On 12/6/2019 1:28 PM, Janko Mihelić wrote:
I think the "forward" and "backward" don't belong in a role of a
relation. Oneway=yes on a way should be enough. In the Wiki discussion
it is said that if there is one little "oneway" way in a big branch,
then all the ways in a branch should be checked to see if the whole
branch is oneway. But that means we are doing the work of a router
directly in the tags.
We should just mark "oneway" ways as such, and leave the rest to the
routers.
If I recall, there were cases of one-way routes whose constituent ways
were *not* one-way. I can't bring one to mind offhand though.
Also, "main" and "alternative" are orthogonal to "forward" and
"backward". We should then have "main:forward",
"alternative:backward", and so on. That doesn't make sense, and is not
what "role" is traditionally used for. Public transport routes used to
use them, but not in the new scheme.
I think you're correct about the orthogonality but I disagree with the
implication. "forward" and "backward" without a prefix can be presumed
to refer to the main trail. If roles like "alternate:forward" and
"shortcut:backward" prove necessary and useful then mappers will and
data consumers will use them, but I don't feel they need to be part of
the initial proposal.
Jason
_______________________________________________
Tagging mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/tagging
