Wouldn't web actually be a good example for browsers to leave the native behind? A lot of native map use has never moved beyond the generation of mapsforge (vectors rendered in the fly into bitmap tiles that are then used as bitmaps), because apparently that's good enough. In the browser, there's everything that evolved around MBTiles and the only thing missing would be some offline repository API filling the gap between a system map and isolated PWA containers.
Sure, there's a lot of mapbox on the native side as well, but that is far more restricted to high profile customers (e.g. the Strava app) than in the browser. At least from my superficial impression, libre forks of mapbox legacy code seem to be far more alive in the subset related to the web than in the subset related to native.
One cause I think is the lure of offline capability: when I'm in the browser I don't mind loading on demand, but as soon as native enters the picture I'll want at least an escape hatch to downloadable maps. And then the Pandora's box of provisioning, backwards and forwards compatibility is wide open and you happily stick to mapsforge or whatever it is you already have.
(ps: and my post above was supposed to start "Wouldn't maps actually" instead of "Wouldn't web actually", but you seem to have read intention instead of words just fine)
> least from my superficial impression, libre forks of mapbox legacy code seem to be far more alive in the subset related to the web than in the subset related to native.
Just based on my superficial impression I think normal apps on the play store pay mapbox if they don't have a big budget and consider maplibre (the fork of mapbox's most recent approach) if they're big enough.
My surface impression is maplibre is used by companies like Facebook who can afford to hire low level graphics devs and fork mapbox's work rather than pay the new license.
For every feature, including offline, I think paying mapbox is the cheapest option with decent quality up to that scale.
Of course, there are projects like mine that aren't motivated by profit. (I'm using mapnik, an independent oss project that got significant mapbox investment years back but which they now consider legacy).
