[OSM-talk] On how to store geospatial data !?; Was : Server slowness
Martin.Spott at mgras.net
Sun Jan 14 16:23:22 GMT 2007
Iv??n S??nchez Ortega wrote:
> Why doesn't OSM use MySQL's spatial extensions, then? If the 'latitude'
> and 'altitude' columns are merged into a single column of type 'Point',
I know, the subject sounds a bit provocative, but I don't intend to
hurt anyone. Please read on, of you like.
A bit of a background: I'm running a respository of "Open Geospatial
Data" for approximatley two years now and the software I'm using for
storage has been proven to be reliable and capable of handling large
datasets, even under load.
I've started by focussing on data that's being used for being draped
over the terrain for the FlightGear flight simulator. Since the early
start, the repository has been enlarged significantly and has been
turned into a multi-purpose storage of different datasets at one
central location. Joins between different datasets are being
implemented (over the time) wherever it is possible and makes sense. A
short overview is presented here:
It's obvious that the primary focus of this repository is different
from the one of OpenStreetMap. The intended use of my repository is
visual display and easy retrieval of large datasets with global
coverage while OSM apparently - please correct me if I'm wrong -
primarily aims at accurate routing in their road network and building
nice road maps. Another significant difference is that I've been
building my repository by using _existing_ data and later started to
prepare for manual adjustment - instead of adding almost everything by
manual work. As a result, apparently I have much more coverage in my
repository while OSM definitely has the nicer tools to work on the
data. Still I see some intersections between these two projects. This
one for example:
>From reading the -talk and -devel mailing lists I realized that OSM has
started importing external data like shorelines into their repository,
using their own home-grown tools. Our (my) FlightGear/OSGeo repository
already contains highly accurate shorelines almost right from the
beginning and importing these into the DB was no big deal.
I'd wish to reduce duplicate effort among two OpenSource projects that
head in the same direction and therefore I'd consider it very helpful
if exchange of datasets would be made easier as it is currently.
Would OSM folks consider it useful to discuss cooperation/collaboration
between these two projects ?
Unix _IS_ user friendly - it's just selective about who its friends are !
More information about the talk