The precision/accuracy is not only limited by the instruments used but also the knowledge used.

For some things OSM has access to very precise data. In other instances it is fuzzy. For some things .. the past entries has been much improved by new data from other sources (sometimes opening of government sources)

No mater the precision/accuracy .. is the information usefull/informative? That should be the criteria for data entry, not its' accuracy/precision. Signifying the accuracy/precision has no formal tag .. I usually enter a note if I am concerned, or if I am really uncertain and want to wave a flag that it should be fixed .. then a fixme tag suits. But I have no objection to 'fuzzy' data ... provided it is usefull/informative.

On 28/03/2016 9:59 AM, Dave Swarthout wrote:
This sort of object is common in Thailand. We have many gated communities here whose boundaries are not exactly known although they are sometimes fairly obvious in aerial imagery because of being surrounded by a wall or fence of some sort. I create a polygon using Bing imagery, tag it as place=neighbourhood, name=* and add a fixme or note tag indicating that the boundary is inexact. Later, if a mapper has better data available they can update that boundary.

Most polygons in OSM are simply not precise enough to define the property boundaries or even the object's position exactly. Such measurements are, practically speaking, beyond the capability of our instruments, and we must accept that in our tagging philosophy. Obviously, forests and woods, wetlands, and the scrub bordering them are not clearly defined. Yet we usually tag them as areas rather than nodes so they will show up in a more useful manner on a map.

I see no problem with this whatsoever.

Cheers,
Dave

On Mon, Mar 28, 2016 at 3:40 AM, Clifford Snow <cliff...@snowandsnow.us <mailto:cliff...@snowandsnow.us>> wrote:

    Fuzzy boundaries do have their place. Currently we use sharp
    boundaries for landuse, but often the boundary is really fuzzy. A
    wooded area would be a good example of a where a fuzzy boundary
    might be employed. But the fuzziness of a wooded area may only be
    a few meters. The fuzziness of "Shakespeare Country" is completely
    different.

    I agree that there are advantages to including fuzzy boundaries,
    but we should first document how to tag these features.

    On Sun, Mar 27, 2016 at 12:59 PM, Colin Smale
    <colin.sm...@xs4all.nl <mailto:colin.sm...@xs4all.nl>> wrote:

        If we can't mark polygons as fuzzy, then we can only allow
        'accurate' polygons. Then we are back to square one, with no
        way of accommodating these regions except for a simple node.

        I think the problem is clear (how do we represent regions
        whose boundaries are not precisely defined). Time to talk
        about solutions.

        The status quo is without any guidelines, possibly leading to
        random creativity according to the whim of the mapper concerned.

        Another option is to not do it, to say such things have no
        place in OSM, and actively reject any attempt to do so (i.e.
        if anyone dares to put "Pays de Bray" or "Shakespeare Country"
        into OSM, the objects will be deleted and the mapper admonished).

        Or we go for the single-node approach, and lose out on any
        clues about the extent of the area concerned.

        Or we accept "best-guess" polygons with "incremental refinement."

        Any offers?

        //colin

        On 2016-03-27 21:36, Martin Koppenhoefer wrote:



        sent from a phone

        Am 27.03.2016 um 21:16 schrieb Anders Fougner
        <anders.foug...@gmail.com <mailto:anders.foug...@gmail.com>>:

        Did you already consider a fuzzy tag (such as fuzzy=yes or
        boundary_fuzzy=yes)?


        that's a makeshift which isn't quite elegant and still has
        similar problems (things that seem to be in might be out and
        vice versa).


        cheers,
        Martin
        _______________________________________________
        Tagging mailing list
        Tagging@openstreetmap.org <mailto:Tagging@openstreetmap.org>
        https://lists.openstreetmap.org/listinfo/tagging

        _______________________________________________
        Tagging mailing list
        Tagging@openstreetmap.org <mailto:Tagging@openstreetmap.org>
        https://lists.openstreetmap.org/listinfo/tagging




_______________________________________________
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Reply via email to