In https://github.com/tilezen/vector-datasource/issues/2019 we stripped names from short road segments (good). But looking at 512px sized 14/9612/6755.mvt I still see a huge amount of names on minor_roads, with names unique in English and Arabic. Most or all of those road features shouldn't have names, and should be merged into a huge multilinestring.
In https://github.com/tilezen/vector-datasource/issues/2019 we stripped names from short road segments (good). But looking at 512px sized 14/9612/6755.mvt I still see a huge amount of names on minor_roads, with names unique in English and Arabic. Most or all of those road features shouldn't have names, and should be merged into a huge multilinestring.