Change the map layer to Transport Map. It’s all gray now.
Sorry, but I have no idea what problem You are trying to solve now. The routing is apparently correct, small detours to “another colour” are normal and your proposition won’t change it at all.
I solved it this way. Still working on the implementation. 
Great to see that the values are similar to what I’m proposing :)
@Tolstoi21 see, it’s possible to classify cycleways into an importance hierarchy.
Nope, mostly because I live someplace where speed limits are artificially high after Republicans with Dunning-Kruger experience in civil engineering decided they know better than the state engineers do.
I focus more on the overall construction contextualized with the engineering goal if I have to think harder about it than “Yup, that’s an expressway” or “No, this is some other kind of highway”.
If we are talking about expressways in the USA, our wiki about these expressways (entered years ago) and its “two-out-of-three-ness” history remains curious, in my opinion:
These are simply a thing unto themselves, a work in progress, let’s say. They are what they are and locally are called expressways. These do fit into a larger network fabric, by being part of County routes in California - Wikipedia . California had a “great renumbering” in 1964 when some numbering protocol sense came about. There’s still a few wiggly things today, nothing terribly broken Room for improvement? Sure. Who is to say an elegant state-level route numbering protocol is even important, right?
Not everything fits neatly (there are ragged edges in California and other states) into classifications. Subjectivity abounds, as does acknowledgement “there are local quirks.” I offer my best efforts to be succinct in my tagging; we all should. That said, highway classification is at best a locally / regionally / nationally rough thing (to do, as we strive for clarity). Local exceptions and “that’s how we do it here” abound. Fitting into each other’s boxes is something we largely do, but don’t perfectly do, that’s OK. Tag your best, please. I do.
How we do things now is OK, but I’ve seen improvements and room for further improvements. I stand back and let that happen as I (“what was that?, we couldn’t hear you, Steve”) Tag My Best.
Whew, this topic is a month plus and 265 posts deep!
Generally, I interpret highway=trunk to mean the most important roads on a national or international scale. As an example, many two-digit Interstates and some two-digit U.S. Routes would be trunk roads, if they weren’t built to freeways standards (the former are).
I’d restrict highway=residential to car-centric streets that provide direct access to residential buildings. Car-centric highways that don’t fall into any of the other categories and aren’t unsurveyed would be unclassified. At least where I live, it’s usually used to mean either a highway built the same as a residential road but not in a residential area or access roads that are wider than service roads (usually with lane markings) and usually named. For example, the various access roads of the Strip and some surrounding the perimeter of a mall.
I’ve heard the roads coming on/off roundabouts as “exits” (e.g., take Exit 1 toward This Street), even if no highway=motorway_junction or junction:reftag is present.
A name is just a name, by the way. It’s still a freeway, if the “I” in “I-495” wasn’t a dead giveaway. It could’ve been called “Long Island Freeway,” but the physical characteristics didn’t change (unless they have a strict set of rules as for what name type the road should be).
Yes, you would say take the 1st exit inside the roundabout. But the entries and exits aren’t referred to as on-ramps and off-ramps.
Has anyone ever wondered whether it might also be possible that the classification on the other side of the border (i.e. not the Czech or Slovak classification criticised here) could be incorrect?
A road leading through a border — one that is not very minor — shouldn’t be tertiary. Such relations like this one should definitely not be tertiary.
This doesn’t change anything though. The Slovak side argues that correcting these is “tagging for the renderer” which is just not true. What’s needed right now is an idea on how to make countries like this one start applying correct criteria for their highway tagging — official category is represented by the ref=*.
I can’t speak to how major or minor this one border crossing to Slovakia is, but I don’t think a border crossing in and of itself necessarily raises the importance of a road.
There are 119 land border crossings between the US and Canada. Some of them are monstrosities with significant infrastructure, multiple lanes, inspection stations, larger buildings, etc.
Some of them… well, you pull up in your car to a traffic light and then sit there. Eventually the border guard in the booth notices your presence. He sighs and puts down his book, reaching for the button to make the light turn green. You pull up. Noticing your license plate from a state that isn’t the border state, he asks, “so, how did you folks find this crossing?” “It was on the map.”1
I think it’s fine for these types of border crossings to be tertiary, especially when there are multiple other border crossings nearby that are more significant.
1 This actually happened
First of all, I respect other communities in other countries and respect that they have agreed on different criteria. Road construction has developed differently in each country (even if it is often similar or countries copy each other). Naturally, this means that the focus for classification is also different. And as a rule, this has little or no effect on routing.
I am very surprised that in your example, the Slovakian section of the road was classified twice as secondary. Each time, it was mappers who usually map in Poland.
For what reason do you consider a higher classification on the Slovakian side of this exemplary cross-border road in your example to be justified? What makes this particular road so important?
After looking at street images from various sources over long distances, I am also surprised that:
translated with deepl