[Talk-us] [Talk-us-newyork] Highway classification guidelines for New York State

[Jmapb]

On 9/11/2021 7:48 PM, Minh Nguyen wrote:
> It seems you're referring to this line:
>
>> Streets in urban areas that serve commercial or industrial districts
>> and do not have a higher classification should also receive
>> highway=unclassified.
>
> This is reminiscent of the rule of thumb that gets tossed about often,
> something about downgrading highway=unclassified to
> highway=residential once there are a certain number of houses per mile
> -- specifically houses. The wiki has a much more convoluted definition
> for both highway=unclassified and highway=residential, so I don't know
> where that simpler formulation comes from; it might just be urban
> lore. Personally, I've never been particularly keen about classifying
> roads based on something that could just as well be captured in the
> abutters=* key. You may be interested in a separate proposal for
> highway=street that would do away with any distinction between
> unclassified and residential roads. [1]

Thanks, that proposal is interesting. Personally I'd prefer
highway=local because "street" implies an urban or suburban context. And
although the name is terrible, I don't think there's any need to get rid
of the concept of highway=unclassified. I'd've prefered
highway=quaternary though, and if I had a time machine I'd do some
shoe-banging to that effect (after a quick trip to the 70s to copy all
the lost Doctor Who episodes.)

As far as downgrading highway=unclassified to highway=residential when
enough houses are built, that seems to miss the point a bit. IMO, as
long as significant traffic is using that road to go from one locality
to another, it should be somewhere on the
trunk/primary/secondary/tertiary/unclassified spectrum.

(I just posted a message on the Brian & Eric thread on this topic as well.)


> access_control=* refers to the degree to which traffic is able to
> enter from side streets and driveways, as in "controlled-access
> highway" or "limited-access highway". If I'm not mistaken, it was
> originally coined by NE2, a former regular on this list, to indicate a
> super-two. [2] More recently, it has come up as a more straightforward
> alternative to expressway=*, though the two keys indicate somewhat
> orthogonal concepts. [3][4] At this point, with over 10,000 uses,
> access_control would qualify as "in use" once someone gets around to
> writing a wiki page for it.

I eagerly await. (Perhaps worth mentioning that NE2 was notorious for
unilateral mass tagging and ultimately banned for uncooperative behavior.)


>>  > Memorial and symbolic highway names
>>
>> Both honorary_name and memorial_name are in use and seem more correct
>> for this purpose, so why is official_name preferable? Personally I like
>> honorary_name because it would cover non-memorial situations, eg, where
>> a living person's name is used. (Currently memorial_name has more uses,
>> but they're all the same road, "USS Indianapolis Memorial Highway".)
>
> Is there an example where a road would have both a memorial name (the
> term used by the MUTCD) and some other kind of official name, where
> the distinction is important enough that we wouldn't want to just dump
> the two names in official_name, separated by a semicolon? I'm aware
> that a road may have multiple official names of various kinds, but the
> distinction may be so minor as to create headaches for mappers in
> other contexts, similar to how we've had to split hairs between
> unclassified versus residential.
>
> Many data consumers handle name keys in the same manner regardless of
> feature tag. Despite being used orders of magnitude more often,
> official_name is also relatively poorly supported by data consumers.
> Are there things other than roads that also need to maintain this
> distinction between memorial and official names? It would be
> unfortunate if, in a quest for precision, we relegate these names to a
> key that would go mostly unconsumed.

Official_name, as I understand it, was invented to record wordy official
versions of country names whose shortened names are preferred in common
use. "United Kingdom" versus "United Kingdom of Great Britain and
Northern Ireland", for example. So ideally the same subset relationship
between name=* and official_name=* would hold on smaller scales as well.

Offhand, for highways in New York, I could imagine "name=Palisades
Parkway" and "official_name=Palisades Interstate Parkway". It's not
tagged like that, but I'd understand if it were.

As far as I know there aren't any honorary names on the Palisades
Parkway, and I don't know of any uncontrived examples of conflict for
highway names. But if OSM has taught me anything, it's that the world is
big enough that corner cases should be expected.

I'm not dying anywhere near this hill though. If the NY community wants
to shoehorn honorary names into official_name=*, I'll tag along.

Cheers, J