Re the roads in Orkneys and Shetlands (whatever specific roads you are referring to here) - if they carry less traffic than other local roads, then yes, I’d argue for downgrading them.
By unreconcilable I meant official, legal status of a road does not reconcile to its OSM tagging in a 100% of instances, which is why those guidelines need to have the disclaimer about OSM rules taking precedence. As a baseline, sure, it makes sense - but let’s not bend reality to fit this.
It wasn’t my intention to speak of anyone or anything negatively, and I wasn’t even referring to A420 specifically with my comment.
To me, those are 2 quite similar concepts - which is why we have the general OSM highway tagging guidelines as they are.
Of course, one could argue this argument is not needed, since routers (ignoring other data consumers) can just take access, surface, lanes, maxspeed etc to assign road weighting
It’s pretty much 99.99% given there’s only two examples of this in the entire country though, so it makes little sense to have these exceptions, and even less sense to make more exceptions!
There is also no “bending of reality” - the reality is this is the A420, classed as a non-primary A road. It’s bending the reality to therefore class it as a lower status.
What takes precedence specifically on routers do you know? Access tags, highway tags, anything else?
Every OSM community that I can think of that has tried to match their “official” road classification with OSM highway values has hit edge cases in which (essentially) the “official” classification doesn’t match reality. All the ones I can think of have gone with “reality” over “officialdom” when picking a value for OSM (often after some convoluted discussion with that country’s equivalent of the “official” faction here).
If someone wants to create a map of “official” road status based on OSM data they have been able to do so since 2015 - that was when a designation=primary tag was added to this road.
8 Likes
reflecting the legal status of the road (highway=primary).
highway=primary is not about the legal status of a road, it is the result of the assessment by mappers. The ref tells you that the road is an A-road
3 Likes
Primary and non-primary status are determined by highway tags in addition
1 Like
Yet there are only two instances here that I still fail to see why they should be tagged otherwise.
Thanks for mentioning the on-the-ground rule. I don’t think anyone in this thread is actually from the Oxford area apart from @Andrew_Chadwick and me (I live a short way away and my kid goes to school 250m from the High Street), so here’s some on-the-ground evidence.
There are no A420 signs in Oxford between Headington Roundabout and Seacourt, on opposite sides of the ring road. The supposed A420 is unsignposted at even majorish junctions like that with Headley Way, or the Plain, or Frideswide Square. A recent article in Oxford’s indie newspaper even called it “Oxford’s hidden A road”.
The High is as much the A420 as some unsigned backwater country lane is the C7653. Yes, it might possibly be noted as such on some planner’s card index (or, more likely, ArcGIS install). But that’s an internal-facing reference and we don’t map that in the ref= tag, let alone allow it to dictate the highway classification.
I am a little amused by your reference to “that’s what the planners intended”. I presume you don’t know any Oxfordshire County Council highway planners? I passingly know a few, as well as a couple of current OCC cabinet members. The notion that they have consciously decided to class the High as an A road is really not what has happened. [1] It’s a historical artefact, a leftover. (It’s also one that happily comes with some funding advantages, so they’re not going to be in a hurry to declassify it!)
(If OCC Transport Policy cared about road numbers they wouldn’t have left 1.5mi of the southern bypass as a randomly isolated stretch of the A423, for starters.) ↩︎
5 Likes
Yes, mapping a nominal A road as highway=tertiary is an exception.
Mapping a road that is closed to cars for most of the day as highway=primary would also be an exception. I don’t know of any other such cases in the UK - maybe there are one or two.
So which exception do we choose? The one that satisfies roadgeeks’ sense of tidiness, or the one that makes for a more useful map for people trying to navigate through Oxford and which has been the solution adopted by the local community for 17 years now? I think the answer is pretty obvious.
7 Likes
To be clear, I added a section to document the usage being discussed. Which, I’ll note, I have not been involved in.
Did you actually check this before posting?
highway=secondary|tertiary|unclassified & ref=A*: 212 ways
highway=tertiary|unclassified & ref=B*: 184 ways
Unfortunately my OT skills aren’t quite up to scratch, so these aren’t unique (i.e., multiple ways of the same road are included) and, sure, there may be some tagging mistakes. But there are certainly far more than just two roads in the UK where mappers have decided the OSM classification does not match the legal designation.
1 Like
according to taginfo geofabrik, there are 155137 highway=primary in “Britain and Ireland”,, so it is 99,86%, or looking at Great Britain, 143653 primaries (99,85%).
Mapping as highway=primary is the exception? Really? Not like hundreds of other restricted roads, dead-end roads, narrow roads, low traffic roads?
I feel mapping as primary is the obvious solution, not the random exception which breaks the pattern of almost every other classified road and therefore makes for a worse map.
I was close enough then!
1 Like
I don’t think anyone here disagrees that in the vast majority of cases, OSM classification and legal definition will match (especially in the UK considering the highway classifications were built on the UK definitions).
However, in the real world, there are a non-zero number of cases where these two don’t match. The actual point being made is: this isn’t just one or two cases as has been stated. There are (likely) hundreds many.
Using pure percentages like that is a case of statistical bias. It is comparing against the wrong population group. We aren’t interested in all highway=primary.
We would need to compare tagging for the types of road in question, i.e., roads where the legal classification does not match its use type.
But this is impractical to do in OSM (since we can’t find those roads on tags alone) and is also biased by the fact that the documentation has only recently been updated to be less rigid.
What I’ve shown is that you were incorrect in assuming there were just two cases. There are far more.
2 Likes
to me it seems it is in the order of tens (because every road will likely be fragmented due to varying speed limits and other changing properties). I would suspect that some A-roads currently tagged as primaries could also be downgraded in OSM, and are only tagged as primaries because the mapper assumed a 1:1 relation.
What I totally agree with: even if the vast majority of A-roads are (and should be) considered primary roads in OSM, it doesn’t mean the few outliers that are not should be retagged to primary just because of this, not at all. Every road must be individually looked at.
2 Likes
I am willing to bet that a lot of those other ways are mistagged sections of road such as link roads across dual carriageways, instead of full roads.
Which opens up more subjective edits and makes the map less consistent.
I didn’t say anything about low traffic, narrow, dead-end roads etc., because we’re talking about Oxford High Street and not, I dunno, some A road on Skye or something. Can you show me many other continuous highway=primary ways in the UK which have a permanent, most-of-the-day point closure to cars?
Which opens up more subjective edits and makes the map less consistent.
if one mapper does it, it’s subjective, if many agree it’s common ground
7 Likes
I did not say there are many other ways that have the exact restriction as this, and not having the capabilities to explore every single possible tag, I can’t say exactly how many.