[Talk-GB] Open data (Was: Parliamentary debate mentions OSM)

[John Aldridge]

On 29-Mar-16 10:19, Rob Nickerson wrote:
> Should we attempt to include everything that is in the open data
> datasets plus our on the ground additions (manually or, unless we
> suddenly get many more mappers, by some form of controlled merge) or
> should we leave the end users with the task of mixing OSM with the open
> data?

I'm generally not in favour of importing 'definitive' data from other 
sources, and it would be better to have some dynamic overlay procedure 
(which must of course available to normal end-users of the map on 
www.openstreetmap.org, not just to sophisticated OSM data processors).

An example is parish boundaries which, I understand, have been imported 
from Ordnance Survey data. The problem with these are that they often 
get inadvertently corrupted in OSM: they tend to lie along other 
features, which means that it's rather easy to get them inadvertently to 
share nodes, which in turn ends up with them being dragged around by 
mistake.

I appreciate that we don't have such a dynamic overlay procedure, and 
nor do I have a solution to offer, but the problem is real: if we import 
data which is definitively specified elsewhere, we are pretty much 
guaranteeing that OSM's version of that data will be inferior to other 
mapping, even if only because it'll get out of date.

I suppose a satisfactory alternative to dynamic overlay would be if any 
such import were required to have an automated procedure for adopting 
regular updates from the definitive source.

-- 
Cheers,
John