Hi Courtney, these are absolutely great observations!
The tile distribution is a neuralgic point. A couple pf personal observations (neither board nor EWG approved opinions) where the OWG may or may not correct me:
First of all, all discussions I have been aware are on the nuclear containment design end of bike shedding, so they wane quickly.
Second, the resources are to large part being given to us from Fastly. So we cannot just sell things we generously have got for free.
Third, use of tiles is extremely uneven between different sites referring to OSM tiles. A tile service for the next door’s physicians homepage would be so cheap that would be to both us and the physician much more overhead than anything else to put a price tag on that. A physician can see probably around 400 people a month, so 5000 page views in a year would be probably below one US dollar. We may want to offer at least an easy path for these site operators to voluntarily pay a little fee but no enforcement to keep the goodwill instead of scraping pennies.
This brings us to fourth, that OpenStreetMap does profit from an ecosystem of service providers that also sell tiles services, with proper service level agreements and thus starting at much larger scales. The reason again being the overhead. OSM profited and profits from these service providers because many of them provided back a lot of innovation (Disclaimer: I’m no such service provider or in special relations, but I acknowledge things like the Geofabrik downloads, or Andy Allans cycle layer, just to name two that now exist for far more than a decade). Just selling tiles would surely cut into that innovation and inflate the OSMF by many clerks for the billing instead of self-employed innovative people, so ultimately not bring substantially more money into the ecosystem.
Fifth, far more users see OSM data processed by a third party than see OSM tiles. If we put too much emphasis on tiles than we feed the myth that OSM were a tile provided while its mission is to be a community that provides geodata.
Sixth and finally, it is quite hard work to chase those that dodge fees if any. Log file analysis is time intensive, draining resources from productive work.
I’m happy to have more options for voluntary recurring donations. Say, one to three layers between the OSMF personal membership fee and the lowest corporate tier. I’m all for cost transparency, and to tell people e.g.: for 50$ a year you pay for running the current OSM core infrastructure for one hour, or you pay for the delivery of OSM tiles for 2 million page views.
But I do not expect there any sane source of income from a tile toll, because this is a nightmare on overhead cost on any level from technical enforcement and billing up to the community damage and goodwill destruction it produces.
OpenStreetMap
Discussion
Comment from drolbr on 12 June 2023 at 12:15
Hi Courtney, these are absolutely great observations!
The tile distribution is a neuralgic point. A couple pf personal observations (neither board nor EWG approved opinions) where the OWG may or may not correct me:
First of all, all discussions I have been aware are on the nuclear containment design end of bike shedding, so they wane quickly.
Second, the resources are to large part being given to us from Fastly. So we cannot just sell things we generously have got for free.
Third, use of tiles is extremely uneven between different sites referring to OSM tiles. A tile service for the next door’s physicians homepage would be so cheap that would be to both us and the physician much more overhead than anything else to put a price tag on that. A physician can see probably around 400 people a month, so 5000 page views in a year would be probably below one US dollar. We may want to offer at least an easy path for these site operators to voluntarily pay a little fee but no enforcement to keep the goodwill instead of scraping pennies.
This brings us to fourth, that OpenStreetMap does profit from an ecosystem of service providers that also sell tiles services, with proper service level agreements and thus starting at much larger scales. The reason again being the overhead. OSM profited and profits from these service providers because many of them provided back a lot of innovation (Disclaimer: I’m no such service provider or in special relations, but I acknowledge things like the Geofabrik downloads, or Andy Allans cycle layer, just to name two that now exist for far more than a decade). Just selling tiles would surely cut into that innovation and inflate the OSMF by many clerks for the billing instead of self-employed innovative people, so ultimately not bring substantially more money into the ecosystem.
Fifth, far more users see OSM data processed by a third party than see OSM tiles. If we put too much emphasis on tiles than we feed the myth that OSM were a tile provided while its mission is to be a community that provides geodata.
Sixth and finally, it is quite hard work to chase those that dodge fees if any. Log file analysis is time intensive, draining resources from productive work.
I’m happy to have more options for voluntary recurring donations. Say, one to three layers between the OSMF personal membership fee and the lowest corporate tier. I’m all for cost transparency, and to tell people e.g.: for 50$ a year you pay for running the current OSM core infrastructure for one hour, or you pay for the delivery of OSM tiles for 2 million page views.
But I do not expect there any sane source of income from a tile toll, because this is a nightmare on overhead cost on any level from technical enforcement and billing up to the community damage and goodwill destruction it produces.