[Tagging] Adding directionality to stop signs

[Paul Johnson]

On Mon, Mar 20, 2017 at 11:36 AM, Volker Schmidt <voschix at gmail.com> wrote:

> Given the effect of the stopping process on the overall travel time
> (mainly for cars), we need a way that can be used by routing algorithms.
>

It's even worse with bicycles, since actual human effort has to happen to
accelerate from a stop.   And even more annoying, far more frequent in some
parts of the country.  Seattle used to be a worst offender, sometimes
having 20+ stop signs per mile on what's meant to be the main bike route
through the area
<http://www.bikingbis.com/2012/02/13/smooth-bike-riding-on-newly-renovated-burke-gilman-trail-section/>
.


> I am inserting them in quantity, but only in the simple way of a node on
> the way. "My" stop signs apply to the nearest junction.
>

I do this as well, though this seems less than ideal for complicated
situations like the Elm Street underpass under the Creek Turnpike
<http://www.openstreetmap.org/way/80022916>, without getting into the fact
that putting a stop sign for bicycles at an intersection every other
direction and even pedestrians in the same direction get a traffic signal
is *intensely* stupid ground truth to start with).


> This should in principle allow a routing algorithm to determine the
> directionality of the sign. I would consider having to create (and manage)
> relations for this tedious, and error prone in case of editing.
>

Short Street <http://www.openstreetmap.org/way/422722387> immediately ruins
this assumption.  The southbound stop sign protecting Canyon Road is almost
in Beaverdam Alley.


> BTW the Stop sign is even more frequent in Italy than in the US, but it
> influences much less the travel time, because drivers simply don't stop
> most of the time, contrary to the practice in the US, or Germany, or the UK.
>

It's getting to be that way in the US, especially for bicycle facilities.
Even USDOT has acknowledged this, and as recently as the 2009 MUTCD,
warrants stop signs only when a yield sign would not be a safe alternative.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20170322/71a444e2/attachment-0001.html>