
02:54 -!- mode/#navit [+o KaZeR] by ChanServ
02:58 -!- KaZeR [~KaZeR@172.56.42.243] has quit [Ping timeout: 244 seconds]
03:12 -!- AGTT [~root@cpe-104-229-91-72.twcny.res.rr.com] has left #navit []
04:01 -!- KaZeR [~KaZeR@c-67-161-64-186.hsd1.ca.comcast.net] has joined #navit
04:01 -!- mode/#navit [+o KaZeR] by ChanServ
04:26 -!- KaZeR_ [~KaZeR@172.56.17.86] has joined #navit
04:30 -!- KaZeR [~KaZeR@c-67-161-64-186.hsd1.ca.comcast.net] has quit [Ping timeout: 272 seconds]
04:37 -!- KaZeR_ [~KaZeR@172.56.17.86] has quit [Remote host closed the connection]
04:38 -!- KaZeR [~KaZeR@172.56.17.86] has joined #navit
04:38 -!- mode/#navit [+o KaZeR] by ChanServ
04:43 -!- KaZeR [~KaZeR@172.56.17.86] has quit [Ping timeout: 276 seconds]
05:30 -!- xenos1984 [~xenos1984@127-166-191-90.dyn.estpak.ee] has joined #navit
05:47 -!- _rd [~rd@pD9E7DC0A.dip0.t-ipconnect.de] has joined #navit
06:00 -!- _rd [~rd@pD9E7DC0A.dip0.t-ipconnect.de] has quit [Ping timeout: 244 seconds]
06:02 -!- jjelen [jjelen@nat/redhat/x-ndqgsktqcfsraddw] has joined #navit
06:33 -!- Robotaxi [3ef5dbf5@gateway/web/freenode/ip.62.245.219.245] has joined #navit
07:59 -!- j_f-f [~quassel@jff-webhosting.net] has quit [Read error: Connection reset by peer]
08:02 -!- j_f-f [~quassel@jff-webhosting.net] has joined #navit
09:00 -!- nudeltofu [~nudeltofu@nat49234.hs-niederrhein.de] has joined #navit
09:37 -!- nudeltofu [~nudeltofu@nat49234.hs-niederrhein.de] has quit [Remote host closed the connection]
10:49 -!- zoff99 [~chatzilla@unaffiliated/zoff99] has joined #navit
11:30 -!- jandegr [57417d8d@gateway/web/freenode/session] has joined #navit
11:31 #navit: < jandegr> hi zoff99
11:31 #navit: < jandegr> https://dl.dropboxusercontent.com/u/93775123/Navit/Wien3.PNG
11:31 #navit: < jandegr> https://dl.dropboxusercontent.com/u/93775123/Navit/Wien4.PNG
11:32 #navit: < jandegr> still without rounding
11:32 #navit: < jandegr> adding rounding might even further improve it
11:36 -!- jandegr [57417d8d@gateway/web/freenode/session] has quit [Changing host]
11:36 -!- jandegr [57417d8d@gateway/web/freenode/ip.87.65.125.141] has joined #navit
12:06 #navit: < zoff99> hi
12:06 #navit: < zoff99> it must be calssified as a mapping error
12:06 #navit: < zoff99> since the coords are too close for WGS 84 projection
12:07 #navit: < zoff99> u can see that its calc-ed correctly. but if you push y-coord together even with lat,lon it changes sides
12:07 #navit: < zoff99> i say this is just mapped too close. on openstreetmap.org you cant even see it properly
12:08 #navit: < zoff99> how did you calc this now to work?
12:18 -!- nudeltofu [~nudeltofu@nat49234.hs-niederrhein.de] has joined #navit
12:49 -!- nudeltofu [~nudeltofu@nat49234.hs-niederrhein.de] has quit [Remote host closed the connection]
12:55 -!- pini [~pini@bou-fi.pustule.org] has joined #navit
13:26 -!- nudeltofu [~nudeltofu@nat49234.hs-niederrhein.de] has joined #navit
13:27 -!- jandegr [57417d8d@gateway/web/freenode/ip.87.65.125.141] has quit [Quit: Page closed]
13:31 #navit: < nudeltofu> hi, i got the map "nrw.bin" in germany, its the only map loaded. i was wondering why a river near prague is shown? i also calculated a bounding box, its obv far too large. why is that?
13:35 #navit: < nudeltofu> here is a image of it: https://www.imageupload.co.uk/image/Z1LE
14:17 -!- jandegr [57417d8d@gateway/web/freenode/ip.87.65.125.141] has joined #navit
14:27 -!- jjelen [jjelen@nat/redhat/x-ndqgsktqcfsraddw] has quit [Ping timeout: 246 seconds]
15:10 -!- zoff99 [~chatzilla@unaffiliated/zoff99] has quit [Quit: ChatZilla 0.9.86-rdmsoft [XULRunner 1.9.1/20090707221522]]
15:22 -!- nudeltofu [~nudeltofu@nat49234.hs-niederrhein.de] has quit [Quit: Leaving]
15:28 -!- KaZeR [~KaZeR@64.201.252.132] has joined #navit
15:28 -!- mode/#navit [+o KaZeR] by ChanServ
16:08 -!- Robotaxi [3ef5dbf5@gateway/web/freenode/ip.62.245.219.245] has quit [Ping timeout: 246 seconds]
16:29 #navit: <@KaZeR> hi there
17:02 -!- bzed [~bzed@shell.bzed.at] has joined #navit
17:44 -!- tryagain [~quassel@178.216.76.43] has joined #navit
18:18 -!- KaZeR [~KaZeR@64.201.252.132] has quit [Read error: Connection reset by peer]
18:18 -!- KaZeR [~KaZeR@64.201.252.132] has joined #navit
18:18 -!- mode/#navit [+o KaZeR] by ChanServ
18:26 -!- _rd [~rd@pD9E7DC0A.dip0.t-ipconnect.de] has joined #navit
18:54 #navit: < tryagain> hi all
18:58 #navit: < tryagain> nudeltofu Rivers you see outside your area of interest are there because of the way planet extractor works. Object which crosses tile boundary is pulled to top level tiles. You download whole tiles which overlap with your area of interest. See also http://wiki.navit-project.org/index.php/Binfile
18:58 -!- _rd [~rd@pD9E7DC0A.dip0.t-ipconnect.de] has quit [Ping timeout: 264 seconds]
19:01 #navit: < tryagain> jandegr how did you manage to improve these intersecting roads? Do you have any clue about actual distortion size (in meters)?
19:01 #navit: < jandegr> hi tryagain
19:02 #navit: < tryagain> hi jandegr
19:02 #navit: < jandegr> are you talking about the Wien case ?
19:02 #navit: < tryagain> yes
19:04 #navit: < jandegr> the internal coordinate system is too coarse
19:04 #navit: < tryagain> internal=binfile?
19:05 #navit: < tryagain> if this distortion is less than 1 meter, it's most probably the rounding, which we do in maptool.  OSM has 1cm accuracy at Equator, while we have 1 meter accuracy there.
19:05 -!- KaZeR_ [~KaZeR@64.201.252.132] has joined #navit
19:05 -!- KaZeR [~KaZeR@64.201.252.132] has quit [Read error: Connection reset by peer]
19:05 #navit: < jandegr> I tried rounding and that is of no use
19:06 #navit: < jandegr> btw there is no rounding in maptool, it is intger floor, but rounding does not help either
19:07 #navit: < jandegr> but the earths radius could be increased with several kilometers and that helps a little
19:08 #navit: < jandegr> anyway just enough to solve Wien
19:09 #navit: < tryagain> interesting way :) I was thinking that different way, to change our coordinate representation to decimetres or even centimetres.
19:09 #navit: < tryagain> in your words, increase earth radius by factor 10 or 100
19:11 #navit: < tryagain> i'm afraid adding a few kilometers just hides this flooring/rounding problem, but with insignificantly changed radius it may show itself in a different place
19:13 #navit: < tryagain> What if we collect 5-10 meters of way segments before we calculate the angles?
19:14 #navit: < tryagain> At that distance, ways should split more significantly and show actual direction...
19:19 #navit: < jandegr> a really good solution would be both change the radius and to increase the coordinate resolution by x10
19:22 #navit: < tryagain> but that means map incompatibility
19:23 #navit: < jandegr> compatibility with what ?
19:25 #navit: < tryagain> with previous versions of navit. We'll have to force users upgrade their binaries.
19:26 #navit: < jandegr> Navit is free
19:27 #navit: < jandegr> AFAIK :)
19:27 #navit: < tryagain> as in "free to flow in direction it wants to" ;)
19:29 #navit: < tryagain> actually that should be done unless we'll find a different solution.
19:31 #navit: < tryagain> What your radius change experiment has shown? Is it a massive code change in different places to keep things working or just one (or a few) constants change?
19:33 #navit: < jandegr> 6378137 and x10 should do it
19:34 #navit: < jandegr> x10 is to improve resolution and 6378137 is to be compatible with others
19:35 #navit: < jandegr> http://tool-online.com/en/coordinate-converter.php
19:37 #navit: < jandegr> iirc that would bring us to WGS84-World Mercator
19:39 #navit: < jandegr> but if you don't mind a few wrong instructions there is no need to change anything
19:40 #navit: < tryagain> i think we should fix such navigation bugs
19:42 #navit: < jandegr> but for now I can blame the coordinate system for some wrong instructions :)
19:43 #navit: < tryagain> so you need this bug to be kept for a while now? :)
19:44 #navit: < tryagain> do you have some osm ids for your Wienna example?
19:51 #navit: < jandegr> http://www.openstreetmap.org/#map=19/48.19379/16.41378
19:52 #navit: < jandegr> from what zoff told me, Navit will connect unconnected roads as well if each has a coord close to another, but I don't have a sample
19:57 #navit: < tryagain> he's right, if two ways have a node with same coordinate values in binfile, they would be treated as connected.
19:58 #navit: < jandegr> OSM 11 millimeter resolution versus Navit 100 centimeter resolution :)
20:02 #navit: < tryagain> that node in Wienn stands in 50 cm from the way it crosses after our coordinate transformation.
20:06 #navit: < jandegr> now that there is some awareness we can watch out for more samples and then consider a change or not
20:25 -!- zoff99 [~chatzilla@unaffiliated/zoff99] has joined #navit
20:25 #navit: < zoff99> hi guys
20:25 #navit: < zoff99> i wonder how easy you talk about changing x10 or x100 the resolution :-)
20:25 #navit: < zoff99> u do know that the end result has to fit into INT range
20:26 #navit: < zoff99> also in maptool there are some function which use unused INT cood values for other stuff
20:26 #navit: < zoff99> so if you increase resultion by x100 dont you think that will overflow? i am pretty sure somebody had some thoughts why it is that way it is now
20:27 #navit: < zoff99> if htere is no overflow, or other drawback near the poles (e.g. finland) then do it
20:27 #navit: < zoff99> who care if people need to download their maps again. it takes only a few minutes
20:31 #navit: < jandegr> that is why I restricted to x10
20:31 #navit: < zoff99> have you checked what the maximum values possible are? for extreme world coordinates
20:32 #navit: < jandegr> and I suppose tryagain can fix that overnight :)
20:32 #navit: < jandegr> mercator does not go to the extremes
20:32 #navit: < zoff99> and there are special values in maptool. to find which coordinates have not been resolved yet
20:32 #navit: < zoff99> in zanavi i dont use this anymore, so good news for me :)
20:50 -!- zoff99 [~chatzilla@unaffiliated/zoff99] has quit [Quit: ChatZilla 0.9.86-rdmsoft [XULRunner 1.9.1/20090707221522]]
20:52 -!- jandegr [57417d8d@gateway/web/freenode/ip.87.65.125.141] has quit [Quit: Page closed]
21:27 #navit: < tryagain> actually, to fit 32 bit value, we have space for factor 107 multiplication. Good point zoff about special values (or special bit, irrc) used in maptool to flag resolved/unresolved nodes.
21:28 #navit: < tryagain> there could be a problem with math we do on coordinates. We could get an overflow attempting to add two coordinate values.
21:29 -!- xenos1984 [~xenos1984@127-166-191-90.dyn.estpak.ee] has quit [Quit: Leaving.]
21:37 #navit: < tryagain> And while we have not matching resolutions in osm and binfile, we could still hit some cases where turn direction should be deducted from a 0.5m way segment  having 1cm inclination, which angle is to be compared to a neighboring long way segment.
21:42 #navit: < tryagain> i think we could start with automatic detection of cases where turn angle difference is below our map resolution, and warn the driver about uncertain maneuver.
21:43 #navit: < tryagain> so we will be fair by not giving wrong directions, an will have some statistics about such cases.
21:51 #navit: < tryagain> actually, osm coordinates have 7 decimal places, so we need almost the whole 32bit signed int range to represent them as int_val*1e-7. http://wiki.openstreetmap.org/wiki/Node
21:52 #navit:  * tryagain is not going to fix that overnight
21:52 -!- tryagain [~quassel@178.216.76.43] has quit [Remote host closed the connection]
22:10 -!- tg [~x@unaffiliated/tg] has quit [Ping timeout: 276 seconds]
--- Log closed Thu May 28 00:00:21 2015