[OSM-talk] Free the postcode maps

[Christopher Schmidt]

On Tue, Jul 04, 2006 at 04:37:13PM +0100, Nick Black wrote:
> On 7/4/06, Barnett, Phillip <Phillip.Barnett at itn.co.uk> wrote:
> >
> >
> >
> >Looking at the list distributed by FTP, it's obviously a typo - whoever 
> >added it missed the 3 off 53. See it in context here
> >53.474080 -2.234340 M1 3BB
> >
> >50.478000 -2.240980 M1 4JY
> >53.475470 -2.242060 M1 5JW 53.477075 -2.240910 M1 6EDFree The Postcode is 
> >a nice idea, but how is it going to be validated, short of running every 
> >postcode past the Royal Mail database, which surely breaks the spirit of 
> >the idea?
> 
> Yes, and the legality too.  Validation will be possible by finding
> rogue postcodes.  For example, if a postcode starting with M is
> surrounded by postcodes starting with P it could be flagged as
> potentially incorrect, with a more complete dataset it should be
> possible to create tolerance levels and automatically exclude some
> postcodes.  This can be done by inference of the data collected and
> from local knowledge.  We can find out where Manchester is, where
> London is etc by looking at the ring road motorways in OSM.  Of
> course, this is not a perfect system, but I that as the dataset gets
> bigger it will be easier to validate it.   Just like OSM -  if you
> want centimeter accuracy, go to the Post Office - if you want it Free,
> get it here.

Just to be antaganistic -- wouldn't using OSM to determine locations
make it in part a derivative work of OSM, which means it can no longer
be public domain? 

Of course, it's just as easy to look at the NASA satellite imagery,
which is public domain, so I suppose there's no real concern with that,
just pointing out that deriving data from OSM probably won't work in
general.

-- 
Christopher Schmidt
Web Developer