foxy

@osm_tech Is it possible to implement a softer blocking method, such as blocking every fourth tile? Perhaps this would make the transition period less painful for users and also serve as a clear hint to app developers.

@BalooUriza :)

I suggest something like this. You can still get your bearings if it's important to someone, but it's obvious that the map will be unpleasant to use and the developer will have to do something. And if this does not help after a while, then block it completely.

@BalooUriza But why not have a more softer mechanism? Imagine that if someone violates OSMF tile attribution, you could not just write an email, but show the seriousness of your intentions by blocking some tiles. Most likely, users will still be able to use the application, but the developer will fix the violation much faster.

@BalooUriza Another example. A bunch of applications send Referer: localhost https://planet.openstreetmap.org/tile_logs/ These are potentially applications used within companies. But perhaps they are some kind of self-hosted applications used by ordinary people. Or applications in the process of being developed on a programmer's laptop. A soft partial block would be much more appropriate here. You can still see the map, but you understand that you are doing something wrong.

@BalooUriza Fan fact, QGIS loads OSMF tile servers more than it's a delivery app. It is important to look at tps_miss, this is the number of requests that did not hit the CDN cache

@BalooUriza what does Google's soft block look like?

@BalooUriza QGIS users are a separate topic for discussion. As much as we like QGIS being open source and so on, they are very problematic tile consumers. All I wanted to say is that sometimes you can reduce problems for unsuspecting app users. Just an idea that if it's easy to implement (potentially it's a simple regular expression), then you can take it into service. p.s. I'm also not a member of working groups, these are public logs