[OSM-dev] Problem with Postgresql / Mapnik / SRTM Contours

[sylvain letuffe]

> No, I'm sorry, 
Then all the excuses have to come from me for being sarcastic, I miss read 
your comment, sorry for that.

> You can
> claim 1 x Beverage of your choice the next time we meet as a reward
> for being eagle-eyed!
The beverage would then be for you for such a very good integration of SRTM 
data, my merit just come from using your tool, wich in turns goes back to 
YOUR merit doing it

> If anyone else wants a similar prize, there's another one for finding
> the most ridiculously and obviously incorrect void filling. 
I'll do my best, but still, it might well be better and less ugly than 
nothing. And trying it, with the time you might have spent for it, is 
valuable data.... for me spending time to try something else ;-)

> No, I don't use it. A few of us have spent time examining google
> terrain layer, and I'm pretty sure it comes from SRTM
How could that be ? google GE's 3D earth model, is, at least in france and 
even more in switzerland close to "perfect". Well, "perfect" in the sense 
that there are no artifacts beneth the simplified terrain model that makes 
cliff ugly and a ~400 meters shit to the S-East, but I can't imagine any 
algorithm that would leads from what we had on the opencyclemap contours to 
the result in GE by just interpolation. 
They just can't have constructed data that "was not" but I'll have a closer 
look to that by places in the mountain I more than perfectly know.


-- 
Sylvain Letuffe liste at letuffe.org
qui suis-je : http://slyserv.dyndns.org