[Talk-us] Update on potential highway classification reform

[Kevin Kenny]

On Tue, May 18, 2021 at 1:47 PM Russell Nelson <nelson at crynwr.com> wrote:

>
> On 5/17/21 5:54 PM, Brian M. Sperlongano wrote:
> > State-specific criteria have been drafted so far for: MA, MS, NH, RI,
> > VT, TX, and WA.
> >
> > In order to demonstrate what the new classification would look like on
> > the map, the New England mappers have put together a temporary live
> > demo[2] which shows what this new arrangement would look like at the
> > motorway and trunk level.
>
> This looks tolerable. I wonder how it would be applied in NY? There are
> several dead-end trunk roads. This seems wrong to me. Also, the entirety
> of the Adirondack Park is empty, which doesn't work for any community
> north of the park.


The definition of 'trunk road' still appears to be 'main route between
regionally important population centers'.  What is 'regionally important'
in northeastern New York will depend, I suppose, on what granularity you
consider for 'region'.  I suspect that OSM intends 'region' to be something
along the lines of 'United States' if not 'North America', rather than
something like 'Saint Lawrence County'. For that reason, I've been doing
some rough sketches (nothing in Brian's server yet) of what the network
might look like in eastern NY. In order to have a reasonably broadly
applicable definition of 'population center' I've been using 'incorporated
community or CDP > 25k inhabitants' (something of an arbitrary cutoff).

It makes sense to me that there are no trunk roads inside the Adirondack
Park apart from the Northway. There's nothing in the park for a trunk to
serve. Are Tupper Lake, Ticonderoga, Dannemora, Saranac Lake/Harrietstown,
or Lake Placid/North Elba 'population centers?"

 I'm finding that even on the north side of the park the population centers
that would define the trunk roads are pretty far apart: Watertown, Orrawa,
Cornwall, Montréal, Saint-Jean-sur-Richelieu, Burlington.  (I haven't dug
deep into the populations of lesser towns on the Canadian side to see if
anything else pops up.)

If we use 'part of the main route joining communities >25k population' as a
working definition of 'trunk', then a few corner cases pop up: NY8, inside
the park, appears so that Burlington will be linked with Utica, for
instance.  There appears to be no 'main route' between Cornwall and Utica;
I'm not all that comfortable with promoting minor county roads into trunks!
You're nearer to there than I am - how _do_ you connect Cornwall with
anywhere else?

If we go by FHWA classifications, NY30 and NY3 appear at least in part, but
NY8 disappears. The suburban arterials of NY85 and NY32 would end in
Bethlehem, because there's no 'population center' beyond there for them to
serve.

I think it's entirely acceptable for trunks to dead-end where the reason
for their existence ends. Thus, NY27 would be a trunk into Southampton
(population >25k) but downgraded past there because there's no longer a
large community beyond that.  The key thing is that we shouldn't have
isolated islands of trunk roads appearing and disappearing simply because
physical characteristics aren't up to spec on short sections.

Going with a tighter definition of 'population center' starts giving
perverse results in both New York and New England.  For instance 'any
county seat is a population center' promotes some Vermont villages of <1000
inhabitants, and in northeastern New York has the effect of promoting
Plattsburgh, Elizabethtown, Malone, Lake Pleasant, Lowville, Canton and
Fort Edward - and I'm not sure I'm comfortable with declaring any of those
communities to be a 'regionally important population center!'  Sorry,
Potsdam, but at least you get to keep US 11 (because it joins Burlington
with Watertown).

We're still struggling with the density extremes of the Big City and the
Big Woods, so your input is welcome!


-- 
73 de ke9tv/2, Kevin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20210518/398c5e8b/attachment.htm>