[OSM-talk-be] GRB hackday december 11th

[Jo]

It wouldn't make sense for anybody to put a deadline on it. It's important
to communicate this to the 'integrators' though, that's the point I was
trying to make. I definitely didn't stress on it enough, when I prepared
the data for UrbIS.

Maybe we should also foresee procedures for cases where it's obvious people
take the data and just 'run with it'? If needed we shouldn't hesitate to
revert changesets. Who's going to follow up on that though? reverts should
ideally happen as quickly as possible.

About landuse that attaches to other landuse or buildings. JOSM has a tool
to nicely stitch those together. It's in the contourmerge plugin.

Jo

2016-12-15 11:38 GMT+01:00 Marc Gemis <marc.gemis at gmail.com>:

> On Thu, Dec 15, 2016 at 8:59 AM, Glenn Plas <glenn at byte-consult.be> wrote:
> >> A model import does not introduce errors that could be avoided by
> >> simply looking at aerial imagery
> >
> > Does, or does not ?
>
> I break the silence, just because you asked a question :-)
>
> every building imported via a model import should be checked
> individually. If you see that the aerial imagery does not show the
> building, look at a third source. Depending on the age of the sources
> take action.
> I have seen buildings that were destroyed, but visible on the
> grootschalige aerial images and GRB, but not in the field or on the
> kleinschalige aerial images. So you do not do an import then.
> Ideally, every building that you add is validated by the mapper. This
> is a slow process.
>
>
> Some of the problems I encountered during the church improvements
> - sub area of the church was marked as tower. This sub area add to be
> adapted to the new floorplan. 3D information will give the same
> problem.
> - almost every grave yard add to be added to nicely fit against the
> wall of the church. Maybe this problem is unique for churches as other
> buildings do not have landuse attached to it ?
> - in some cases you have to move paths as they would otherwise lead
> through the church. Also occurs for footpaths between 2 buildings.
>
> Other problems
> - addresses that do not match, the geometry merge plugin is not really
> helpful to see which address comes from GRB and which one was in OSM.
> - replacing multipolygons with individual buildings and somehow
> keeping the history.
> - Take a look at https://xian.smugmug.com/OSM/Screenshots/AGIV-Problems/
>
> so there are a lot of things to check while you add the data, all of
> them will slow you down during the import.
>
> @Jo, I am glad you do not put a deadline on it. Thanks !
>
> regards
>
> m
>
> _______________________________________________
> Talk-be mailing list
> Talk-be at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-be
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-be/attachments/20161215/c82f323a/attachment.htm>