[Tagging] decide on a meaning for what is not documented [way: Expressway=yes/no versus new tags "dual_carriageway=yes/no", "limited_access=", "grade_separated"=?]

Joseph Eisenberg joseph.eisenberg at gmail.com
Thu Feb 20 12:02:34 UTC 2020


Taginfo shows that the tag has been used in just a few places (mostly
one city in Poland) so it's not very well established, in spite of
several thousand uses:
https://taginfo.openstreetmap.org/keys/dual_carriageway#map

There are a couple similar tags that are a little less common:
* https://taginfo.openstreetmap.org/tags/carriageway=dual (688 uses,
the value =single is more common)
* https://taginfo.openstreetmap.org/tags/type=dual_carriageway (1144
uses, mainly relations - so this appears to be a type of relation,
perhaps meant to group together the 2 directions of a dual carriageway
highway?)

But fortunately the key name has a clear meaning in British English: a
"dual carriageway" should be a road where the lanes in each direction
are separated by a central reservation or barrier, aka a "divided
highway" in American English.

As pages like Key:highway and Tag:highway=trunk have long recommended,
these sort of highways are usually mapped as 2 separate ways +
oneway=yes.

As we can see in taginfo (combinations), the key dual_carriageway=* is
always used with highway=* and almost always with oneway=yes, so that
works: https://taginfo.openstreetmap.org/keys/dual_carriageway#combinations

Checking overpass-turbo shows that the tag is used on pairs of ways
which are parallel, as expected. For example
http://overpass-turbo.eu/s/QTv (try Poland for the biggest number)

It also shows that most occurences of "dual_carriageway=no" are with
highway=trunk (or rarely highway=motorway or highway=primary), and
almost all "dual_carriageway=yes" are highway=primary, =secondary,
=tertiary or =trunk.

So, that's what I've based the page text on. I have not added mention
of this tag to any other wiki pages or Map features.

If people are interested in using it, I might make a proposal. But
perhaps we will decided that database users should interpret the
geometry and name=/ref= tags of parallel highway ways to add this
information in post-processing. Does anyone know if that is feasible?

On 2/20/20, marc marc <marc_marc_irc at hotmail.com> wrote:
> Hello,
>
> Le 20.02.20 à 07:41, Joseph Eisenberg a écrit :
>> I've created a page for Key:dual_carriageway based on existing usage
>> in the database:
>> https://wiki.openstreetmap.org/wiki/Key:dual_carriageway
>
> how did you proceed? is the tag added by only one contributor to whom
> you asked the question to create the page based on his answer?
> or did several contributors suing this tag answer consistently ?
>
> Because I noticed it with the depreciation of the undocumented tag
> camp_site=camp_pitch : when a tag is undocumented, the real meaning of
> the objects in the database depends on what was going through the heads
> of the different contributors (for this particular tag, we have found 5
> meanings each used more than 100x and it took hours and numerous
> messages to ).
>
> The great advantage of an undocumented tag is that we know that the
> meaning is not documented. the contributors aiming at quality can
> therefore avoid it. the users of the data too.
> Conversely, a tag documented "by hidden guessing" degrades the quality :
> Some people will believe that this is really the meaning of the objects
> in the database when it's only a guess. at the very least, it would
> require a big banner "a contributor guess that the meaning of the tag"
>
> Regards,
> Marc
> _______________________________________________
> Tagging mailing list
> Tagging at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging
>



More information about the Tagging mailing list