[OHM] Identity through time
SK53
sk53.osm at gmail.com
Wed Apr 29 09:05:23 UTC 2015
H Mattias,
Just as OSM has no notion of a named road (it can be lots of linear pieces)
and the named road is deduced by proximity, so in OHM we have no notion of
a single object changing through time, instead such links need to be done
by both geographical and chronological proximity.
Te reason for this is fairly straightforward. In a full relational model
the pieces of a road would be in a 1:m relationship with the road object.
Similarly in a temporal relational database for a boundary there would be
many entries each with a distinct non-overlapping time range. This type of
direct relationship is not possible with tags, although you could use
relations.
What we do not know is how easy it is to pull boundaries for different time
periods together.
Jerry
On 29 April 2015 at 05:20, H MK <sockenkartor at gmail.com> wrote:
> Hello everybody!
>
> In the Stockholm area there are now boundaries for many ecclesiastical
> parishes. In the city of Stockholm, the census districts for the time they
> existed (1877–1926) should be done, even if they need a lot of fact
> checking.
>
> In order to separate different areas of the same entity, I have named both
> the parishes and the census district with the years they had that
> territorial extension. This is not satisfactory. Reasonably, a parish is
> the same object after a minor territorial change. As it is now, they become
> a new object whenever there is a change. Or, is it ok to have several
> objects with the same name? I would like to strip the years from names.
>
> /Mattias
>
> _______________________________________________
> Historic mailing list
> Historic at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/historic
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/historic/attachments/20150429/010d217a/attachment.html>
More information about the Historic
mailing list