flooded regions on motorola defy, razr i and xperia active

271 views
Skip to first unread message

christian orjeda

unread,
Apr 14, 2013, 9:19:11 AM4/14/13
to osm...@googlegroups.com
Hello,
on three different phones the osmand is showing the same error (the
region of Schaffhausen Switzerland and and a part of the lake of
Constance region is flooded (great parts of land are tainted blue). I
don't know why, even a new install and a new map-download didn't help.
The 3 phones are motorola defy and razr i and a sony xperia active; all
3 with jb 4.1.
The same error hapens within map droyd. This error remains all over the
last 3 or 4 map actualisations.
Excuse my bad English.

Any ideas for solving this problem?

Hardy

unread,
Apr 15, 2013, 3:01:05 PM4/15/13
to osm...@googlegroups.com
Hi, I think I have seen some discussion on this before, but am not aware of the latest status. If you search this forum for the "Bodensee" issie, you may find some explanations and information.
 
Best,
Hardy

Victor Shcherb

unread,
Apr 15, 2013, 4:04:32 PM4/15/13
to osmand
I believe this issue hardly could be solved in OsmAnd, simply because data is partially present in obf file . 

First of all if the lake is defined as coastlines it would be better rendered on higher zooms, but on lower zooms it will not be rendered because of broken polygon.

Secondly, if the lake is defined as relation of ways then it could not be solved in easy way. All previous attempts failed. However there is simple solution include this lake in all countries/regions. If someone could provide such boundaries, we could generate the map and see if the issue has gone.

Victor


--
You received this message because you are subscribed to the Google Groups "Osmand" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

Manfred

unread,
May 8, 2013, 6:57:57 PM5/8/13
to osm...@googlegroups.com
Hi Victor!

Nothing new on this matter?

There is so much going on (different colours for houses with/without numbers, routable world map,.........., bells and whistles nobody needs) but showing a lake on a map correctly seems to be impossible.

Here a list of threads dealing with this problem.
Maybe there can be found some hints or persons who could contribute.

https://groups.google.com/forum/#!topicsearchin/osmand/bodensee/osmand/CjDHy0V-aY4
https://groups.google.com/forum/#!topicsearchin/osmand/bodensee/osmand/hPAb929NTao
https://groups.google.com/forum/#!topicsearchin/osmand/bodensee/osmand/3p9kGsfr2q0
https://groups.google.com/forum/#!topicsearchin/osmand/bodensee/osmand/95THMLcAt_w

Regards
Manfred

Harry van der Wolf

unread,
May 9, 2013, 11:33:09 AM5/9/13
to osm...@googlegroups.com


2013/5/9 Manfred <alfman....@gmail.com>

Nothing new on this matter?

There is so much going on (different colours for houses with/without numbers, routable world map,.........., bells and whistles nobody needs)

well, that depends on how you want or need to use Osmand. That's very personal.
I dare say that OsmAnd will calculate a nice route and that you can still drive over the "flooded" roads even though the fact that you don't have a 4-wheel drive :)
It has been mentioned a couple of times already: It is the OpenStreetMap that contains a bug, not OsmAnd. Try one of the other OSM based apps, and they show the same. So comparing an OSM bug with "all the bells and whistles" for OsmAnd is not correct.
See <https://groups.google.com/forum/?fromgroups=#!topic/osmand/95THMLcAt_w>
 
Why not become a contributor yourself?
OSM is an open-source project and if you feel an itch, you simply scratch (yourself).
That's also how I became a contributor: I needed something but nobody did it for me, so I built up the knowledge and did it myself.
Open-source doesn't mean that you get everything for free without doing anything, saying that you don't have the knowledge. That's how commercial software is working.

So please join the contributor community.

Harry

Manfred

unread,
May 10, 2013, 10:45:05 AM5/10/13
to osm...@googlegroups.com
Harry van der Wolf!

Maybe I am not as smart as you, but I am sure, you are not as smart as you think you are.

There seems to be a lot you don't know about OsmAnd an the Bodensee problem.

For example, that OsmAnd once stood for OSM for Android, and its main feature was to display locally stored vector maps.
This was and is a basic functionality it should do without such problems.

Or that about a year ago, Bodensee was dry.
Until Sanderd17, a guy that knows more about OSM than me and you together, built a Bodensee.osm, which, converted to obf and saved in the osmand-folder solved the problem.
Unfortunately only for a short time. After some changes ins OsmAnd (not in OSM-data!) parts of the Swiss coast is flooded.

So why can you say, that is a problem of OSM data only, and not of OsmAnd?
I think, there are things to do on both sides.

Why looks:
http://www.openstreetmap.org/?lat=47.641&lon=9.178&zoom=11&layers=M
O.K.?

What I want is, that this problem is solved. If I could, I would have done already.
Experts from both sides should discuss what cause the problems in detail, and what by whom could/should be done to solve.


BTW:
I contribute longer than you, but not so loud. ;-)

Und dänn sottist d'Schauörö nid so witt ufrupfö, wändt mit arwachsna redast, du Rotzschauörar, du gröüna!


Regards
Manfred

Victor Shcherb

unread,
May 10, 2013, 11:12:47 AM5/10/13
to osmand
To be honest Bodensee is well known problem. I think, the way Sander approach I should admit is not the best.

Sander work on OsmAnd code to get it working when relations are incomplete. But it doesn't solve the issue completely as you see. One of my proposal and still it is, to map in OSM as coastline. In that case on zooms > 15 (when coastlines will not be broken) it will show a nice waterline as we see with seas and oceans.

Another solution will be just make this lake relation complete in that case we should include it into all boundaries for these countries.  Fortunately now we don't dependent very much on geofabrik (we have own processing) we can do it. If some one can provide updated poly's for Switzerland, Austria... https://github.com/osmandapp/OsmAnd-misc/tree/master/osm-planet/polygons/europe feel free.


Shortly saying, we should not address a change in OsmAnd for now, but for OSM (one big coastline or many small) or restructure countries polygons.


Best Regards,
Victor


--

Harry van der Wolf

unread,
May 10, 2013, 12:28:06 PM5/10/13
to osm...@googlegroups.com
Hi Manfred,

2013/5/10 Manfred <alfman....@gmail.com>

Harry van der Wolf!

Maybe I am not as smart as you, but I am sure, you are not as smart as you think you are.

I never said I was that smart. I'm also asking questions in this group because I don't know a lot of things.


There seems to be a lot you don't know about OsmAnd an the Bodensee problem.

For example, that OsmAnd once stood for OSM for Android, and its main feature was to display locally stored vector maps.

I know.
 
This was and is a basic functionality it should do without such problems.

Or that about a year ago, Bodensee was dry.
Until Sanderd17, a guy that knows more about OSM than me and you together, built a Bodensee.osm, which, converted to obf and saved in the osmand-folder solved the problem.

Again: I didn't say I was smart or very knowledgable. That's exactly what I mentioned in the thread I was referring to: That I did look into it and that I didn't have enough knowledge about these large polygons. I have been using my (limited) knowledge on other aspects.

 
Unfortunately only for a short time. After some changes ins OsmAnd (not in OSM-data!) parts of the Swiss coast is flooded.

So why can you say, that is a problem of OSM data only, and not of OsmAnd?
I think, there are things to do on both sides.

Why looks:
http://www.openstreetmap.org/?lat=47.641&lon=9.178&zoom=11&layers=M
O.K.?

What I want is, that this problem is solved. If I could, I would have done already.
Experts from both sides should discuss what cause the problems in detail, and what by whom could/should be done to solve.


Agreed.


BTW:
I contribute longer than you, but not so loud. ;-)


Very nice that you're a contributor (no pun intended), but you didn't sound like it: You mentioned an issue and you wanted to have it solved: plain like that.
Hence my reaction as I've seen that in many Open Source projects.

And yes: I don't contribute very long to OsmAnd, but there have been and still are many OS projects since I joined the first in Oct 1994 (and some were and still are my own).
I don't mention this to brag (seriously not), but as an argument for my next sentence.
What I have communicated loudly before and also did in my previous post ;) , is that Open Source projects need contributors as they can only survive, grow and progress of the work done by these volunteer contributors, and the fact that you are one of them is great and that you were addressed as not being one is  regrettable.

Please translate the swiss part for me. I don't know whether it is just a signature part or meant for me: if so I like to know whether it stimulates me or whether it stimulates me even more.

Harry


Manfred

unread,
May 10, 2013, 2:33:27 PM5/10/13
to osm...@googlegroups.com
Hi Victor!

Pleased to hear, that Bodensee is a well known problem.
I thought, it was a forgotten one ;-)

Glad to to hear the "... we can do it"
Who can do the changes in OSM data (what skills are necessary)?
And who tells the ones, that could do, that they should do?


Best Regards
Manfred

Victor Shcherb

unread,
May 10, 2013, 4:10:49 PM5/10/13
to osmand
Hi Manfred,

No skill is required to edit for example poly file in JOSM with openstreetmap underlay. You can take poly file http://download.geofabrik.de/europe/switzerland.html and use 2 simple activities osm2poly & poly2osm get loaded into JOSM and change it. After that we can put it here https://github.com/osmandapp/OsmAnd-misc/tree/master/osm-planet/polygons/europe and try to regenerate it  

No particular skills is required to replace tag from way to coastline, but you will need local OSM community approve for that.

P.S. : if that is a known issue, it doesn't mean that I'm doing anything about it :) I presume it is a data issue rather than software (preparation step or design step).

Victor

Manfred

unread,
May 10, 2013, 8:21:34 PM5/10/13
to osm...@googlegroups.com
Hi Victor!

"No skills" sounded good, but it took me hours to find and install all I need: Perl for Windows, the scripts and switzerland.poly and find out the syntax of the scripts.

Now I have converted switzerland.poly to *.osm and opened it in JOSM, with OSM Mapnik as background.

What should I do next?

Change the Swiss poly line to include the whole Bodensee?

Regards
Manfred
Message has been deleted

Harry van der Wolf

unread,
May 11, 2013, 7:28:13 AM5/11/13
to osm...@googlegroups.com
Could the JOSM scanaerial plugin (http://wiki.openstreetmap.org/wiki/Scanaerial) be a good help here?

I had a look at it during the previous mail exchange (end of March) but didn't dive further into it

It might simplify and reduce the amount of work.


Harry van der Wolf

unread,
May 11, 2013, 8:13:24 AM5/11/13
to osm...@googlegroups.com
As it takes quite some python dependencies which might be quite hard to install on windows I did the same on my Ubuntu box as linux is python enabled by default.
Please find attached a screen shot. I had to use scanaerial twice. Once for the bodensee and a second "click" for the untersee as well.
Please find attached a screenshot to display what has been captured (including the island Mainau), and an osm poly file for the bodensee and the bodensee+untersee.

As the bodensee is way out of "my region" I will stop now, but I see this also as a learning curve. If tools/options are readily available on linux but not on python I can step in again.



2013/5/11 Harry van der Wolf <hvd...@gmail.com>
bodensee_and_untersee.jpg
bodensee.osm
bodensee_and_untersee.osm
Message has been deleted
Message has been deleted

Victor Shcherb

unread,
May 11, 2013, 7:13:14 PM5/11/13
to osmand
Yes, please open osm file and edit it to include bodensee (use osm maps for underlay) !  After that convert back to poly and send it to me.
Or if it is not difficult press "+" button here https://github.com/osmandapp/OsmAnd-misc/tree/master/osm-planet/polygons/europe (fork and edit) and put content of the file - create pull request :) That would be great. 

I will update script to include switzerland as cut https://github.com/osmandapp/OsmAnd-misc/blob/master/osm-planet/script.sh and in 5 days it willappear here http://download.osmand.net/osm-downloads/ ready for osmand indexing.

Sorry I know for Windows user, it is just a hell. That's why 2 years ago I switched to Linux where it is just 2 command lines (copy/paste).

Manfred

unread,
May 11, 2013, 7:48:09 PM5/11/13
to osm...@googlegroups.com
Hi Victor!

I hope, I did it right.

Are there the self made OSM-files (not from geofabrik):
http://download.osmand.net/osm-downloads/
?

Regards
Manfred

Victor Shcherb

unread,
May 12, 2013, 7:40:06 AM5/12/13
to osmand
I scheduled map special generation http://new.osmand.net:8080/job/GenerateIndexes/163/console .

So now you can try to download and test.

Victor

Manfred

unread,
May 12, 2013, 7:47:44 AM5/12/13
to osm...@googlegroups.com
Hi Victor!

I did generation over night, am going to test right now.

Regards
Manfred
Message has been deleted

Manfred

unread,
May 12, 2013, 9:30:13 AM5/12/13
to osm...@googlegroups.com
Hi Victor!

I tested, but things didn't get better.
In different zoom levels, different amount of land is flooded, in some, parts of the lake are dry.
But I think that is not new, happened with old Switzerland.obf too.

Regards
Manfred


Am Sonntag, 12. Mai 2013 13:40:06 UTC+2 schrieb V S:

Christof Knu

unread,
May 12, 2013, 3:02:14 PM5/12/13
to osm...@googlegroups.com
and it's not only a Bodensee problem.
Same (or simular) things happen at Lago Maggiore (regions flooded), Lago di Lugano (one branch is dry), Lago di Como, which is completely inside Italy (one branch is dry)...
It happens with water and vegetation (s. attachment).
In my eyes really an issue, that should be solved (but don't know, how...).

Regards,
Christof
Lago_di_como.jpg

Victor Shcherb

unread,
May 12, 2013, 4:41:01 PM5/12/13
to osmand
Indeed this is different and that require special investigation. 
I think the multipolygon was not correctly processed, but it would be good at least to know if it happens on zoom >= 15 or lower or both.

Victor

Christof Knu

unread,
May 13, 2013, 6:57:11 AM5/13/13
to osm...@googlegroups.com
Sorry, I have to apologize the things I mentioned seem to be a problem of old maps.
I downloaded the new Italian map, and all water problems I mentioned have gone.
The only thing remaining is a vegetation-cut at lago di como, as shown in the picture.
Regards,
Christof

Victor Shcherb

unread,
May 14, 2013, 1:39:47 PM5/14/13
to osmand
Hi,

I also can confirm, now it looks good :) Except 13th zoom when it is sometimes (!) you see dry area instead of lake, but there is no flooded areas.

Victor


On Sat, May 11, 2013 at 6:30 PM, Max <openstre...@nurfuerspam.de> wrote:
I realy don't understand it, all my messages are deleted immediately.
Automatic deletion?

christian orjeda

unread,
Jul 3, 2013, 2:34:20 AM7/3/13
to osm...@googlegroups.com


Am Sonntag, 12. Mai 2013 15:30:13 UTC+2 schrieb Manfred:
the bodensee, schaffhausen-flooded problem remains. with every map-update, i've the hope it will be fixed, but it remains. on my pc with the map-ditor the flooded regions aren't visibl and my knowledge to edit is beetween zero and basic.

christian orjeda

unread,
Jul 3, 2013, 3:07:06 AM7/3/13
to osm...@googlegroups.com
bodensee-flooded.png

christian orjeda

unread,
Jul 3, 2013, 3:15:05 AM7/3/13
to osm...@googlegroups.com
schaffhausen-flooded.png

Harry van der Wolf

unread,
Jul 3, 2013, 10:47:14 AM7/3/13
to osm...@googlegroups.com
Please find attached two screenshots of the relevant regions of the 01-07-2013 Switserland map, running in OsmAnd+ 1.4.1 beta on a 800x480 screen.
To reduce size for slower and/or Android clients I converted them to relatively low-quality jpg.
The maps are fine.



Schaffhausen_correct.jpg
Bodensee_correct.jpg

Manfred

unread,
Jul 3, 2013, 2:23:00 PM7/3/13
to osm...@googlegroups.com
Hi Harry!

Looks nice, but is totally different from what I am seeing on my Device.
Those Regions are flooded!

Latest nightly, latest map of Switzerland, Baden-Württhemberg Bayern and Austria

Regards
Manfred

Harry van der Wolf

unread,
Jul 3, 2013, 3:23:47 PM7/3/13
to osm...@googlegroups.com


2013/7/3 Manfred <alfman....@gmail.com>

Hi Harry!

Looks nice, but is totally different from what I am seeing on my Device.
Those Regions are flooded!

Latest nightly, latest map of Switzerland, Baden-Württhemberg Bayern and Austria


That means two things:
- The .obf is correct, otherwise it would render incorrectly on my device as well.
- The OsmAnd app displays the (correct) map incorrectly on some devices.

My resolution is 480x800 (portrait). I did see that Christian's images are 540x960, so in another resolution. I have no clue whether that has an influence on the rendering inside OsmAnd, but that's simply not clear yet.
Maybe we should check at which devices it displays correctly, on which OsmAnd version (although I don't think that's really relevant as it seems to happen in all versions as of 1.2.1 and newer), in which resolution, with or without the "high resolution enabled" when relevant , and on which android version.

So if a couple of users could check this we can hand-over a better bug report to hopefully simplify his bug fixing task.

So please report:
- Android device
- device resolution
- Android version: Android settings->about the phone; version and kernel (kernel with regard to possible level of graphic (opengl and other) drivers; might not be relevant though).
- OsmAnd version
- Use of high resolution enabled/disabled, and does it make a difference.

If I forget important things please add them as well.

For me that is: Alcatel 993D, 480x800, Android 4.0.4 on kernel 3.0.8, OsmAnd+ 1.4.1 beta, no "high resolution" option available due to my device's resolution.
Note that on my previous Switserland map (12-06-2013 I think), I also had  also a correct rendering of the area.

Harry

Manfred

unread,
Jul 3, 2013, 3:59:39 PM7/3/13
to osm...@googlegroups.com
Hi Harry

ZTE, 480x800,Android 2.3.5

High resolution does not matter

Regards
Manfred

Manfred

unread,
Jul 4, 2013, 6:31:14 AM7/4/13
to osm...@googlegroups.com, osman...@gmail.com
Hi Harry!

The "Bodensee-problem" is not device dependent (which I never thought).
If you add the map of Baden-Württemberg you will see, what I see: a lot of flooded land in Switzerland.

So the conclusion is, that the wrong display of regions in Switzerland is caused from map of Germany!
Maybe, somebody knows a way to fix this?

Regards
Manfred

Harry van der Wolf

unread,
Jul 5, 2013, 10:16:51 AM7/5/13
to osm...@googlegroups.com
I downloaded the Baden-wuerttenberg map and I see the flooded regions as well. So you are completely right.
 
As you are aware the polygons used to create the country maps run as close as possible around (=outside) the country border, but you always have a small overlap at borders (in every map/navigation application).
The administrative border between Germany and Switserland is running through the Bodensee and should have some overlap in the bodensee but also in/at Konstanz. In this case OsmAnd is using it's own polygon for Switserland including the entire Bodensee. You created this extended poly some time ago specifically to prevent this flooding.
So this is what I tried:
- I tried with the OsmAnd Switserland map (which was extended to include the Bodensee) and I still have the flooding.
 
- I tried the Switserland map with my own Germany_major_roads (which I published some time ago along with some others) and that doesn't show the flooding. Point is that the major_roads maps I made don't display landuse. So there is no rendering conflict (if it is a rendering conflict) between the Germany map and the Switserland map as only Switserland shows landuse.
 
- I downloaded the original switserland map osm.pbf from geofabrik.de and converted it to obf. Using the geofabrik original Switzerland map still gives me flooding. So, using the Osmand extended switzerland or the geofabrik switzerland map doesn't make a difference at all: it is still flooded.
 
- I downloaded the Baden-wuerttenberg poly from geofabrik and did the same what you did for Switzerland: I included the entire Bodensee and a very small part of Switserland inside the Baden-wuertenberg poly and created a map for this extended Baaden-wuertenburg poly. This extended Baaden-Wuertenburg poly in combination with both the Osmand extended Switserland map, and with the "normal" geofabrik Switseland map still gives flooding. (I created this map by first merging the geofabrik Switserland.osm with the Baden-wuerttenberg.osm using osmconvert. From the combined Bw_sw.osm I created the extended Baaden-wuerttenburg map using my extended Baaden-Wuerttenburg poly).
 
- Finally I created a combined map from the combined Switserland-Baaden-Wuerttenburg.osm (which I created in the previous step). This combined Switserland-Baden-Wuerttenberg.obf (bw_sw_2.obf) displays it all correctly. No flooding at the coast (Konstanz, Münsterlingen, Romanshorn, etc) and no flooding east of Shaffhausen. Note that it took quite long to generate this map, hence my late reaction.
 
- I downloaded Be-On-Road and NavFree including the Germany and Switserland maps. They do not show the flooded regions. (Just to make sure it really is an OsmAnd issue and not something else)
 
 
The other problem which causes the flooding East of Schaffhausen (I think) is the fact that Baden-Wuertenbrg contains an isolated spot of land inside Switserland. See attached "Baaden-wuertemburg_around_Schafffhausen.jpg" image. The Baden-wuerttenberg poly simply includes this in the map by using a big "bulb" that includes this area (not a separate poly) causing a big overlap. I think this area is incorrectly rendered by OsmAnd thereby creating the flooded region on top of Switserland.
Out of curiosity I checked this with "Baarle-Hertog", which is a minor Belgium village within the borders of Nederland, but that one is rendered correctly. Belgium consists of multiple polygons. One for Belgium itself and also a minimal for Baarle-Hertog.
 
Harry
 
 
2013/7/4 Manfred <alfman....@gmail.com>

--
Baaden-wuertemburg_around_Schafffhausen.jpg

Harry van der Wolf

unread,
Jul 5, 2013, 1:07:16 PM7/5/13
to osm...@googlegroups.com


2013/7/5 Harry van der Wolf <hvd...@gmail.com>

 
- Finally I created a combined map from the combined Switserland-Baaden-Wuerttenburg.osm (which I created in the previous step). This combined Switserland-Baden-Wuerttenberg.obf (bw_sw_2.obf) displays it all correctly. No flooding at the coast (Konstanz, Münsterlingen, Romanshorn, etc) and no flooding east of Shaffhausen. Note that it took quite long to generate this map, hence my late reaction.
 


The rendering is still not correct. At zoomlevel 13 the water suddenly turns to land. zoomlevel 12 and below is OK, and zoomlevel 14 and up is OK as well. I tested with several rendering profiles. It really must be an OsmAnd bug or a rendering profile bug in default.render.xml (or rendering_types.xml?), but I can't find it.
In this case I think it is another bug then the flooding although it might be related.


Harry

Victor Shcherb

unread,
Jul 7, 2013, 8:01:39 PM7/7/13
to osmand
I remember that Bodensee was correct from point of Switzerland a month ago and I know that there was no changes about rendering from OsmAnd side. 
Could we trace back the problem? Is it possible that current polygon is broken? Were some changes in that area.
Please provide information how it is registered now : multipolygon? coastline?

Regards,
Victor


--

Manfred

unread,
Jul 8, 2013, 6:13:28 AM7/8/13
to osm...@googlegroups.com
Hi Victor!

I think Bodensee never was O.K. since a lot of months.

Regards
Manfred

Victor Shcherb

unread,
Jul 8, 2013, 6:54:41 AM7/8/13
to osmand
Look in the thread :) So far it is like that then it is very minor issue.
>>>>>>>>>> Victor
>>>> 13th May

Hi,
I also can confirm, now it looks good :) Except 13th zoom when it is sometimes (!) you see dry area instead of lake, but there is no flooded areas.

>>> Harry 7th May
The rendering is still not correct. At zoomlevel 13 the water suddenly turns to land. zoomlevel 12 and below is OK, and zoomlevel 14 and up is OK as well. I tested with several rendering profiles. It really must be an OsmAnd bug or a rendering profile bug in default.render.xml (or rendering_types.xml?), but I can't find it.
In this case I think it is another bug then the flooding although it might be related.


Manfred

unread,
Jul 8, 2013, 10:08:29 AM7/8/13
to osm...@googlegroups.com
Hi Victor!

I read that.
But I cannot believe, that you meant the Bodensee problem was solved, but the one with Lago Maggiore (old Italian maps)

But I read also, that only if Switzerland and Baden-Württemberg are within ONE map, most of Swiss coast aren't flooded any more.
For me, it would fit, if there would be ONE map of Switzerland, Austria, Bayern and Baden-Württemberg (all countries around teh Bodensee), but there seems to be something wrong with the Baden-Württemberg and/or the rendering of it.

Regards
Manfred

Harry van der Wolf

unread,
Jul 8, 2013, 11:31:05 AM7/8/13
to osm...@googlegroups.com
Hi Victor and Manfred,

Manfred is right. When I create a map combining Switserland and the Bodensee, there is no flooding of the regions at all!
And indeed: at zoomlevel 13 most water "turns into" land. I think that is a separate issue (although it can be related). I did check the deafult.render.xml but I can't find an error, but I'm not the real expert.

I did some careful checking of ways and polygons and the Bodensee polygon is really closed. The Baden-Wuerttenberg polygon is correct as well.

I think the issue is that when (big) water makes part of the border of any map, OsmAnd treats it as "outer water" of an open polygon, thereby extending it into the next map and flooding real land (also at Schaffhausen with the Rhine (der Rhein).
When I create the combined map, the Bodensee is of course one closed polygon as it is inside the entire map.
I think that the Italian border lakes suffer from the same issues: polygons that are really closed when inside the map, but now open due to the fact that they cross map borders. Either OsmAnd doesn't render them right or OsmAndMapCreator doesn't create them correctly.

This issue should be solved (says someone who can't program in C/C++ and/or Java).
Another solution in the mean time could be to generate the so-called DACH map. Geofabrik calls it Alps (http://download.geofabrik.de/europe/alps.html). It does cover the entire Bodensee. I didn't check/created it yet myself (I'm currently on holiday), but I think that some of the Italian border lake problems are also gone with this Alps (DACH) map as these lakes now also fall inside that map.

Harry


2013/7/8 Manfred <alfman....@gmail.com>

Harry van der Wolf

unread,
Jul 8, 2013, 11:32:44 AM7/8/13
to osm...@googlegroups.com


2013/7/8 Harry van der Wolf <hvd...@gmail.com>


Manfred is right. When I create a map combining Switserland and the Bodensee, there is no flooding of the regions at all!

Sorry, I meant of course Switserland and Baden-Wuerttenberg. sorry for this sloppy mistake.

Harry

Victor Shcherb

unread,
Jul 8, 2013, 5:39:58 PM7/8/13
to osmand
Sorry Harry, 
I missed who is right. Isn't it proves my point? Because I'm saying that Switzerland including Bodensee - no issue!
And our map is ok because we maintain exactly that switzerland.poly.

So there are might be issues if you download other maps with Switzherland, Germany ... because combination of maps also can give flooded areas. Basically we need to maintain all maps in order to take support easy and don't see flooded areas. 

Other solutions could be to realize during runtime that polygon is broken and don't draw anything. To be honest coastline could be a solution but it is much less error prune than completely included multipolygon.

Regards,
Victor


--
You received this message because you are subscribed to the Google Groups "Osmand" group.

Manfred

unread,
Jul 8, 2013, 6:08:19 PM7/8/13
to osm...@googlegroups.com
Hi Victor!

There are issues,not only might be.

Swiss map alone -> O.K.
Add Baden-Württemberg map -> Swiss coast is flooded.

There are no problems with Austria and Bayern, which meets each other on the eastern part of Bodensee.
Why, or better, why not?

Regards
Manfred

Victor Shcherb

unread,
Jul 8, 2013, 6:27:49 PM7/8/13
to osmand
>> Add Baden-Württemberg map -> Swiss coast is flooded.
This is expected. We can do the same trick as we did for Switzerland https://github.com/osmandapp/OsmAnd-misc/tree/master/osm-planet/polygons/europe

Harry van der Wolf

unread,
Jul 9, 2013, 11:07:13 AM7/9/13
to osm...@googlegroups.com
Hi, I don't care who is wrong or right if we can solve it together.   Helping each other to get in the right direction is what counts.

2013/7/8 Victor Shcherb <victor....@gmail.com>

Sorry Harry, 
I missed who is right. Isn't it proves my point? Because I'm saying that Switzerland including Bodensee - no issue!
And our map is ok because we maintain exactly that switzerland.poly.

No that is not correct.  a separate Switserland incl. bodensee map, like Manfred made some time ago, still has an issue if on your device with the Baden-wuerttenberg map.

The Switserland-with-Bodensee map and a separate Baden-wuerttenberg doesn't work. that gives flooding.
One single combined map including Switserland and Baden-Wuerttenberg works.

Harry

Manfred

unread,
Jul 9, 2013, 11:38:42 AM7/9/13
to osm...@googlegroups.com
Hi Harry!

Maybe he thought of making a Baden-Württemberg map including the Bodensee too?

Regards
Manfred

Harry van der Wolf

unread,
Jul 9, 2013, 12:40:27 PM7/9/13
to osm...@googlegroups.com
Hi,

I already did that, but it doesn't help.
Please (re)read my mail of 5 July 15:16. It describes what I tried.
Please let me know if I forgot options. I'm willing to try, but I 'm on holidays so it's at a very low priority right now.  If I have absolutely nothing else to do I will try and otherwise after my holidays.

Harry

2013/7/9 Manfred <alfman....@gmail.com>

--

Victor Shcherb

unread,
Jul 9, 2013, 12:52:47 PM7/9/13
to osm...@googlegroups.com
Hi,
Let me be absolutely clear. If one map contain broken Bodensee than all maps will generate errors. To narrow down that map you need to deactivate everything except it.

My point is Simgle Switzerland is ok. All other could be broken


--
Sent from Gmail Mobile

Harry van der Wolf

unread,
Jul 9, 2013, 2:34:31 PM7/9/13
to osm...@googlegroups.com


2013/7/9 Victor Shcherb <victor....@gmail.com>

Hi,
Let me be absolutely clear. If one map contain broken Bodensee than all maps will generate errors. To narrow down that map you need to deactivate everything except it.


That's also exactly what I'm saying, but I come to a different conclusion.
The bodensee poly is absolutely correct.
When you look at the maps, the Bodensee is "cut in half" by Switzerland and Germany. For OsmAnd or OsmAndMapCreator this means probably a broken Bodensee, but that is an OsmAnd or OsmAndMapCreator error. Therefore these maps contain flooded areas.
If I merge the "Switserland.osm" and the "Baden-Wuerttenberg.osm" into one single "Switserland_Baden-Wuerttenberg.osm" with osmconvert and then create a "Switserland_Baden-Wuerttenberg.obf" it works fine amd the map does not contain flooded areas.

What I am saying is that all the polys that are used (Switserland, Baden-wuertenberg and the Bodensee) are correct. But OsmAnd or OsmAndMapCreator treats the Bodensee obviously as a broken polygon as it is only partly in these maps. That is not a polygon error, that's an OsmAnd error.

 
My point is Simgle Switzerland is ok. All other could be broken


I tested the polys for Switserland (the OsmAnd extended one AND the geofabrik normal one) and I tested the Baden-wuerttenberg poly and I tested the Bodensee poly and they are all correct. This means no broken polygons.
What's more: I downloaded and tested it also with Be-On-road and with Navfree and both work fine as well.
The night before I went on holiday I compiled Navit and created with their mapcreator tool also a Switserland map and a Baden-wuerttenberg map (from geofabrik data). Navit also displays correctly. (And their mapcreator tool is very fast by the way)
Only OsmAnd does not display it correctly!
 
It is not an error in the (original) mapdata. It is an error in OsmAnd and/or OsmAndMapCreator.

Sorry if I seem too pushy, but due to the fact that we are all non-native english speakers I try to make it as clear as possible and repeat some things.

best,
Harry
 

Victor Shcherb

unread,
Jul 9, 2013, 4:00:01 PM7/9/13
to osmand
Hi Harry,
- What do you mean Baden-wuerttenberg poly correct? What is Bodensee poly correct?

You probably missing the main point. All programs have different rendering engines and different multipolygon processors. This is not about correct or incorrect polygons. All programs need to have some specifics for poly boundaries. Most of the program do triangulation for rendering multipolygons. OsmAnd does it differently. 

And let me be explicit you can treat of course as OsmAnd error or OsmAndMapCreator, but that can not be fixed there. With current rendering engine it is just not possible. 

The only possibility is to prepare "correct" boundaries for all regions connected to Bodensee.

P.S. : May be it was a big mistake to include in the file broken polygons and let the lake be completely dry, but at least it will be very visible how to fix it .


Regards,
Victor




--

Manfred

unread,
Jul 9, 2013, 4:33:34 PM7/9/13
to osm...@googlegroups.com
Hi Harry!

As an Austrian, I must say it:

The Bodensee is shared by Switzerland, Germany (Baden-Württemberg and Bayern) and Austria.
There are no problems at the eastern part of Bodensee, where Austria and Bayern meet (there is no connection between Bayern and Switzerland).

As I wrote earlier, I have installed all those maps.
Deactivating Baden-Württemberg removes the flood on Swiss coast.
I don't know, how these things work, but this map seems to be the bad one.
Or makes the trouble somewhere else visible.

Regards
Manfred

Harry van der Wolf

unread,
Jul 9, 2013, 4:49:19 PM7/9/13
to osm...@googlegroups.com


2013/7/9 Victor Shcherb <victor....@gmail.com>

Hi Harry,
- What do you mean Baden-wuerttenberg poly correct? What is Bodensee poly correct?

You probably missing the main point. All programs have different rendering engines and different multipolygon processors. This is not about correct or incorrect polygons. All programs need to have some specifics for poly boundaries. Most of the program do triangulation for rendering multipolygons. OsmAnd does it differently. 


I didn't know that. So I was indeed "missing that point". It does not make it clear why it does not work correctly.
 

And let me be explicit you can treat of course as OsmAnd error or OsmAndMapCreator, but that can not be fixed there. With current rendering engine it is just not possible. 

The only possibility is to prepare "correct" boundaries for all regions connected to Bodensee.


Can you explain a little more? I already created a polygon for Baden-wuerttenberg that included the bodensee, like the OsmAnd Switserland poly does. My Baden-Wuerttenberg poly didn't solve it either. I must say that I didn't include the Untersee. I went up at Konstanz. I will redo it and include the Untersee as well.
Anyway: please explain what should be done.

Harry

Harry van der Wolf

unread,
Jul 9, 2013, 4:51:42 PM7/9/13
to osm...@googlegroups.com


2013/7/9 Manfred <alfman....@gmail.com>

Hi Harry!

As an Austrian, I must say it:

The Bodensee is shared by Switzerland, Germany (Baden-Württemberg and Bayern) and Austria.

Sorry, you are right. I have Austria installed otherwise I got a white area.
It's just that it doesn't matter with Bayern in place (what I forgot to mention in my first mails) and Austria doesn't matter either.

Harry

Nico W

unread,
Jul 11, 2013, 4:09:37 PM7/11/13
to osm...@googlegroups.com
Harry, I do see the same thing on my Motorola Defy, and that is on the world map. I do not have those parts of Switserland or Gemany downloaded, so the problem starts very early. The lake is visible down to zoomlevel 11 with borderline going straight through the middle of it. At zoomlevel 12 the lake disappears but the border remains, and all I see is the text Bodensee.
 

On Sunday, April 14, 2013 6:19:11 AM UTC-7, christian orjeda wrote:
Hello,
on three different phones the osmand is showing the same error (the
region of Schaffhausen Switzerland and and a part of the lake of
Constance region is flooded (great parts of land are tainted blue). I
don't know why, even a new install and a new map-download didn't help.
The 3 phones are motorola defy and razr i and a sony xperia active; all
3 with jb 4.1.
The same error hapens within map droyd. This error remains all over the
last 3 or 4 map actualisations.
Excuse my bad English.

Any ideas for solving this problem?

Victor Shcherb

unread,
Jul 13, 2013, 10:17:36 AM7/13/13
to osmand
Hi,

I probably should say it one way or another.
We need to maintain own boundaries for all countries around Bodensee i.e. Germany, Austria, Switzerland (Italy?) and make them include Bodensee. 
That's it! There is no another solution.

P.S. From my side now I can provide a fix, that will no be flooded areas, but Bodensee will be dry (also not good).
So let's update boundaries and that's it.

Regards,
Victor


--

Manfred

unread,
Jul 13, 2013, 11:07:41 AM7/13/13
to osm...@googlegroups.com
Hi Victor!

That would be:
Austria
Switzerland
Germany (only the parts Baden-Württemberg and Bayern)
NOT Italy

I will try it for Baden-Würtemberg.


On Lac Lemmond (Genfersee) the same Problem exists.
The border running through is between Switzerland and France.
But this lake is to far away to be of interest for me ;-)

Regards
Manfred

Manfred

unread,
Jul 14, 2013, 6:05:49 AM7/14/13
to osm...@googlegroups.com
Hi Victor!

Cannot find new Baden-Wuerttemberg.
I think, it should have been generated?

Regards
Manfred

Manfred

unread,
Jul 14, 2013, 6:01:04 PM7/14/13
to osm...@googlegroups.com
Hi!

Seems, that new Baden-Wuerttemberg map did it!
The great Bodensee flood is over! Hallelujah!

Regards
Manfred

Manfred

unread,
Jul 14, 2013, 6:04:35 PM7/14/13
to osm...@googlegroups.com
Hi!

But the lake seems to be dry in zoom level 13 :-(

Regards
Manfred

Harry van der Wolf

unread,
Jul 15, 2013, 12:01:15 PM7/15/13
to osm...@googlegroups.com
Am Montag, 15. Juli 2013 00:01:04 UTC+2 schrieb Manfred:Hi!

Seems, that new Baden-Wuerttemberg map did it!
The great Bodensee flood is over! Hallelujah!

Great! But who did what?  :)



2013/7/14 Manfred <alfman....@gmail.com>

Hi!

But the lake seems to be dry in zoom level 13 :-(

Regards
Manfred



Yes, but that's already quite some time the case. Osmand error or?


@Victor: You mentioned that coastline would be better.
Does that mean that the Bodensee polygon should be transfered to a Coastline polygon? Or another polygon next to the Bodensee polygon itself?

Harry

christian orjeda

unread,
Jul 22, 2013, 8:41:03 AM7/22/13
to osm...@googlegroups.com

can cofirm, the flood is over. Thank you very much, who ever solved it.

pl

unread,
Jul 31, 2013, 6:39:44 AM7/31/13
to osm...@googlegroups.com

Google maps has also problems with rivers and borders: link :)
Danube river missing between Silistra (Bulgaria) and Chernavoda (Romania) (31.07.2013)
Reply all
Reply to author
Forward
0 new messages