skbun wrote:Okay. So there is a gray cities layer, and a layer that shows names of cities for the gray layer. If memory serves me, they're 'cities_01' and 'cities_names'. As best I can reverse engineer, these are only generated at map tile generation time. They're used in the following ways:
Waze client (Android at least): BOTH
Waze Livemap: Only the placement of city names
If this is the case, then why is M, Ohio still appearing in the client?
Even though "M" is not showing at this zoom level, that giant kite shape laying waste to the great state of Ohio is the aforementioned "M, Ohio". GizmoGuy and I took care of this by Nov. 19, it no longer appears in the colored WME City layer, and here it is still in the client. And yes, I refreshed the tiles...
Even more interesting, however, is when I zoom in around me...
There are seemingly random squares of "M, Ohio" that appear regularly along my commute between Bowling Green & Toledo. Even better, depending on the zoom level, some disappear while others appear. These segments all have no city attached to them (confirmed in WME).
skbun wrote:If the above is how this works, and Waze could tell us either way, right?, IMHO, it'd be to our advantage to have cities_p update as often as Waze will do it. Two big reasons:
1. The closer to real time this layer is run, the more often we can check the cities layer to see if smears have been resolved or new bogus cities have appeared - particularly if they increase the number of colors and boldness of borders to make them easier to see; and they may be caught/removed before they're ever even seen on the clients. But, especially,
2. I'm almost certain that 'The highlighted road is too far from the city it was added to' is dependant on the cities_p layer, not the one we see on the live map tiles. Put another way, if the city polygon layer for WME were run with each edit, we'd never have the 'M, Ohio covers half the state' problem ever again**. In any case, the more often it runs, the less we'll see it. I can happily add a fictitious 'T, California' on two segments, between saves, 150 miles away from one another, so this really leans toward 'It's about that multicolored polygon layer'. To me, it seems Waze doesn't know 'how far too far is' unless there's a polygon to check against.
Personally, I don't understand why the WME & client city layers couldn't be the same thing. Editable polygons that could be fixed by CMs as needed and then left. After all, how many city boundaries change on a weekly basis? Then, the need to type in a city for a segment is eliminated and, therefore, these phantom cities. I know I'm not a fan of city boundaries being determined by city names per segment.
Am I missing something here? Is something like I'm thinking of simply a pipe dream? Please, feel free to educate me!