[Tagging] Tagging Digest, Vol 117, Issue 36
354lbr at gmail.com
Thu Jun 13 08:41:16 UTC 2019
> Personally I would prefer a more inclusive definition which requires for
> lanes to be recognizable, which could be either through lane markings or
> through traffic observation (if the vehicles drive in two lanes it is a
> 2-lane road also in absence of road markings).
> Cheers, Martin
I agree with this. "Only marked lanes" has been annoying for me several
First, in my old United States neighborhood, many roads are unmarked, with
about 10% having a double stripe down the middle. Those 10% are really no
different from the rest, and often times the paint won't be reapplied when
new asphalt is laid down. So really, there isn't any difference between
these roads except for how I'm supposed to tag them.
Second, when creating the connectivity relations proposal, people asked
about what lane numbering system unmarked roads should be given. I thought
that they should be assumed to have 1 lane in each direction, and should
accordingly just contain the lane #1.
The main thing that is hard here is deciding at which width 1 lane becomes
2. In the United States, unmarked roads are usually wide enough for 2 cars
to easily pass each other, so are practically speaking 2 lanes. Here in
Europe, many rural roads are only wide enough for 1 car, meaning that a car
must pull off the road in order to let another one pass.
I really like the "observable lanes" definition, but think that switching
to this one would create the need for some extra tag to indicate that the
lanes are not actually marked to help data consumers that care about that
kind of detail.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Tagging