[Tagging] Feature Proposal - RFC - Bridge types

Christopher Hoess cahoess at gmail.com
Sat Feb 2 21:34:31 GMT 2013


Response to selected comments:

On Sat, Feb 2, 2013 at 12:57 PM, Martin Koppenhoefer
<dieterdreist at gmail.com> wrote:
>
>
> +1
> these are all bridge values with more than 100 occurrences, my
> comments inline after the percentages:
>
> yes
> 1 656 829 97.79%  the very most
>> null

Not surprising, given that other than "viaduct" (and "1", "true") it's
the only one supported by the OSM mapnik setup.

> viaduct
> 24 314 1.44%   not sure if we need this, what is the difference to a
> bridge? It mostly tells that there is a road or railway going over it
> (what you can already see from the data, as you can see the length)
>> A ''long'' rail, road, or other bridge made up of many short spans.

You may be focusing on the wrong attribute: it's the many short spans,
not the length, that make a viaduct. I think I'd file it under
bridge:type, like a trestle (which also has many short spans).

[...snip...]

> aqueduct
> 1 084 0.06%  prefer historic=aqueduct for historic aqueducts (also
> fragments) and would rather introduce a tag similar to "power" for
> water if I wanted to map modern aqueducts. Anyway an aqueduct has not
> much to do with bridges
> -
> null

Ambiguity in the word. In the US, it's used both for long structures
carrying drinking or irrigation water, above or below ground (which
seems to be your use), but also for elevated structures carrying
canals, usually navigable, across lower terrain. The latter use of the
word seems very much like a bridge.

[...snip...]

> abandoned
> 556 0.03%  could be an idea to keep these. There is also the
> alternative way to tag it abandoned:bridge=yes
> -
> null

As I said, I haven't yet tried to systematize the various ways of
saying a bridge is unused/damaged/removed and so on, so I wouldn't
interfere with these.

Anyway, I agree with your classifications of other values that I haven't quoted.

> I agree, looks as if bridge=yes is the only remaining value if we
> introduce type and structure and prefix abandoned, disused, etc.

Since it's already renderer-supported, that would be quite helpful.

Yours,

-- 
Chris Hoess



More information about the Tagging mailing list