Happy Birthday, OpenStreetMap

It easily happens that employees do whatever it takes to get their internal tool to show the right thing, or use an automated tool, without ensuring they don't destroy or duplicate existing data. (made-up example: someone is refining the mapping of streets, but destroys information about adjacent/connected footpaths in the process because they don't matter for their delivery-vehicle use case. Or maps street types as they would in country A, but in country B a different norm has been established, so the edits don't fit the surrounding data)

Data quality can be a problem, where the company data source does not match reality (anymore). Or it's difficult for the community to verify your data and be convinced your quality is good, especially if there are other concerns about your edits.

OpenStreetMap operates on soft consensus and can have multiple acceptable ways of doing a thing. Mass-editing that converts one form to other is frowned upon unless wide agreement that one way is deprecated has been sought. Often related to the first point, when company tooling prefers one style or employees only have been taught about one method, but in an area the other mapping style is widely used.

Similarly, communication to resolve conflicts can be lacking: if another editor challenges what you are doing (because they think you made a mistake or are overstepping norms), you need to react and fairly consider their point. OSM has clear expectations that if you are doing coordinated edits you clearly identify and announce your work and respect feedback.

Data sources/licenses can be a concern: what you find acceptable to import into your company database might not be acceptable to include into OpenStreetMap, and how does the community tell that your data is good and won't get the project in trouble later?

What did FB do on a OSS project with AI?

If I may be so bold, scrolling around the Netherlands and what is covered of Germany, it seems pretty useless.

- The resolution is so low, it's really hard to tell if there might be missing roads/routes or if this concerns water/rail.

- There are so many water/rail mistakes, there is red everywhere also in areas where I know every road in existence is on there. You have to evaluate each tile manually for all of the covered landmass.

- Just taking out air/rail/water might not solve the problem. Looking in residential areas with no water or rail nearby, there are still yellow tiles popping up at random with no indication as to why. Having the offending trace(s) would be really helpful here, but those are of course not shareable for privacy reasons. Perhaps short cutouts could be shared that are between 25 and 75% of the route (so not near the source or destination)? That would make it very clear if the person is on a bus (that obviously won't follow the shortest path in most cases) because the trace would go past the bus stops and often linger there, or if it was a cyclist for example.

As a frequent OSM contributor I love such initiatives and I'm very happy to see TomTom getting more involved. TomTom's map quality is so far behind OSM (globally), I was wondering who'd even still considering buying from them so it makes a lot of sense to combine forces instead. User data is the main advantage Google has over OSM, for both live traffic and purposes like these, so I'm very happy to see innovation here! I'll definitely be checking out updates to the site.

Can anyone recommend good interface to OpenStreeMaps? I installed OsmAnd on my Android Radio (9" 720p) and find menu options/icons are too tiny (and too many) to operate while driving.

Looking at the places I'm familiar with, there's also a pattern of places like high traffic intersections (probably lots of people avoiding those?) and places with lots of foot/bike traffic, where people don't travel along the shortest path but along the most beautiful or the one with the slowest elevation change.

I guess there's lots of useful information for improving routing algorithms in the raw data. I too am skeptical about its usefulness for mapping.