<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body>
<div>Hi, Mateusz,</div>
<div><br>
</div>
<div>Discussion regarding marking secondary vs tertiary was long ago. Thread (in Lithuanian starts March 2010): <a href="https://lists.openstreetmap.org/pipermail/talk-lt/2010-March/000188.html">https://lists.openstreetmap.org/pipermail/talk-lt/2010-March/000188.html</a> 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. :-)</div>
<div><br>
</div>
<div>I would like to combine arguments from that discussion and addressing your concerns.</div>
<div><br>
</div>
<div>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.</div>
<div><br>
</div>
<div>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.</div>
<div><br>
</div>
<div>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.</div>
<div><br>
</div>
<div>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.</div>
<div><br>
</div>
<div>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)</div>
<div><br>
</div>
<div>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.</div>
<div><br>
</div>
<div><span>
<pre>-- <br></pre>
<div data-evo-paragraph="" class="">Aidas Kasparas</div>
</span></div>
<div><br>
</div>
<div>On Tue, 2022-06-07 at 13:21 +0200, Mateusz Konieczny via Talk-lt wrote:</div>
<blockquote type="cite" style="margin:0 0 0 .8ex; border-left:2px #729fcf solid;padding-left:1ex">
<div dir="auto"><a href="https://wiki.openstreetmap.org/wiki/Lithuania" rel="noopener noreferrer" target="_blank">https://wiki.openstreetmap.org/wiki/Lithuania</a> page has a proposed<br>
</div>
<div dir="auto">schema for road classification and as far as I see<br>
</div>
<div dir="auto"><a href="https://wiki.openstreetmap.org/wiki/Lt:WikiProject_Lithuania" rel="noopener noreferrer" target="_blank">https://wiki.openstreetmap.org/wiki/Lt:WikiProject_Lithuania</a> has<br>
</div>
<div dir="auto">equivalent proposal<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">Some parts are a bit suspicious and likely to result in problems<br>
</div>
<div dir="auto">(rigid adherence to official road classification seems to be proposed,<br>
</div>
<div dir="auto">and official road classes at least sometimes completely mismatch<br>
</div>
<div dir="auto">their actual importance)<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">But especially<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">"<br>
</div>
<div dir="auto">The following mapping is proposed:<br>
</div>
<div dir="auto">secondary -- county roads (paved)<br>
</div>
<div dir="auto">tertiary -- county roads (unpaved)"<br>
</div>
<div dir="auto">"<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">has problematic results where treated robotically:<br>
</div>
<div dir="auto">highway=tertiary with small paved segments ends being tagged<br>
</div>
<div dir="auto">with isolated highwa=secondary what makes no sense and<br>
</div>
<div dir="auto">is extreme oddity.<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">Note that highway=* road values such as primary/secondary/tertiary/unclassified<br>
</div>
<div dir="auto">are about road role/importance and not direct tagging of road quality<br>
</div>
<div dir="auto">and such disjointed road classes are odd at best<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">It seems that such approach to road classification is considered as weird<br>
</div>
<div dir="auto">not only by myself, see <br>
</div>
<div dir="auto"><a href="https://wiki.openstreetmap.org/wiki/Talk:Lt:WikiProject_Lithuania">https://wiki.openstreetmap.org/wiki/Talk:Lt:WikiProject_Lithuania</a><br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">See for example <br>
</div>
<div dir="auto"><a href="https://www.openstreetmap.org/way/518530445#map=16/55.6427/22.4486&layers=N">https://www.openstreetmap.org/way/518530445#map=16/55.6427/22.4486</a><br>
</div>
<div dir="auto"><a href="https://www.openstreetmap.org/way/757876997#map=16/55.9277/24.0718&layers=N">https://www.openstreetmap.org/way/757876997#map=16/55.9277/24.0718</a><br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">(I sadly found about it via drama in which I also participated , see
<br>
</div>
<div dir="auto"><a target="_blank" rel="noopener noreferrer" href="https://lists.openstreetmap.org/pipermail/talk/2022-June/087549.html">https://lists.openstreetmap.org/pipermail/talk/2022-June/087549.html</a><br>
</div>
<div dir="auto">(that is how I become aware of that in the first place)<br>
</div>
<div dir="auto">and<br>
</div>
<div dir="auto">
<div dir="auto">
<div dir="auto"><a href="https://www.openstreetmap.org/changeset/121943595">https://www.openstreetmap.org/changeset/121943595</a></div>
</div>
</div>
<div dir="auto"><a href="https://www.openstreetmap.org/changeset/121951869">https://www.openstreetmap.org/changeset/121951869</a><br>
</div>
<div dir="auto"><a href="https://www.openstreetmap.org/user_blocks/6043" rel="noopener noreferrer" target="_blank">https://www.openstreetmap.org/user_blocks/6043</a><br>
</div>
<div dir="auto">
<div dir="auto">)<br>
</div>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">I made <a href="https://www.openstreetmap.org/changeset/121951869" rel="noopener noreferrer" target="_blank">
https://www.openstreetmap.org/changeset/121951869</a> edit as<br>
</div>
<div dir="auto">this classification schema was described at OSM Wiki page<br>
</div>
<div dir="auto">as proposal.<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">If, for some reason, Lithuanian community supports creating<br>
</div>
<div dir="auto">such unusual disjointed highway=secondary fragments it would be <br>
</div>
<div dir="auto">useful to clarify it. Though I want to note that such tagging <br>
</div>
<div dir="auto">is extremely weird, unusual and mismatching how road classication<br>
</div>
<div dir="auto">works in OSM.<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">
<div dir="auto">I am not proposing a new tagging schema, but I would at least considering<br>
</div>
<div dir="auto">that short isolated strips of paved highway=tertiary should not be upgraded
<br>
</div>
<div dir="auto">to disjointed highway=secondary (surface=paves/asphalt/concrete/... obviously<br>
</div>
<div dir="auto">still should be tagged on such roads to provide surface data).<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">It seems to have no benefits whatsoever and is problematic from cartographic,<br>
</div>
<div dir="auto">consistency, QA and other perspectives.<br>
</div>
<div><br>
</div>
<div>I would also consider using in general highway=secondary to be used as indicator<br>
</div>
<div dir="auto">of importance, not as surface=paved - and from looking through Lithuania<br>
</div>
<div dir="auto">it seems that some quite small roads like<br>
</div>
<div dir="auto"><a href="https://www.openstreetmap.org/way/243485255#map=15/55.7693/24.0950" rel="noopener noreferrer" target="_blank">https://www.openstreetmap.org/way/243485255#map=15/55.7693/24.0950</a><br>
</div>
<div dir="auto">are getting tagged as highway=secondary<br>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">And more widely - rigidly basing highway=* road classification on official<br>
</div>
<div dir="auto">schema will cauyse problems, and should have escape hatch at least for<br>
</div>
<div dir="auto">cases where official classification is blatantly wrong.<br>
</div>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">PS Ideally, road classification would not change on borders:<br>
</div>
<div dir="auto"><a href="https://www.openstreetmap.org/#map=17/54.22583/23.38369">https://www.openstreetmap.org/#map=17/54.22583/23.38369</a><br>
</div>
<div dir="auto"><a href="https://www.openstreetmap.org/#map=17/54.09060/23.51764&layers=N">https://www.openstreetmap.org/#map=17/54.09060/23.51764</a><br>
</div>
<div dir="auto"><a href="https://www.openstreetmap.org/#map=17/54.38052/22.81872&layers=N">https://www.openstreetmap.org/#map=17/54.38052/22.81872</a><br>
</div>
<div dir="auto"><a href="https://www.openstreetmap.org/#map=17/54.39621/22.92510">https://www.openstreetmap.org/#map=17/54.39621/22.92510</a><br>
</div>
<div dir="auto">this road is not magically losing/gaining importace and it would be<br>
</div>
<div dir="auto">nice to discuss such cases to have continous road classes<br>
</div>
<div>_______________________________________________<br>
</div>
<div>Talk-lt mailing list<br>
</div>
<div><a href="mailto:Talk-lt@openstreetmap.org">Talk-lt@openstreetmap.org</a><br>
</div>
<div><a href="https://lists.openstreetmap.org/listinfo/talk-lt">https://lists.openstreetmap.org/listinfo/talk-lt</a><br>
</div>
</blockquote>
</body>
</html>