[OSM-dev] GPX in the database

Tom Hughes tom at compton.nu
Tue Sep 25 21:36:42 BST 2007


In message <ad40d23d0709251225p1a0e9c48v6217fdaad9fbb8de at mail.gmail.com>
          "Ted Mielczarek" <ted.mielczarek at gmail.com> wrote:

> I'm new to OSM, but I've been poking around the mailing lists and such
> lately.  I noticed some conversations about the GPX storage in the
> database this month.  Maybe this is a silly question, but *why* are
> you storing all the points from each uploaded GPX?  Wouldn't it make
> more sense to just store a centroid or bounding box per-GPX, and then
> make the original file available via HTTP for clients to retrieve?
> It's not like GPX is terribly complicated to process.  As far as I can
> see, the only use of the GPX data is for tracing, etc, so it seems
> like requiring the clients to process GPX files themselves would be an
> easy tradeoff to relieve a lot of load from the database.

One reason is that you may wind up downloading far more data
if you download the whole of every trace that overlaps the
bounding box you wanted - some of the traces we have are truly
enormous.

Another reason is anonymity - about 10% of the traces we have
are marked as private which means you can only get the points
from them and can't tell what order they are in, who they came
from or the timestamps associated with the points.

Tom

-- 
Tom Hughes (tom at compton.nu)
http://www.compton.nu/




More information about the dev mailing list