[Talk-ca] Sinking islands...

2010-09-23 Thread G. Michael Carter
I was importing data in the Georgian Bay area and I noticed something. If an island is natural=land it renders in mapnik as white, regardless of the outer multipolygon. if you have natrual=wood or possibly others, it sinks if it's not role=inner. This is at least what I've observed so

Re: [Talk-ca] Sinking islands...

2010-09-23 Thread Sam Vekemans
try using the Bing maps version of osm, i think it renders faster. Its a handy way to check, i also use the Marble Desktop, as it renders pritty fast also. cheers, sam On 9/23/10, G. Michael Carter mi...@carterfamily.ca wrote: I was importing data in the Georgian Bay area and I noticed

Re: [Talk-ca] Ottawa - OpenData - Map Feature Catelogue

2010-09-23 Thread john whelan
I'm not certain what the point would be if the licensing doesn't line up. What would be nice would be if some one could take one or two of the data sets, flip it into OSM format and email it to me. That way I can merge it with a copy of the cities OSM data on a USB stick and show a couple of

Re: [Talk-ca] Sinking islands...

2010-09-23 Thread G. Michael Carter
In the cases where we have all three and they are identical. Seems better to have the three merged. In any case good to know. I'll start removing them on import. On 23/09/10 10:00 AM, Bgin, Daniel wrote: Bonjour Michael, 2010-08-16 11:54 [Talk-ca]

Re: [Talk-ca] Sinking islands...

2010-09-23 Thread G. Michael Carter
I've been taking all three objects and merging them. So the inner is natural=wood, with name. http://www.openstreetmap.org/?lat=45.45968lon=-80.42647zoom=16layers=M http://www.openstreetmap.org/?lat=45.45968lon=-80.42647zoom=16layers=M Seems to be rendering ok. Childs and Osawa are

Re: [Talk-ca] Sinking islands...

2010-09-23 Thread Bégin , Daniel
Cool, So, what should be done for the next release? - Keep the natural=land areas - everybody will have to merge each island to get the proper rendering; - Transform the natural=land areas into points - no merging necessary, no name lost; - Remove the natural=land areas - name will be lost

Re: [Talk-ca] Missing islands and coastline

2010-09-23 Thread G. Michael Carter
The mix of natual=water and natural=coastline is because their dual objects. The natural=coastline is needed as the great lakes (as far as I know) is connected to the ocean. So deleting the coastline would delete portions of the Atlantic Ocean. What I'm doing is enclosing the Canadian

Re: [Talk-ca] Sinking islands...

2010-09-23 Thread G. Michael Carter
I'm fine with making the island name a point. It would show up in POI applications. On 23/09/10 11:28 AM, Bgin, Daniel wrote: Cool, So, what should be done for the next release? - Keep the natural=land

Re: [Talk-ca] Missing islands and coastline

2010-09-23 Thread Bégin , Daniel
Hi, From my understanding, natural=coastline is just a way to map large water areas. - The advantage of using this tag is that it can be composed by multiple ways without the need of a relation type=multipolygon or doesn't need that all ways be joined togeter to get a single closed way. -

Re: [Talk-ca] Sinking islands...

2010-09-23 Thread Lennard
On 23-9-2010 15:09, G. Michael Carter wrote: I was importing data in the Georgian Bay area and I noticed something. If an island is natural=land it renders in mapnik as white, regardless of the outer multipolygon. if you have natrual=wood or possibly others, it sinks if it's not role=inner.

[Talk-ca] natural=land

2010-09-23 Thread G. Michael Carter
Is there any case where we'd be using natural=land any more? I've been thinking of taking the osm data in my db and creating a list of common errors. Something like http://keepright.ipax.at http://keepright.ipax.at/ but as just a check list. One thing I'm going to scan for is identical

Re: [Talk-ca] Ottawa - OpenData - Map Feature Catelogue

2010-09-23 Thread Sam Vekemans
I'm asking as the data will 'eventually' be available, it already is ok to use for other OpenMaps projects, and it's available for Garmin Devices already. Just needs to be converted. ... and tags need to be matched up :) If knowone has seen, or gone through the catelogue to match up the features,