Interpretation of parent_station versus pathways

[Stefan de Konink]

Hi,

I would like to have a discussion on inference of a concept we call
pathways.txt in the GTFS specification.

It is a very common practise that on street bus systems have two stops on
the opposite side of the street. Typically they are grouped for example in
Hastus this is called "place", on the logical level in NeTEx it would be a
StopArea and on the physical level a StopPlace. In GTFS we can group stops
via a parent_station.

I would like to know - in your opinion - if the following should be done in
GTFS;

If two stops on the opposite of the road are less than 100 meters apart
(bla bla, preventing that > argument), should they be grouped via a
parent_station? From maps perspective a parent_station could be shown in
higher zoom levels. The counter argument "but it is not a station" does not
hold since we already have decided all stops with information such as
entrances and exits should have a parent_station anyway.

Now if someone respons yes, could they tell me if the parent_station
grouping is required *because* of the pathways or that is it a good
practise to have an implicit transfer area. Where pathways could be
considered an explicit transfer area, or as the definition defines it
"exhaustive".

--
Stefan