[Tagging] Semi-colon separated route values

Martin Koppenhoefer dieterdreist at gmail.com
Sun Jan 23 22:07:19 UTC 2022


Am So., 23. Jan. 2022 um 13:15 Uhr schrieb Andy Townsend <ajt1047 at gmail.com
>:

> > The obvious alternative is to create quasi-duplicated relations with a
> > single route=* value. It seems simpler, easier to dealt with for data
> > consumer and also to maintain (what if I'm a cyclist and don't know
> > about horse-riding specificity ?).
>
> As a data consumer it makes essentially no difference to me.  As you
> said, the tagging is just a fact, so as a data consumer, you just deal
> with it.
>
> As an example, The Pennine Bridleway
> https://www.openstreetmap.org/relation/50288 in England is tagged like
> this and the extra work to render routes of this type is one line of
> code:
>
> https://github.com/SomeoneElseOSM/SomeoneElse-style/blob/master/style.lua#L7626
>


this is the line:
( keyvalues["network"] == "ncn;nhn;nwn" )) then

it is just one line if you only want to support network=ncn;nhn;nwn
but if you also want to support the (identical meaning?)
network=nhn;nwn;ncn
and
network=nhn;ncn;nwn
and
network=nwn;nhn;ncn
and
network=nwn;ncn;nhn
and
network=ncn;nwn;nhn
and the similar
network=ncn;nhn
and
network=ncn;nwn
and
network=nwn;ncn
and
network=nwn;nhn
and
network=nhn;ncn
and
network=nhn;nwn

then it will be more lines of code, and potentially your performance might
slow down if you do it for all values and all tags (it meaning split by
semicolon into an array and look what you got).


>From a tagging point of view, where it does make sense to have separate
> routes is if some sections differ for different modes.



even if some (potentially not interesting properties) differ, e.g. a route
that was opened in 1986 for hikers and cyclists and in 1992 became
accessible and signposted for horse riders as well, would have different
start_date tags, although the current situation is probably not that
different from a route that started as a combined route for all 3 modes of
transport and did never change. (And someone will also come up with a tag
like start_date:horse=1992, because you can always invent a new key in OSM).

At first, it might seem almost the same amount of work to produce 3 almost
identical relations (simply duplicate one and adjust the tags) or to have a
combined one. Which variant is better for later maintenance will depend on
the kind of modification you expect. If there will be variations for one
(or more) means of transport (e.g. horse riders at this one spot take a
different route for some meters (or more)), the individual relations are
easier and less work to modify (less errorprone also, and the modifications
are much more likely to be done at all), but if the modifications are
always relating to all means of transport, then it is much lighter to have
just one relation to care for (also less weight in the db).

Without being able to back it with actual research it seems obvious that
people will be more hesitant to make modifications for just one means of
transport if it is tagged as combination with others, and it is more likely
that the object will not be interpreted correctly or even skipped
completely if there are multiple values in a relevant tag.

Cheers,
Martin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20220123/3f4b1cb2/attachment-0001.htm>


More information about the Tagging mailing list