[Tagging] bridge:name and tunnel:name

Arne Johannessen arne at thaw.de
Sat Aug 22 21:20:38 UTC 2020

Martin Koppenhoefer <dieterdreist at gmail.com> wrote on 15 August:
> On 15. Aug 2020, at 17:33, Arne Johannessen <arne at thaw.de> wrote:
>> Therefore, the tunnel's name is the primary name for that particular way, and thus belongs into the name=* tag.
>> The full name tagging for a road tunnel should usually look like this:
>> name=The Tunnel
>> highway:name=The Road
> I would see this as an awkward exception to the whole system if we followed your reasoning and said that in the case of highway=* + a specific property this property would change the semantics and the property would define the feature while the highway (or waterway) would become secondary.

That's not what I'm saying at all. In fact, I'm only applying *exactly* what's currently documented on the wiki's name=* page, which considers pragmatics instead of semantics.

In other words, instead of focusing on the objective meaning of tags, it focuses on their meaning in context of real-world usage.

In particular, as documented, name=* should contain the "common default name" of an element, whatever it may be. This means that for any particular element which e. g. has the two names Foo and Bar, but which is most commonly referred to by locals only as Bar, the Bar name should go into name=* and the Foo name into another appropriate name tag (alt_name=*, xyz:name=*, whatever fits).

That's not an "awkward exception" – it's the current system of the name tags.

You seem to suggest there is a restriction for name=* to only apply to the primary tag key ("highway" etc.). However, such a restriction doesn't currently exist.

> To me it seems clear that a tunnel is often more than just the road leading through it, so that the logical consequence is that the tunnel=yes is interpreted as a thing being inside a tunnel (i.e. tunnel is implicit), just as it is the case with bridges (man_made=bridge is the bridge, bridge=yes means on a bridge).

You know what, your use of "implicit" might be tripping me up here.

Please consider <https://www.openstreetmap.org/way/195986873>.

In your view, is the actual tunnel "Lærdalstunnelen" itself (as opposed to the E16 road) currently included in OSM data – yes or no?

> Also note that highway:name is objectively an unused tag with only 188 occurrences for a total of 178 million highway objects, […]
> Also compare this to 12815 occurrences of tunnel:name.

Those numbers don't appear to be meaningful in this discussion. But they piqued my curiosity, and I looked into current practises for road tunnel naming in OSM. Anyone overly interested will find the results here:

The primary conclusions are:
- Both tagging variants (tunnel name in name=* vs. tunnel:name=*) are very common.
- There are significant regional differences in the use of these variants.

Consequently, neither variant should be dismissed out of hand.

> I see your interpretation as a change in paradigm and would invite you to formally propose it with the proposal process in order to check the support of the community, if you really believe this definition would be beneficial.

Neither stated OSM policy nor actual tagging practise bears out the position that using name=* for tunnel names is somehow a "change of paradigm".

Arne Johannessen

More information about the Tagging mailing list