On Mon, Feb 9, 2009 at 1:45 PM, Feargal Hogan (E-mail) <
[email protected]> wrote:

>  Is there a case for examining the quality of the existing OSM boundary
> data and trying to improve it?
>

Yes.


>
> Is it worth/possible to create a project to do this, either within the
> current OSM framework or without?
>

Yes, either would be great.

>
> I'm thinking about something similar to www.mappa-mercia.org, which popped
> up on twitter this morning, but with the focus more on a particular dataset
> rather than a locale.
>
> There are a few possible data sources which might help improve the data,
> legitimately!
>
> 1. I wonder if the euro-election constituency boundaries might be available
> without any restrictive licences? It might be that the EU would have rights
> to such data which would override any OS imposed licencing issues.
>
>

Worth some investigation, although our experience is that the Ordnance
Survey's copyright is pretty inviolable


> 2. There are lots of physical features from which boundaries are derived -
> rivers, roads, etc. Surely if these were mapped as geo-features the data
> could be copied to a boundary layer where it was known that the boundary
> followed the feature, rather than the other way round.
>
>
Yes, where the boundaries follow a road or river it is relatively easy.



> 3. There are lots of "Welcome to Surrey" type signposts around. Mapping
> their locations would start to give us some join-the-dots style boundaries.
>
>
Definitely worthwhile.  I've derived a lot of boundary points from signs
like this.  Often the sign is a few metres away from the actual boundary for
convenience/safety reasons.  But you can often guess that, for example, if
there's a sign just before a bridge then the river is the actual boundary.



> 4. Lastly, if I drew some crude "from-memory" style boundaries at a very
> low zoom level, they could be amended and improved whenever anyone noticed
> they were wrong. Would that breach anyone's licencing?
>
>
If you have a photographic memory and used to work for the Ordnance Survey
then it might, but otherwise it is reasonable that a person has a working
knowledge of their locality.  Crude data is much preferable to no data and
stepwise refinement is one of OSM's guiding principles.



> Just a thought.
>
> Feargal
>
>
>
> OSM does have some administrative boundaries but they are one of the harder
> things to derive.
>
> Some county boundaries can be obtained from old out-of-copyright maps, but
> these can and do change over time so it's not an infallable method.
>
> Borough boundaries have been derived, by some people, in urban areas from
> on-the-ground inspection of street furniture (street lights, rubbish bins,
> manhole covers, whatever) and then extrapolating from there.
>
> Ward boundaries are very difficult, I don't know of any general method for
> capturing these.
>
> So, OSM does contain some boundary data, but its not very complete and may
> not be very accurate.
>
>
> _______________________________________________
> Mailing list [email protected]
> Archive, settings, or unsubscribe:
> https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public
>
_______________________________________________
Mailing list [email protected]
Archive, settings, or unsubscribe:
https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public

Reply via email to