Agreed. Happens on google maps too, and that's just so irritating
I put in a random street (mission street in San Fransciso), this how it looked for me on desktop:
Then I started filling out streets with the continuous name of the street (first copy and pasting the name in small squares, then by hand adding red dots to separate them). This is how it looked after:
I was pretty careless/moved quickly, but that should match some of the effect of imperfect algorithmic mapping decisions. The key insight is that even if one specific crook looks horid (not enough space to lay it out neatly), humans will just look at another part of the street to determine the name. But if it's really a little nook where the crooked letters don't contain enough information we'll pinch/zoom in to make it big enough to display legibly. If you really look closely at my picture you will see places where I totally garbled street names through careless copy and pasting and overlap. No problem at all.
The "after" is much more beautiful/useful.
But notice two streets are blank: https://imgur.com/a/I24REM0
This is because their name didn't appear even a single time in my view. that means even though I'm on desktop and there's plenty of space, I would have to zoom or scroll to see them! They're literally information that is missing from the map.
For other streets I had to look a fair way away to copy and paste, since the street name wasn't anywhere near where I was working. (For example, looking at my first, original map, suppose you're at the top-left point, on Jersey street: how long do you go down Jersey street to the right, before you hit the Jersey and Church street intersection? This is a super common question!)
I think the resulting map is extremely repetitive, but pretty useful/beautiful. Humans can filter out repetitive information really easily. It's just no problem for us at all. Tasks such as the one I mentioned, where you are about to go down Jersey street and need to know when you'll hit Church street, are some of the main types of navigation tasks we do every day.
So I don't see the issue with making the repetitive map. Instead of roads, it should just repeat the name of the road.
(On the other hand, those of you who disagree can at least see what the Google engineer was talking about.)
For me, it's not hard to read. I'd prefer it, anyway.
Still, if you wouldn't prefer it then we can see what the Google engineer meant, as this is exactly what you get if you show as much information as possible!
(it just says 'roadroadroad', 'houseshouseshouses', 'riverriverriver' ... :))
And if Kopec in the lower center were the name of a street, I think it's great that it's written there.
The only thing missing is the red dots, so instead of roadroadroadroad it said road·road·road·road (by the way you might wonder why I made the dots red in my example: it's so the eye can just focus on the red dots to sparsely get the road's path.)
I just don't see the issue at all. It's way better, at that zoom level, to have to use my human ability to notice and ignore patterns and ignore noise while concentrating on what I want, than to use two fingers to try to zoom in enough for the map to place the names. I just don't see the issue with the view you presented. It looks great!
For non-slavic speakers, I especially like how it stops at the border, so you can get half a screen of normal map and half a screen of strange words.
I often want road names in Tokyo but they are also competing for space of transit lines as well as places with double or triple decker differently named things like there's a freeway named A over aroad named B which is over a Subway line named C
Road names are usually not as important in Japan as they are not used in addresses but they come in handy when explaining to a cab driver how to get somewhere
I showed what would happen if you literally follow the rule "whenever [...] there's enough room on a visible street to put a label, then put the label".
That's the specification. If you show me a twisty road with no room for a label, then according to this specification it doesn't get the label. The specification (for what we're testing) says "whenever there is room". Whenever there is enough whitespace on the road.
Here's an app of a company I worked for, that cares about non-shopping use cases: https://windymaps.com/app
Maybe they would appear if you zoomed in even further, but at that point the map would be pretty useless.
