[Tagging] Feature Proposal - RFC - "is_sidepath" as a sidepath concept

[Tobias Knerr]

On 02.12.22 15:02 Mateusz Konieczny wrote:
> Or is it justifiable because it is outright impossible to do reliably
> with automatic preprocessing?

I would say that's the reason. Of course, it's hard to prove that 
something is impossible. But so far no one has built a working solution 
for inferring this information where it isn't explicitly mapped.

> If yes, can you give examples where it would be beneficial
> and explicitly ask to introduce such tagging solely in
> cases where preprocessing is doomed to fail?

Given the current lack of working preprocessing solutions for even the 
easier situations you've linked to, I'm not sure it's easy to tell where 
the dividing between "solvable" and "unsolvable" is. It would also have 
to be defined and explained in such a manner that mappers (and editing 
tools like StreetComplete who might some day ask users to fill this in) 
can easily understand where it is or isn't needed.

> If it is second case of preprocessing being impossible -
> why do massive duplication and ask to duplicate ref value
> AND name value AND highway value?

The duplication-free solution of referencing the road's way ID as 
is_sidepath:of = w4087515 is going meet some resistance from people 
pointing out the need for, and current lack of, editor support to make 
such values usable.

(Not from me, though, I'd love that approach and have dozens of ideas 
for other places where it would be helpful. :))