On 2020-01-07 21:14, Paul Allen wrote:

> On Tue, 7 Jan 2020 at 19:42, Colin Smale <colin.sm...@xs4all.nl> wrote: 
>> I'm glad you said "probably", because it is of course not always true. And 
>> these edge cases are what we need to accommodate. Limiting the discussion to 
>> just handling the easy cases is cheating.
> I know it's not true because I've had to deal with some of these oddities. 
> Sometimes we can have a scheme that handles oddities in its stride without 
> imposing unnecessary difficulties on the normal cases.  I haven't seen 
> anybody 
> suggest anything like that (yet) for addresses.  But I still think we should 
> optimize 
> for the common case and not optimize for the abnormal case.  Make the easy 
> things 
> easy and the hard things possible rather than make everything hard.

Optimisation should be phase two. First define a model that works, then
optimise. Ignoring the edge cases just delays the pain. We are long past
the point of handling the simple cases. 

>> Bit of a philosophical question: What is an address? In the UK, the Post 
>> Town and Postcode are for the purposes of delivering mail. If they happen to 
>> be useful to other parties, that's great, but it is only a side-effect.
> Post town is actually the opposite of useful.  People put the post town in 
> their address 
> rather than their nearest named locality which makes it hard to find them 
> when looking 
> at a printed map.  Actual nearest locality is far more useful whether looking 
> at a 
> printed map or making a nominatim query.  Post Town is no longer necessary 
> even for 
> delivering mail, it's just a historic artefact that serves no useful purpose 
> any more.

Royal Mail do not say the Post Town is optional. RM also know of
localities and dependent localities, which may or may not bear any
resemblance to an inhabitant's perception of where they live. 


>> Administrative boundaries are not relevant in UK addressing, unlike many 
>> European countries (I know about NL, DE, BE, FR) where "places" have defined 
>> boundaries.
> Administrative boundaries are not usually relevant but are often given and 
> often 
> required when filling in forms.  They sometimes are relevant; there are 
> several 
> localities called Tarbert (sounds like a Dilbert character) in Scotland and 
> without a 
> postcode you need a county to figure out which one is which.  There are other 
> places 
> in the UK where the county is needed to disambiguate, and even some where you 
> need 
> more than just the county.

Postal counties have been deprecated for years, but are still in many
people's minds. Metropolitan counties are no longer "administrative".
Traditional counties maybe? But almost certainly not administrative

>> The relationship between buildings and postcodes is N:M. If we replace the 
>> word "building" with "premises" and saying that an address refers to a 
>> "premises" may get us a bit closer, given that a "premises" may consist of 
>> part of a building, a  whole building, multiple buildings or any combination 
>> thereof.
> I simplified, a little.  For anything that has a postal address in the UK, 
> the building(s) 
> number or name, plus the postcode, uniquely identifies it for the purposes of 
> postal 
> deliveries.  But the other stuff can be useful for other purposes.

So if I am now more explicit about my intention to help this discussion
towards a conclusion..... What an address refers to, is different in the
UK compared to other countries. We will never find a single model to fit
the whole world that is not abstracted to the point that it becomes
useless. Let's stop chasing our tails, and accept that. 

Back to the philosophical question: Is a normal "address" in OSM: a) for
delivering letters, or b) for navigation, or c) an identifier of a
building/premises, or d) something else? Should/could we cater for these
different definitions of "address", e.g. by having tags like
addr:{address_type}:{address_element}? This is a question that IMHO is
probably best addressed at global level; then let each country have its
own model within that framework.
Tagging mailing list

Reply via email to