[Talk-lt] secondary/tertiary in proposed definitions at WikiProject Lithuania
Aidas Kasparas
a.kasparas at gmc.lt
Tue Jun 7 13:36:39 UTC 2022
Hi, Mateusz,
Discussion regarding marking secondary vs tertiary was long ago. Thread (in Lithuanian starts March 2010): https://lists.openstreetmap.org/pipermail/talk-lt/2010-March/000188.html Later this was discussed live with some drinks in the hands. And results were written into wiki page. Yes, that page was not marked as mandatory, but remember that internet standards are also just Requests For Comments. :-)
I would like to combine arguments from that discussion and addressing your concerns.
First, the official classification outside cities in Lithuania is quite good. I know just one road which do not have road number when it probably should have. I don't know location where change of classification would be appropriate. We would be glad to make an exception for road or two which would produce illogical result.
Second, every year some country roads are paved. This is done taking into account importance of the stretch of the road (plus, if it goes over some village). So, important stretches gets paved and upgraded in the OSM database. If more gaps are left, then the whole road is less important in country's or wider context. So state of map reflects state on the ground. In our book this is how things should be.
Third, there are different legal requirements while driving on paved roads and on unpaved roads -- 90 vs 70 km/h max speed, right of the way for driving on first vs second.
Forth, in all practical applications we could come up at the time of the decision we expected that one will work with set where all or none of secondary and tertiary will be included. So, we considered this to be safe.
Fifth, number of map errors in LT is way lower than in neighboring countries. And this is a result of automatic tests run nightly, education of new mappers. Guess who does this hard job for longer than a decade? At his own expense in time and equipment. Thank you, Tomas, very much! (we often forget to tell this)
But this does not mean that he could do whatever he wants. I do remember when mine and his opinions were different. But technical arguments always prevailed. On some occasions mine were accepted, on others -- Tomas'. So, he is not a dictator here.
--
Aidas Kasparas
On Tue, 2022-06-07 at 13:21 +0200, Mateusz Konieczny via Talk-lt wrote:
https://wiki.openstreetmap.org/wiki/Lithuania page has a proposed
schema for road classification and as far as I see
https://wiki.openstreetmap.org/wiki/Lt:WikiProject_Lithuania has
equivalent proposal
Some parts are a bit suspicious and likely to result in problems
(rigid adherence to official road classification seems to be proposed,
and official road classes at least sometimes completely mismatch
their actual importance)
But especially
"
The following mapping is proposed:
secondary -- county roads (paved)
tertiary -- county roads (unpaved)"
"
has problematic results where treated robotically:
highway=tertiary with small paved segments ends being tagged
with isolated highwa=secondary what makes no sense and
is extreme oddity.
Note that highway=* road values such as primary/secondary/tertiary/unclassified
are about road role/importance and not direct tagging of road quality
and such disjointed road classes are odd at best
It seems that such approach to road classification is considered as weird
not only by myself, see
https://wiki.openstreetmap.org/wiki/Talk:Lt:WikiProject_Lithuania
See for example
https://www.openstreetmap.org/way/518530445#map=16/55.6427/22.4486<https://www.openstreetmap.org/way/518530445#map=16/55.6427/22.4486&layers=N>
https://www.openstreetmap.org/way/757876997#map=16/55.9277/24.0718<https://www.openstreetmap.org/way/757876997#map=16/55.9277/24.0718&layers=N>
(I sadly found about it via drama in which I also participated , see
https://lists.openstreetmap.org/pipermail/talk/2022-June/087549.html
(that is how I become aware of that in the first place)
and
https://www.openstreetmap.org/changeset/121943595
https://www.openstreetmap.org/changeset/121951869
https://www.openstreetmap.org/user_blocks/6043
)
I made https://www.openstreetmap.org/changeset/121951869 edit as
this classification schema was described at OSM Wiki page
as proposal.
If, for some reason, Lithuanian community supports creating
such unusual disjointed highway=secondary fragments it would be
useful to clarify it. Though I want to note that such tagging
is extremely weird, unusual and mismatching how road classication
works in OSM.
I am not proposing a new tagging schema, but I would at least considering
that short isolated strips of paved highway=tertiary should not be upgraded
to disjointed highway=secondary (surface=paves/asphalt/concrete/... obviously
still should be tagged on such roads to provide surface data).
It seems to have no benefits whatsoever and is problematic from cartographic,
consistency, QA and other perspectives.
I would also consider using in general highway=secondary to be used as indicator
of importance, not as surface=paved - and from looking through Lithuania
it seems that some quite small roads like
https://www.openstreetmap.org/way/243485255#map=15/55.7693/24.0950
are getting tagged as highway=secondary
And more widely - rigidly basing highway=* road classification on official
schema will cauyse problems, and should have escape hatch at least for
cases where official classification is blatantly wrong.
PS Ideally, road classification would not change on borders:
https://www.openstreetmap.org/#map=17/54.22583/23.38369
https://www.openstreetmap.org/#map=17/54.09060/23.51764<https://www.openstreetmap.org/#map=17/54.09060/23.51764&layers=N>
https://www.openstreetmap.org/#map=17/54.38052/22.81872<https://www.openstreetmap.org/#map=17/54.38052/22.81872&layers=N>
https://www.openstreetmap.org/#map=17/54.39621/22.92510
this road is not magically losing/gaining importace and it would be
nice to discuss such cases to have continous road classes
_______________________________________________
Talk-lt mailing list
Talk-lt at openstreetmap.org<mailto:Talk-lt at openstreetmap.org>
https://lists.openstreetmap.org/listinfo/talk-lt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-lt/attachments/20220607/1414b865/attachment-0001.htm>
More information about the Talk-lt
mailing list