richard at systemed.net
Tue Jan 18 15:24:26 GMT 2011
Chris Moss wrote:
> Richard's point that OSM is a data project rather than an end user
> map website puzzles me. I can understand that in server terms
> there are huge demands but isn't one of the ways of coping with
> that by being selective about the data that's given out.
We're a data project - not even a data distribution project.
OSM's main output is here: http://planet.openstreetmap.org/ . That's the raw
data for the whole world.
Others can, and do, make excerpts of this available. Many people, for
example, work with the excerpts provided by Cloudmade and Geofabrik. At
present these are regional excerpts - usually country-specific, so you can
download (say) the UK alone. But there's no reason why thematic excerpts
couldn't be made available if someone saw the need to provide them.
To provide targeted data distribution services on osm.org itself, over and
above simple planet downloads, would be computationally expensive -
especially as it would only encourage people to use osm.org for their
downloads rather than third-party excerpt mirrors. You only have to look at
the on-going problems of XAPI, a non-core selective download API, to see the
challenges. Far better that people download from well-funded OSM supporters
like Cloudmade and Geofabrik, rather than placing additional funding demands
on OSM itself.
> As I understand it, all the data - supermarkets, bus routes, power
> grids etc. is being uploaded to OSM and is also downloaded at
> some level of detail - except that it's virtually useless because it
> can't be distinguished from the other stuff. Or do I misunderstand
> what's going on?
Yes, you do.
You download the data (as a planet file or regional extract) and then you
extract the bits you want from it.
This is generally done by filtering the data on tags. To take a very simple
example, if you were interested in waterways but not roads, you would
extract all the data from your planet file that contained a "waterway=*"
tag, but discard any that contained a "highway=*" tag.
Most people use a program called Osmosis for this, but you can also write
your own fairly easily if you have any programming expertise.
> When I look at openstreetmap I get a choice of 4 base layers: Mapnik,
> Osmarender, Cycle Map and NoName. The first two appear to be
> rendering engines, and I can find NoName in the wiki, but what is
> Cycle Map and why is it there?
> Please don't think I'm denigrating cyclists - I'm one myself. But
> shouldn't there be an opportunity for 100 such layers?
There is - if you want to create the 100. See
but I think you might be getting ahead of yourself here. :)
View this message in context: http://gis.638310.n2.nabble.com/invisible-tp5933851p5935908.html
Sent from the Great Britain mailing list archive at Nabble.com.
More information about the Talk-GB