Given the cultural origins of OpenStreetMap, some silliness is to be expected 
4 Likes
I believe that residental and unclassified should be on the same level? I think unclassified is basically “residental but outside of residental areas".
2 Likes
Apparently the best way to kick off a talk about highway classification in OSM ist to start a topic asking not to talk about highway classification in OSM … 
19 Likes
why you think it would make it resistant to differing interpretation of tagging schema in sub-communities?
2 Likes
yes, it could be better
and highway=unclassified is unduly confusing
but
3 Likes
Aaand that’s a reason for a switch in the highway scheme. The key names are just ambiguous on their own.
There would have to be clear guidelines that for official classification you have highway:category and for actual importance you have importance=* which must be coherent (always connect to a road of equal or higher importance with only a couple exceptions).
As I said at the beginning of the other thread. It really is possible. If there was no unnecessary negative feedback that’s attacking the ambition, we could’ve just tagged importance=* everywhere and the most cutting-edge software would adapt to the new scheme and in 10 years we can deprecate the current highway=* importance hierarchy values.
2 Likes
This method works in real life too. Every time any politician say “let’s not talk about it” people start talking about it even more.
(In all seriousness, making this thread was not a good idea)
3 Likes
Well I had to respond to some of those claims and then everyone got a feeling that they must reply to me. But I don’t know why this is a new thread instead of piling it onto the other one. Or even better why not just not interact with the thread if you don’t like the concept.
Frankly, the fact that we can even describe a large cohort of mappers as classification specialists is a community health problem in itself. We classify roads out of necessity and sometimes for fun, but tinkering with classifications endlessly suggests a lack of things to do, which hopefully isn’t the case for a project as ambitious as OSM.
11 Likes
The subject of this thread is meant to be “endless tagging disputes sap energy”, not “but my tagging idea is ACTUALLY great and here’s why”. Somehow you have again diverted it to the latter, thereby proving point 1.
11 Likes
Okay, but can we just discuss this one specific tagging dispute across the USA/Canada border?
I’m kidding, I’m kidding.
@pavvv, not every tagging dispute has to be solved OMG right now this very minute. There are areas where mappers in different places disagree and that’s OK. You are talking a lot and not listening very much. One mapper in Poland, ignorant of North American community discussions, is not going to just waltz into the situation and say “Aha! I have solved it. Tag it this way because it looks trunk to me.” No.
What I see is that you suffer from a collection of pitfalls here:
- Proposing solutions with a painful inability to clearly state a problem
- Thinking that these discussions are easy and can be solved quickly
- Not working hard enough to understand other points of view
- Lack of awareness of the intensity of highway tagging debates and all of the many community discussions which have happened in the past
And probably others. That’s resulting in you bursting into the scene like the Kool-Aid Man, with a rather naïve understanding of the situation. You seem to think that nobody has thought about any of this stuff before.
I appreciate that you are trying to solve real problems and are experiencing the frustration that goes along with building consensus in a complex community of competing ideals. But please, take a step back and do some more reading to understand how we got here before dazzling us with brilliance like we’re all morons.
7 Likes
That’s exactly what I’m saying. Everyone talking about no chance of implementation think I want to implement it right now but instead I want to make the scheme as good as it can possibly be, soon enough start using the tagging and in several years, when most software adapts, deprecate the current scheme.
Well lately people have kept asking for reasoning with all the why’s while there hasn’t been too much concrete feedback but just “your scheme sucks with no chance of implementation stop wasting my time”.
See? The term ‘trunk’ is extremely ambiguous. It seems like the majority of the world saw an expressway and said “yep, looks like a trunk to me!”. My goal is a coherent proposal that uses as little ambiguity as possible.
Honestly, I don’t think it’s an issue on my end that I have to explain why inconsistency is bad.
I’m not thinking that. Of course it takes time and I haven’t figured out the whole scheme myself which is why I’m engaging in these threads. But making posts like the first one in this thread just does not help at all.
Really, I do understand them but it seems like others don’t understand mine. I get that in the US the highway tagging is fine, but for me, who sees how different the tagging between Poland, Germany, Czechia, Lithuania and Ukraine is, it’s clear that there’s no consistency whatsoever.
Could you show me any of the notable ones then? Most of them have probably been about living_street, motorway and maybe expressway but nobody has ever questioned primary and secondary.
1 Like
Obviously the solution there is to tag importance:forward=* and importance:backward=*
12 Likes
But yet, you felt it was appropriate to write a whole proposal on the subject. Stop doing that.
Honestly, I’m a little irritated that you can’t be bothered to start typing search terms into Google and the forums and looking back into the mailing lists. Maybe stop rushing to respond to every little thing and spend some more time searching and reading. You’re making a proposal, don’t you think it’s therefore your job to locate and link all of the relevant background discussions that have happened and how they went?
Do you have such vanishingly little regard for everyone else’s time that they need to do your research for you? I will indulge, slightly. If you want to know how this evolved in the United States, this is a sample of your required reading list:
- The Talk-us May 2021 Archive by thread
- [Talk-us-massachusetts] Massachusetts Highway Classifications
- https://www.youtube.com/watch?v=jWQUldGg43A
- Slack
But that is just a single country, and these are just some of the many conversations that have happened. If this is an area you are serious about working on, step 1 is actually understanding why things are the way they are.
3 Likes
No, I literally have not made a proposal. If I had, there would’ve been a normal proposal page and maybe even an rfc instead of a sandbox page and a generic-titled thread.
Maybe I searched the forum, I don’t remember. I don’t check or read mailing lists because the GUI is just horrible and not user-friendly. You provided links to discussions in the US community when I know that the scheme has been fixed there and the biggest issue is in continental Europe. What I’m looking for is a thread like this one about the highway tagging scheme on international scale.
1 Like
Or how about:
- Canada side:
highway=primary / trunk,highway:CA=primary,highway:US=trunk - US side:
highway=trunk / primary,highway:US=trunk,highway:CA=primary
Brilliant! I’ve solved cross border classification disputes.
Joking in case it’s not obvious. This would be terrible.
5 Likes
I’ll never forget what my boss told me soon after I started my first job as a computer programmer working on a mature codebase. He congratulated me on resisting the temptation to refactor a decades-old routine to use the correct data structures and language style that I had learned from school, on making a surgical change to fix a bug without going wild tinkering with purely stylistic changes all around it. Trimming all dangling whitespace, converting all the spaces to tabs, or futzing with the Hungarian notation was considered the mark of an inexperienced programmer who poses a risk to the project and might need to be shown the door eventually. Fortunately we’re a lot more patient and welcoming than that.
(I eventually got called out for spending too much time tidying up the internal wiki, but that’s another matter.)
You should definitely want OSM to be as good as it can possibly be. Making the scheme as good as it can be would confuse the means with the ends.
7 Likes
Hang on, what if we went with:
highway:uid:2622285=trunk
highway:uid:38090=primary
highway:uid:219297=unclassified
.
.
.
highway:uid:165=road
Data consumers could calculate whatever “average” they wanted to use and it would end all highway classification debates forever.
Anyhoo, got to go and pop some 
6 Likes
the parentheses are there because service is usually said to be for access roads, roads that lead to something (and end, or continue as track, etc.), i.e. a functional classification, but this also usually implies they are less important than residential roads.
Trunk is used differently, in countries like Germany or Italy, it is a physical class, to decide between primary and trunk we look at level crossings and access ramps, divided carriageways. In other places, trunk is for roads that are higher in the importance hierarchy than primary roads. The parentheses are meant to say it can be hierarchical but not necessarily, depends on the country
1 Like
oh, we have a lot of other things in the same key as well, that are very different from roads, with respect to these, highways which are determined legally instead of by agreeing on their importance are very much the same in this context, just to name a few:
rest areas
bus stops
stop obligations
pedestrian crossings
raceways
corridors in buildings
platforms
traffic mirrors
milestones
and much much more
The keys in OpenStreetMap are not necessarily to say everything you can find there is the same kind of thing, it is just a rough separation, particularly those keys that were there first, (before we discovered hstore ;-) ), amenity, highway, etc.
7 Likes