Hi Colin, Dependent and doubly dependent localities are technical terms and without having access to PAF most mappers wouldn't know which one to use. And if they did, that could be considered a copyright infringement. Also, it just doesn't sound right. No one asks "which dependent locality do you live in". I agree it matches PAF very well, though.
I agree towns and villages are less precise but since we already have them as admin levels that would be the easiest and most intuitive solution. We already have addr:suburb and addr:hamlet so that would be a natural extension, and one which is already in use in the UK. I agree with you on addr:parentstreet. The issue here is that house numbers and names are associated with either addr:street or addr:place. So if we were to introduce addr:substreet or addr:campus that convention would have to be changed. For this reason I suggested using addr:place as a dependent thoroughfare.This would only require allowing both street and place to be defined together. Best regards, Andrzej On 28 January 2019 07:05:40 GMT+08:00, Colin Smale <colin.sm...@xs4all.nl> wrote: >Hi Andrezej, > >I would oppose addr:village for the Dependent Locality as it invites >incorrect usage. There is no reason to overload an existing tag with a >different meaning to its current usage. In the UK, a village is not >simply a neat subdivision of a town. I think addr:locality and >addr:sublocality would be better, as this would (correctly) imply a >possibly fuzzy boundary which possibly crosses formal admin boundaries. > > >Regarding streets/thoroughfares, the main thoroughfare is addr:street - >that is clear and established usage. We are looking for a solution for >a >"substreet" and moving the main thoroughfare up to "addr:parentstreet" >to make room for the dependent thoroughfare in "addr:street" feels >wrong >as it gives addr:street different semantics under some conditions. Note >also that the word "thoroughfare" has probably been carefully chosen to >allow application to things other than simple streets with adresses >neatly on each side. I would also instinctively expect a campus to be >more of a locality (subarea of a Town) than a super-street. Maybe >someone with access to PAF data can see what data is in what fields for >some address on the CSP. > >It wouldn't surprise me if subbuildings were used for "Unit 1", >"Building A" etc. That doesn't sound/feel at all unreasonable. > >On 2019-01-27 23:27, Andrzej wrote: > >> Hi Colin, >> >> This is broadly in line with Robert's proposals. However, it raises >questions about: >> >> 1. tagging "dependent localities" - they can be towns or villages. >Are you happy with addr:town, addr:village for this purpose? Reaching >consensus on that would be a major step forward. >> >> 2. Tagging "dependent throughfares". I think they could be used to >tag "building name, Cambridge Science Park, Milton Road, Cambridge". >This could be addr:place except in OSM addr:place should not be >combined with addr:street. Or, like in Robert's proposal, >addr:street+addr:parentstreet. Except that CSP is a campus, not a >street. >> >> 3. Tagging subbuildings. Addr:unit is available but is fairly limited >(unit names?) and vague. >> >> 4. PO Box - I haven't thought about it. Is that something that we >would include at all in a geographical database? Perhaps if it is >associated with a business that has a known location but uses PO Box as >its address? >> >> Best wishes, >> Andrzej >> >> On 28 January 2019 05:21:36 GMT+08:00, Colin Smale ><colin.sm...@xs4all.nl> wrote: >> >> Assuming the post code is seen in OSM as a way of addressing post (as >opposed to a geographic subdivision or an indication of location) then >I suggest following Royal Mail's address structure, which can be seen >in the description of the Postcode Address File on Wikipedia [1]. If we >cannot map a full-format address onto OSM tags, we need a description >of how to deal with this (i.e. which bits to leave out or combine). >> >> I have taken the table from wikipedia and added a column for the OSM >tags where known. Most of these fields are actually optional, or not >always present, depending on the exact address in question. >> >> How do we fill in the blanks? >> >> ELEMENT >> FIELD NAME >> DESCRIPTION >> MAX LENGTH >> OSM >> >> Organisation >> Organisation Name >> >> 60 >> n/a >> >> Department Name >> >> 60 >> n/a >> >> Premises >> Sub Building Name >> >> 30 >> >> Building Name >> >> 50 >> addr:housename >> >> Building Number >> >> 4 >> addr:housenumber >> >> Thoroughfare >> Dependent Thoroughfare Name >> >> 60 >> >> Dependent Thoroughfare Descriptor >> >> 20 >> >> Thoroughfare Name >> Street >> 60 >> addr:street >> >> Thoroughfare Descriptor >> >> 20 >> >> Locality >> Double Dependent Locality >> Small villages >> 35 >> >> Dependent Locality >> >> 35 >> >> Post town >> >> 30 >> addr:city >> >> Postcode >> Postcode >> >> 7 >> addr:postcode >> >> PO Box >> PO Box >> >> 6 >> >> [1] https://en.wikipedia.org/wiki/Postcode_Address_File >> >> On 2019-01-27 21:40, Andrzej wrote: >> >> Hi, >> >> When working on post codes in East Anglia I realised the current >address tagging scheme is insufficient for even fairly basic scenarios. >I have already discussed the issues with some of the most experienced >mappers and like to bring these issues to your attention. Robert has >summarised his ideas in >https://wiki.openstreetmap.org/wiki/User:Rjw62/UK_Address_Mapping >> >> The bottom line is, I would like to be tag commonly used addresses >without losing information and without resorting to addr:full. >> >> Issues: >> 1. Post towns (most pressing one because there is a lot of confusion >around it). The UK is fairly unique in that not every town is a post >town. This makes it impossible to tag e.g. Station Road, Histon, >Cambridge CB24 9LF. >> Wiki recommends addr:city to be used for tagging post towns >(Cambridge) but then how do we tag Histon? >> - Robert recommends sticking to the current meaning of addr:city and >using addr:town and addr:village for town and village names, which, >although not in wiki, are already being used in the UK. I like this >solution because it is very explicit in what each addr: key means and >it doesn't redefine addr:city. >> - SK53 prefers using addr:city for everything (towns, even villages) >and either not tagging post towns (they can be seen as a an internal >detail of a closed Royal Mail database) or using a new tag for it, like >addr:post_town. It is a simple solution, results in Histon being called >Histon and not Cambridge (without introducing new tags for town and >village names) and is commonly used. It is also a bit confusing (what >exactly is a city?) and I think we we should at least support tagging >post towns. >> >> Key questions: >> a) addr:city for post towns or towns and villages? >> b) how to rag remaining information (respectively, towns and villages >or post towns,) >> >> 2. Tagging addresses within campuses, business parks etc. There is >addr:place but it is supposed to be used instead of addr:street. Again, >Robert has a fairly decent proposal for that using addr:place or >addr:locality and addr:parentstreet. Please comment. >> >> 2a. should buildings in campuses be tagged with >addr:buildingnumber/name or addr:unit? I would prefer >buildingname/number (as they are often subdivided) but these seem to be >associated with addr:street. >> >> 3. Similar to (2) but for buildings. Tagging buildings that have e.g. >a single name but multiple house numbers? >> >> Best regards, >> ndrw6 >> >> _______________________________________________ >> Talk-GB mailing list >> Talk-GB@openstreetmap.org >> https://lists.openstreetmap.org/listinfo/talk-gb > >_______________________________________________ >Talk-GB mailing list >Talk-GB@openstreetmap.org >https://lists.openstreetmap.org/listinfo/talk-gb
_______________________________________________ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb