[Tagging] reference_point and landmark for addresses

Felix Delattre linux at delattre.de
Wed Mar 21 19:09:16 GMT 2012

On 03/21/2012 07:06 AM, Steve Bennett wrote:
> I seem to recall having read either the articles Felix posted, or
> similar ones. The point is, in some countries, these informal
> descriptions actually *are* genuine addresses. There's no other
> addressing system in place, so over time they become the de facto
> standard. So what I think Felix is suggesting is being able to define
> the reference points that addresses are constructed from, in exactly
> the same way as we define name=* for a highway=*, or for a place=*.
> I think it's worthy of discussion get this right. landmark=* is
> problematic because as noted there actually may not be a landmark
> (like the little tree which is actually not visible). Some kind of
> addr:reference_point=*? Or maybe a kind of place=*?

This is exactly the point what I wanted to touch and consult how we can
define the best generic way. Then documenting it and putting into
practice in these countries.

I would like to recap:

* The tag landmark is not suitable for all possible reference points
* A new tag would be a good option

addr:reference_point=* or reference_point=*
What could be the right values? true, yes, popular, confirmed,...? Any

On 03/21/2012 08:38 AM, Nathan Edgars II wrote:
> On 3/21/2012 9:06 AM, Steve Bennett wrote:
>> The harder question is if you want to try and define actual addresses,
>> like actually putting a unique address description on each dwelling
>> ("From the church, 400m south", "From the church, 380m south with the
>> blue door"). But maybe leave that harder question till later :)
> This is where I get confused. Is the "address" created like this
> actually unique, or are there any number of descriptions that are
> equally valid?
> (Aside: I'm reminded of metes and bounds descriptions of property:
> begin at the stake in the old tree at the northeast corner of Bill's
> property, run south 59 degrees east for 600 chains for the point of
> beginning, then run by the following courses... Here of course there
> are many possible starting points and ways to describe the route to
> the point of beginning.)

This goes way to far and, in my opinion, is too complex for considering
it for mapping. I don't think it is necessary (or even possible) to use
exact addresses. Having at least reference points marked in a way
computers can process them would be a big improvement. Let's stick to that.

Thank you for all your responses.

More information about the Tagging mailing list