[Talk-GB] Data Model for Address
colin.smale at xs4all.nl
Sat May 30 08:43:00 UTC 2015
...or we identify use cases we want to handle (and thereby take a
conscious decision to "de-scope" other use cases), and ensure that our
model is fit for (our) purpose. Then map the "official" data model onto
the "internal" model. This mapping will probably be a one-way mapping,
i.e. it is unlikely to be possible to go from the (simplified) OSM model
back to the standard model in a reliable way.
I found this paper (2003) which shows some more of the vagaries of the
system. There is a short section on BS7666. It's a few years old now so
things have moved on a bit, but I reckon a significant part still holds
On 2015-05-30 10:00, SK53 wrote:
> If the world is complicated it may be necessary to reflect that in the data model (or alternatively make the data model so general that complications are pushed to applications). I think the Royal Mail approach to addresses is already too Procrustean!
> However, there is a British Standard for addresses, which I'm told by those who know it well provides a high degree of flexibility for the numerous cases which do not reflect housenumber, supplement, street, place model.
> On 29 May 2015 at 15:26, Colin Smale <colin.smale at xs4all.nl> wrote:
>> If anyone is interested in the data model used by Royal Mail in UK addresses, this will tell you loads:
>> http://www.poweredbypaf.com/wp-content/uploads/2015/02/Latest-Programmers_guide_Edition-7-Version-6.pdf 
>> Warning: you may find yourself uttering things in "rather unparliamentary language" when you read this.
>> Contrast this with the Dutch address model: Street, Housenumber (numeric), Housenumber Extension (alphanumeric), Postcode (9999XX) and Town. Probably much the same in Belgium and Germany too.
>> Talk-GB mailing list
>> Talk-GB at openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk-gb 
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Talk-GB