On Thu, Aug 14, 2008 at 09:46:45AM +0200, Marcus Wolschon wrote: > On Thu, 14 Aug 2008 07:56:26 +0200, Jochen Topf <[EMAIL PROTECTED]> wrote: > > On Thu, Aug 14, 2008 at 03:44:57PM +1000, Brett Henderson wrote: > >> Karl Newman wrote: > >>> Even still, it's a valid concern and there are other operations (i.e., > >>> cutting into tiles or tag transforms) that can manipulate the data > >>> into a form that probably shouldn't be uploaded. > >>> > >>> Karl > >> Yep, agree. If it wasn't the default behaviour it would just be far > >> less likely to occur :-) > ... > > But people *want* to use the data they get from some kind of extraction > > process to base decisions upon. And they should be able to. Not all the > > data is invalidated by an extraction, only some of it. It makes sense > > to mark those parts invalid that are, so that an automated process can > > decide what it uses and what not. > > I propose to not modify ways by default and to remove the way-id of > ways that are modified during extraction. > What should a tool care if the way references node-ids that it has > no nodes for because they are outside the bounding-box? It has to ignore > them. > > The removal of the ID is because this is no longer the way with > that ID but a new and artificially created way that contains a part > of the original one.
If you remove the ID the way can't be referenced any more from relations which you might also have in the OSM file. But I agree that the way is somehow different, not the same anymore so it might not have "a right" to this ID. In a more general case, if you remove some tag from a node and leave of the ID because of that, you can't reference it any more from ways, which might make the node useless. There are good reasons for removing, say all "created_by" tags, because they take a huge amount of space and most people will not need them. There is another option, I want to bring into the discussion: In the case of the Osmosis extract we could keep the node list, but not include the nodes itself. This would keep the way object intact for the price of breaking referential integrity. This is the solution employed for relations, which are not resolved recursively (or only to a certain degree). This way we don't have to make objects incomplete. Still wouldn't solve the imcomplete tag thing. Jochen -- Jochen Topf [EMAIL PROTECTED] http://www.remote.org/jochen/ +49-721-388298 _______________________________________________ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
