[Tagging] Are tunnels only below ground? (Was

Andy Robinson (blackadder-lists) ajrlists at googlemail.com
Wed Nov 4 10:11:55 GMT 2009

Richard Bullock wrote:
>Sent: 04 November 2009 9:39 AM
>To: tagging at openstreetmap.org
>Subject: Re: [Tagging] Are tunnels only below ground? (Was
>> I concede.
>> In fact my OLD Encyclopadia Britannica states that a tunnel is excavated
>> underground and a "cut and cover" is not truly a tunnel.
>> So the question now is how to tag an above ground "tunnel-like" structure
>> to properly indicate it's characteristics, that is "completely enclosed
>> all sides, save for the openings at each end".
>We don't *have* to stick to dictionary definitions here when tagging, as
>long as the meaning is clear;
>If it walks like a duck, talks like a duck, then define it to be a duck.
>I wouldn't hesitate to tag a cut-and-cover structure as a tunnel in OSM.

Agreed, the cut and cover approach is just the method of construction. The
end result is still a tunnel in most instances as opposed to a long bridge.
One way of telling the two apart is that typically the horizontal deck of a
bridge sits on top of but is not integral with its side supports
(abutments). In a cut and cover tunnel the deck/lid is normally integral
with the sides to form a composite structure.

>A passageway through a building (but, say, without being inside that
>building) is, to all intents and purposes, a tunnel. It doesn't necessarily
>matter whether the "tunnel" is through a brick-built structure instead of,
>say, a man-made earth embankment or natural hill.

The question here is whether the structure is a building with a hole in it
or a tunnel with a building attached. The tagging should logically fall out
when you think about the overall structure first.

>By all means expand the tunnel key like others have done with bridge=* to
>describe the tunnel properties;
>e.g tunnel = cut_and_cover / tunnel = avalanche_tunnel etc.

Agreed, though as a tunnelling engineer I'd probably call an avalanche
tunnel a bridge in some cases on inspection, but I'd also expect the layman
to tag it as a tunnel in most instances.

>But I think the key here is that tunnel=yes should be allowable to get the
>basic meaning across without an editor requiring to be an expert in tunnel

Also agreed. The meaning is obvious if it has that tagging

>Of course there will be edge cases, but we don't have to go and invent a
>million new keys to describe similar, but not identical, classes of object.



More information about the Tagging mailing list