[OSM-talk] A warning about gates and other barriers

[Steve Bennett]

On Mon, Sep 20, 2010 at 4:54 AM, Frederik Ramm <frederik at remote.org> wrote:

> You cannot make the mapper responsible for deciding whether or not a car, or
> bicycle, routing can go through a barrier; he might not know. Treating all
> un-tagged gates as access=no will lead to mappers entering bogus access tags
> (or not adding the gate at all) in such cases. This is hardly desirable.

The best thing OSM can do is to define how such cases are to be
interpreted, and then for software using the data to follow those
rules (or to break them for specific reasons).

If you look at this thread, it's essentially "here's how I read this",
"oh really? I do it differently", "oh me too, but you know, everyone's
opinion is valid!" - which is pretty much useless.

Personally, I think I prefer the default that access is open to anyone
who could be on both sides. But anyway, could we try and reach
consensus, and then document that?

Steve