[Talk-us] US highway classification

[Anthony]

On Sun, May 29, 2011 at 12:36 PM, Richard Welty <rwelty at averillpark.net> wrote:
> On 5/29/11 11:59 AM, Anthony wrote:
>> Anyway, why argue about it?  If you have a reason to start
>> aggressively collecting data the missing maxspeed data, just do it.
>
> argue in the sense of a civil discussion of two distinctly different points
> of view is not an unreasonable thing to be doing right now, we have a
> bunch of distinctly different points of view and so far we're not entirely
> uncivil about it (although sometimes pushing it a bit.)

I'm not sure where we disagree, though.  I've got no problem with you
aggressively collecting maxspeed data.

> here is the problem with the state by state classification trick. there's
> been a standing effort do just that for a couple of years now:
>
> http://wiki.openstreetmap.org/wiki/OSM_tags_for_routing/Maxspeed#United_States_of_America
>
> as you can see, maxspeed data for exactly 3 states has been entered
> in that time frame. i'd be tempted to call this effort a dismal failure.

Do you really think Mapquest couldn't collect maxspeed information for
all 50 states in a day or two?

> there are two additional problems in addition to the obvious one that
> the tables aren't filled out and seemingly won't be:
>
> 1) getting the routing engine authors to pay attention to this data
>
> 2) the sometimes tremendous variation between highways which default
> and highways with speed limits that are at variance with these tables.
> in surveying primaries, secondaries and tertiaries in Rensselaer County
> (upstate NY) i frequently see explicit speed limits that are between
> the default 55mph and the residential presumption of 30mph.
>
> i really don't think there's any substitute for actually collecting
> maxspeed for the network of major roads. sure, try a temporary
> hack based on classifications, state-by-state if you want, but
> i suspect that effort won't be very profitable.

Well, you just said you know that MapQuest's router uses a temporary
hack based on classifications.  I would assume that hack is state by
state, since max speeds are set by the state.

> collecting maxspeed, by comparison, while very tedious and time
> consuming, goes directly to the issue at hand in routing engines.

When that effort to collect maxspeeds for all ways is completed then
we can have this discussion all over again.  Of course, by that time
we probably won't have use for roads as most people will just use
their teleporters.

> and it also allows us to focus the classification discussion on making
> "pretty" maps, which is probably where it should be focused.

I'd say "useful" maps rather than "pretty" ones, but yeah, that's
where the classification really matters.  Even beyond the maxspeed
distinction, the classification dispute only seems to be an issue when
there's no other route for miles and miles anyway.  So I don't think
the routers particularly care about this dispute.

> and the distinction we're then discussing is a simple one:
> classification as trunk based on physical characteristics vs
> perceived function in the road network. the former is much
> less subjective than the latter, but the latter does have its
> passionate supporters. but i don't see the point in discussing
> "major intercity routes" without a definition of what constitutes
> an urban area that qualifies as an endpoint or midpoint.
> that's why, for example, i'm not clear on whether NY 7/VT 9
> from Troy, NY to Bennington, VT deserves trunk rather than
> primary. at this point, i'd call it primary, but NE2 is sure it's
> trunk (incidentially, i'd probably also upgrade NY 2->MA 2
> to primary as well.)