[OSM-dev] Extracting Great Britain and Ireland from planet.osm
osm at randomjunk.co.uk
Tue Mar 13 12:14:19 GMT 2007
On 13/03/07, Ludwig M Brinckmann <ludwigbrinckmann at gmail.com> wrote:
> I think you misunderstood me. I perfectly understand the reasons why the
> API gives the 400 error.
> What I thought was 'insane' however was the attitude that my desire to
> 'download the whole of London' was somehow not justified rather than just
> pointing out a temporary shortcoming of the software. Just because it did
> not work and there was no trivial workaround (apart from download the whole
> planet, install ruby, install another DB, configure a few scripts...) did
> not mean that my request was somehow outlandish. Remember, 'the customer is
> always right' ;-)
> But with these scripts I think we will get a very good workaround that
> will reduce workload on the server and give people what they want: easy
> access to open data. Best of both worlds really. Over and out.
Hehe... I download the whole of London every week (
I just tile the requests into about 300 queries run over the period of a few
hours. That runs to about 60MB these days which is a significant sized
downoad from any web server... even without the backend database activity
that entails. Doing something with that, like rendering it with osmarender,
takes the better part of 2 hours and 2GB of RAM.
So TBH, custom queries of that size are a little crazy. But yeah, a big dump
of areas people might like neatly packaged would be cool. Preferrably in
some way that meant you didn't need to buy (or steal) so much ram that you
could actually run windows vista on the machine (obviously not at the same
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the dev