I can see this working. It’s a combination of my proposal and what @Langlaeufer proposed further up in this thread.
On footway
, oneway=yes
by itself is unclear, but oneway=yes oneway:foot=no
is clear, as is oneway=yes oneway:foot=yes
.
With the proposal approved, there will be two ways of tagging oneway
on footway
unambiguously. One is avoiding bare oneway=
and just tagging the mode-specific one-way tags, e.g. oneway:foot=
and oneway:bicycle=
. The other is generic oneway=
plus oneway:foot=
.
The implications of both tagging options for editors and data consumers are the same: bare oneway
without oneway:foot
should be flagged by validators. Presets should discourage mappers from setting just oneway=
. (Only talking about footways here.) Pedestrian routers and pedestrian maps only need to look at oneway:foot
.