it's not clear to me whether we remove the street entirely, but i opted for just access:no for now. there are no street markings, just bollards. i didn't take note of whether the two blocks on the larkin side exist

the building deliberately does not connect to smith hall to the east; the overhanging section does not meet the neighboring building.

there are two bits that are roof-only that overhang the path. that roof is at level 1, while the enclosed building has two levels. the existing crossing points were correct. also, there was an impossible one-way section in the car park, for which i closed the loop

this is based on a site visit, but i went only from ocean to the building, so can't speak for anything north. bing aerial is accurate, google maps is not, and other site plans within the ccsf site are not.

additional notes: stop at hayes/masonic was relocated, as it uses the same nextbus ref and no other lines share it. route:disused probably isn't the right marker, but didn't see any official way for lifecycle tagging. after issues are noted and cleared, i'll do the inbound side

there are no traffic lanes south of ellis. i changed the crossings to 'tram/path' from 'rail/path' so they could be marked as having signals. removed no-left from powell to ellis, as ellis is one-way. please review/correct

this removes some of the turn-restriction relations around market toward southbound church, which now is tagged as cycle-only. this also was flagged by a note. could use confirming eyes

stockton st doesn't have actual bends until market st, but osm has had a meandering line for it. i am about to survey the market st intersection anyway, as the lanes in osm don't look right

when van ness brt was mapped, some of the van ness ways got a rather meandering look, particularly south of market and north of sacramento. i'm trying to put the way down the center of the traffic lanes

these aren't true crossings; they're an artifact of having tracks in traffic lanes