OpenStreetMap
Sometimes these are actually unnamed, or haven't been assigned one yet (maybe?); is there a way to distinguish these from the roads that we just don't know about yet? 'cos otherwise different people could end up making a point to visit them again and again, only to find out there is no listed name to see anywhere or they're too small to be noted as one.
People using the map might also think it doesn't look good to see roads without names (if they don't know its unnamed), or if you're on one of these roads and trying to figure out where the hell you are; you could spend ages looking for a name somewhere without realising its a fruitless search. If you knew it didnt list one, you'd then be able to hunt out another landmark earlier like postboxes with the ref. no (ok, just my personal thing!) or telephone box and its number (again...), or a pub/road crossing it/... You get the point.
So, has this been considered before?
Should I post on the wiki rather than suggest here?
name=unnamed/false/no/-1 doesn't work as programs won't know (although -1 isnt too bad, but its not consistent/following convention enough for me); unnamed=true might be ok.
PS: I asked in my immediate-last post about how to get a map of unnamed roads. I know you can search in josm for "label:highway -label:name" or similar; not sure how to convert that into tiles/a download for one of the progs I use on my phone (j2me, current faves are WhereAmI+ and GMJ, but Mom also possible). ie. in gpx or similar for waypoints (maybe start/end points? dunno how I'd do it).. Or tiles that I can hack into one of their db's.
Discussion
Comment from amapanda ᚛ᚐᚋᚐᚅᚇᚐ᚜ 🏳️⚧️ on 25 August 2008 at 10:41
You're not the first person to think of this. Steve Coast brought this up in July on the talk@openstreetmap.org mailing list. There were a few suggestions thrown around, "noname=yes", "name:absent=yes" "name=__none__", "unnamed=yes", "named=no", "nameless=yes".
Comment from Breezer on 25 August 2008 at 12:51
name=__none__ is very very pythonic!
Err... No! Unless we're suddenly becoming python developers of course :)
I like name:absent best, as it is then a 'subset' of the 'named' info and that seems to fit. Im hunting out the discussion to make these comments on...
Comment from Breezer on 25 August 2008 at 12:55
Further thought I wanted to put here...
The namespace way is also more happy with possible extensions like 'name:old' or :previous or :alternate if something's changed name recently but still referred to by old name a lot, and it means any renderer can understand that these other tags may need to be searchable or at least what they're to do with and can be displayed in association with, rather than needing to know about 'unnamed' or 'oldname' or other various (relatively) random tags that could be used. The more I think about this the more I like the namespace thing!
Comment from LivingWithDragons on 25 August 2008 at 14:51
Yep, stay away from just using name=something because that appears ugly on the map!
At a London no-name mapping marathons we decided if you add a lot of detail (building outlines/names, telephones, postboxes) around the roads you've checked then it looks more like someone has been there. Generally there will only be one or two isolated no-named roads in real life so that can be accepted, unlike a big clump of unnamed roads which were probably traced from aerial photography.
In the mean time you could just drop "note=there really is no name" in for anyone looking at the data before going out.
Comment from IgnoredAmbience on 25 August 2008 at 16:53
(I'll just note that there's also been large arguments over calling 'namespaces' namespaces, since apparently although they appear like them, they aren't namespaces, again, refer to the mailing lists)
Comment from Breezer on 25 August 2008 at 17:53
I get that - namespaces are heirarchical and usually if you alter a sub-domain of it, it knows to automatically affect the others and link them; whereas it actually doesn't care what the tag name is so using the colon is not a specific feature as such, its just a human-added abstraction layer we've made a convention.
Comment from BlueMM on 26 August 2008 at 01:15
BTW, the new HTML 5 draft spec is planning on using![{}]()
for truly empty field. Though I think some kind of "namespaced" solution would be better than {}. Hopefully can find something a bit more generic than :absent
Comment from smsm1 on 26 August 2008 at 09:50
Do you know about the no names map? http://dev.openstreetmap.org/~random/no-names/
Comment from Breezer on 26 August 2008 at 12:54
Aha! Thankyou very much smsm1! That's the one I was looking for!
but after a quick look - does it also mark main roads like A-classified dual carriageways as unnamed if they only have the ref (eg "A12" but no name for the individual sections that may well be named too, although I know some bits of it are "Ipswich Road" everybody always just says "A12"). If it doesn't it would be slightly nicer to mark them as less bright red 'cos the names aren't so essential as often the number is all that's used to refer to them. But that's being picky :)
'Namespace' (well if not that what else do we call the colon-ifying (lol) that people will understand) suggestion for the name stuff (and would work for other things too): how about name:status=old or =absent or =deprecated or whatever for the different states? We could standardise a selection of values. Do we have a standard for alternative names/nicknames (aliases) for a road? They're usually very local if there are any. I know the different languages are name:=....