On Mar 10, 2020, at 3:17 PM, Michal Migurski <m...@teczno.com> wrote: > Thanks Mikel for your followup! > > In response to your recommendation that we publish what *didn't* make it > through the filter process, we released an OSC diff file: > > > https://daylight-map-distribution.s3.amazonaws.com/pbf/planet-2020-03-06_v0.1.osc.bz2 > > At this early stage of trying to determine what’s useful for the community, > it’s pretty raw. The OSC does not yet differentiate between things we > filtered out categorically because we don't show them on our maps, vs. things > we filtered out individually because we found a problem with them. Individual > tree nodes are an example of the former: they don't get shown or labeled so > they’re not in Daylight. > > I appreciate everyone’s questions about this data release. The FB team behind > Daylight and our other mapping efforts is well aware that OSM is a tough and > curious community! > > -mike.
I realize that OSMCha (as Mapbox describes its use) and the compressed planet file extract (as Facebook publishes and begins to describe its use) are (respectively) ways to "flag changesets/features with reasons" and are characterizable as in "early stages of trying to determine what's useful for the community (pretty raw)." I'd call both of these a reasonable start, especially as Mapbox's pipeline seems a bit further along in its development and Facebook's data is in an early state, still needing more feedback from the community. I encourage the community to provide this feedback, perhaps on an ongoing/continual basis, and thank both Mapbox and Facebook (well, Mikel and Michal) for their participation and good communication here. Good QA improvement can and does work: in a project like OSM with corporate participants endeavoring on major sub-projects, it is usually based on very open communication with very wide community like this. Yay! (And conversely, as these get better established and "tightened up" into what the community expects and can participate in improving, the volume knob here can be turned down as these discussions don't have to be quite so public). SteveA _______________________________________________ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
