[OSM-talk] Measuring success of OSM

[Gregory Williams]

> -----Original Message-----
> From: talk-bounces at openstreetmap.org [mailto:talk-
> bounces at openstreetmap.org] On Behalf Of Brendan Morley
> Sent: 12 December 2009 14:03
> To: Talk OSM
> Subject: [OSM-talk] Measuring success of OSM
> 
> In the interests of healthy debate,
> 
> I disagree with some of the sentiments around imports.  In particular
> that imports reduce the amount of OSM contributors, and that that is a
> Bad Thing.
> 
> IMHO success should be measured by the accuracy of the data (to
> reality) and to the pervasiveness of its use by people and corporates.
> What if suddenly
> Ordnance Survey merged all their excrutiating detail of data under OSM
> terms?  Is that a good thing or a bad thing?  I say good, because there
> will always be
> need for tracking of changes.

What makes you think that the OS will track changes better than people on
the ground in the area? As I understand it OS primarily is notified of
changes by local authorities etc. combined with the odd aerial survey. I've
surveyed many a feature near me before they're even officially open, whereas
the same features still haven't appeared in the OS data that I can access.

There are lots of things that OS don't cover too. Cycle parking, the current
occupants of shops, etc. We still need people on the ground to collect and
maintain that data.

If there's largescale importing of data not collected by somebody in OSM
then it's unlikely that somebody is going to feel the same sense of
ownership of it. Also in those areas where the data has essentially
exclusively come from an import instead of our own survey it's less likely
that we'll attract new contributors in the first place. Surveying highways
is much better for getting somebody hooked on OSM that surveying POIs.
Adding highways has an immediately obvious impact. To have the same level of
impact with POIs will need at least an order more work.
> 
> Perhaps OSM would turn into a "heads up" service for government - you
> know, a local would identify a tree planted "somewhere around this GPS
> reading" and
> then an OS employee would know to come around and get a more accurate
> fix.  I'm only speculating here, I don't know how OS actually detects
> changes to its
> dataset.
> 
> If OSM ever reached that level of detail, then it'd be pointless for
> Google to "steal" our data, they'd always be playing catchup.

I've already pointed out that OSM is richer than OS in some domains.

Oh, and it'd be really easy for Google to have a near-real-time mirror of
OSM's data with updates rendered shortly behind. Lots of people already
replicate from OSM, and we've shown that it only takes extremely modest
hardware in Google's terms to keep up with the rerendering necessary.

Gregory