[OSM-dev] Autocorrectable data inconsistency
gregory.williams at purplegeodesoftware.co.uk
Tue Aug 19 08:55:06 BST 2008
I've only seen this thread partway through, so please tell me if I've completely misinterpreted the topic.
As I see it there is a case, at least on a practical level, for storing a series of nodes along a straight line in a way. Imagine a very long straight road that crosses multiple z12 tiles. If this were stored in its simplest geometric form then there wouldn't be nodes within (or potentially in the extra margin that t at h downloads) a z12 tile. Therefore the tile would be rendered as if that road doesn't exist.
Similarly if somebody working in an editor doesn't download a sufficiently large area to encompass at least one of the way's nodes then they may get the impression that the road hasn't yet been entered into OSM's data. That might even lead to them drawing their own road along the same path!
> -----Original Message-----
> From: dev-bounces at openstreetmap.org [mailto:dev-
> bounces at openstreetmap.org] On Behalf Of Jochen Topf
> Sent: 19 August 2008 08:24
> To: Joachim Zobel
> Cc: dev at openstreetmap.org
> Subject: Re: [OSM-dev] Autocorrectable data inconsistency
> On Mon, Aug 18, 2008 at 11:17:53PM +0200, Joachim Zobel wrote:
> > Am Montag, den 18.08.2008, 13:30 +0200 schrieb "Marc Schütz":
> > > Is there actually a valid use case for having the same node
> > > times in a row?
> > I don't see any.
> > I could write a script to korrect the existing ones. Do the
> > have to be in sequence (which means the ones above need to be shifted
> > down) or does a simple DELETE per dupe do it.
> All changes should be done through the API. Do changes on the database
> directly only leads to inconsistencies. The API does not expose the
> sequence number, so the order alone is fine.
> Jochen Topf jochen at remote.org http://www.remote.org/jochen/ +49-721-
> dev mailing list
> dev at openstreetmap.org
More information about the dev