[Talk-lt] secondary/tertiary in proposed definitions at WikiProject Lithuania

[Tomas Straupis]

2022-06-07, an, 19:34 Mateusz Konieczny via Talk-lt rašė:
> in my experience highway=secondary is rather used for connection
> between bigger settlements, with highway=tertiary for connection
> between smaller ones

  In ideal homogenic environment - maybe. But reality is different and
some small settlement might have some a higher classification road.
For example road to Ignalina Nuclear Station is secondary. Some
settlements have two more or less parallel roads connecting them, one
paved, another unpaved, it is clear that the paved one is more
important, better maintained and therefore should be more prominent on
a map.

  While being close, I would disagree that this is a case of mistaging
for the renderer. This is just a mapping, we're mapping a set A with a
set B using attributes {a1,...,an} from a set A and attributes
{b1,...,bn} on a set B. It just happens that one of the objective
parameters is surface (otherwise we would have had no objective
tertiary in Lithuania at all). if I remember correctly, all of these
roads there tertiary, but we have raised the classification of all
roads and tertiary was left "empty", we needed a objective criteria to
separate it from secondary.

  In our secondary/tertiary case higher classification also means
higher standards of road maintenance, not only surface. Small span of
paved road in a settlement is cleaned/repaired more than unpaved
outside of a settlement. This important information would not be
possible to communicate with surface/max_speed tags (this does not
apply to <100m stretches of course)

  One thing missing in current thread but which was and still is
important - getting rid of subjectivity as much as possible. If we
cannot map classification to some objective criteria (like official
classification + what we see on the ground) there will be situations
like it was before this classification was introduced: people
unknowingly were changing classification back and forwards, because
you know, that road leading to my forest hide-out is really cool so it
should be secondary ;-) (this is actually a real story, I was
struggling with one guy trying to map a field road as tertiary for
years, because it led to his remote house in a field and he is "easily
doing 90km/h on that road").

> I would just propose considering unifying classification at least for extremely
> tiny sections (<100m?) where displaying something as paved is unlikely
> to matter much and such isolated highway=secondary is quite silly.

  With this one I agree. We will discuss this further, but my
suggestion is to find such short spans of different surfaces which are
either on bridges or in some very rare occasions some short spans of
paved road in the middle of nowhere (have no idea why, maybe some
pavement testing as some had short spans of altering different
pavements). After finding it manually we will add a tag exception=yes
(augment the data) and thus the automated rule will know what to do.

-- 
Tomas