[OSM-dev] One value per key? Why not one pair per object?

Matthias Julius lists at julius-net.net
Tue Feb 3 18:45:49 GMT 2009

Jochen Topf <jochen at remote.org> writes:

> On Mon, Feb 02, 2009 at 03:25:57AM +0100, Stefan de Konink wrote:
>> What is the motivation to go to the one value per key thing again. 
> Practically everything we have been doing with OSM tags until now has
> assumed that keys are unique. Nearly all the software I have seen (like
> editors, renderers, etc.) treat keys as if they are unique. Yes, some
> things could maybe be cooler if we allow multiple keys, but the reality
> is that basically every mapper out there and most of the software assumes
> they are unique. By making them really unique we just enforce this now
> and turn flaky software into solid software without much work. :-)
> For years now we had the option of using multiple tags with the same
> key, but didn't. From a practical standpoint I don't see how this would
> change. Why would suddenly everybody change over the software to cope
> with multiple tags with the same keys. So if thats not going to happen
> anyway, why not acknowledge that by making keys actually unique?

Now that users know that the API supports duplicate keys they would
probably push software authors to implement that feature (if they
wouldn't also know that this goes away in API 0.6).

There is some need for having multiple values for the same key and the
current recommendation is to separate them with a semicolon.
Unfortunately, no software that I am aware of supports that neither.

IMHO, duplicate keys are the cleanest solution.  But, the downside is
that it requires every software to deal with it.  String separators
don't need to be supported by editors for example.  One can always
enter them by hand.


More information about the dev mailing list