[talk-ch] Semantics through abstraction (Re: Pls, don't delete sidewalks, visual impaired people might use them)

marc marc marc_marc_irc at hotmail.com
Fri Aug 25 10:00:41 CEST 2017

Le 25. 08. 17 à 08:13, Marc Mongenet a écrit :
> I don't understand the sidewalk accuracy argument.
> With street width, the sidewalk location can accurately computed, no?
Yes the width of the road would greatly improve the geographical 
precision, but there are 2 problems :

- there is a problem of semantic : what is the width of a street ?
For me, width on an object is the width of that object. Therefore,
for a road with sidewalk, it is the width including the sidewalk.
For other, the width of the road is "car driven" and it is
the total width of the road without the sidewalk.
I did not see (but did not seek) a proposal to try to remove
the ambiguity between the 2

- I have almost never seen a width tag on a road. Measuring the width
of a road in town is not easy due to traffic. A laser measurement is 
required. An estimate can also be made by crossing the road and counting 
his footsteps, trailing before to know the distance of a stride.
We can, however, also use "lane" which would allow an accuracy of about 
50cm (half the difference between a narrow road and a wide road).
An accuracy of 50cm is often better than the gps error.
I started last week by adding it to street that do not have a default 
lane value like residential.

More information about the talk-ch mailing list