[Tagging] Reviving the conditions debate

[Tobias Knerr]

On 14.06.2012 13:30, Colin Smale wrote:
>> motor_vehicle:forward:(Mo-Fr 16:00-18:00) = agricultural
> At first glance this looks like a motor vehicle going "forward" between
> those times is considered "agricultural". It doesn't feel very
> intuitive, based on the established key=value paradigm.

Putting a group of users into the value is not my own invention, that's
established tagging. Mappers are using things like
motor_vehicle=agricultural or access=destination right now.

Look at the distribution of values of motor_vehicle, for example:
http://taginfo.openstreetmap.org:8001/keys/motor_vehicle#overview

I agree that it is not a particularly nice solution, and you could of
course split this into separate tags with "yes" and "no" as the values.
But I don't think it's useful to exclusively focus on the one thing that
is not new at all when discussing an example.

> motor_vehicle=yes
> motor_vehicle:forward:(Mo-Fr 16:00-18:00)=no
> motor_vehicle:backward:(Mo-Fr 06:00-09:00)=no
> hgv=yes
> agricultural=yes

Unfortunately, this doesn't work with the conflict resolution approach
currently defined by "Extended Conditions". These state that the "most
specific" rule counts (in terms of boolean logic, the one that implies
the others). In this case, there's sometimes no "most specific" rule.
When you compare

motor_vehicle:backward:(Mo-Fr 06:00-09:00)=no
hgv=yes

then for an hgv at 08:00, the time is more specific on the first rule,
but the vehicle class is more specific on the second. This means that
the most restrictive value is relevant, which is "no".

Tobias