While editing the OSM map I discovered many POI’s with no tags or description that were not present before. See: http://www.openstreetmap.org/?lat=52.068&lon=4.45537&zoom=17&layers=M
They are visible while in EDIT-mode in Potlatch 1.4 and Potlatch 2. Where do they come from and are they ‘ghosts’ or real POI’s that should be removed because of having no meaning? No… I’m not drunk. I’ve tried to cleanup a small area, but they seam to re-appear on different spots. Weird…(?)
It seems like “3dShapes” is uploading a big changeset (33.000 nodes). Just wait and look tomorrow again :-).
Bye, Andreas
Thanks, I’ll do so.
And please, next time, don’t just remove nodes, as they may be part of an ongoing edit. Like in this case. Also, these are not ‘POIs’. A POI is a node that carries some information about the object in the form of certain descriptive tags. A node that is (or will be) in a way is not a POI, but is technically the same (without the tags, of course).
The general advice would be to wait at least a day (or two) and then maybe clean them. An even more sane solution would be to contact the mapper that uploaded them, to ask him/her to clean it up.
Sorry, if I’ve messed up something. Didn’t know this. Been too ‘enthousiastic’ and ’ -again- I will bind my hands on my back for the time being…
Don’t worry too much. We’re all still learning new things everyday (I hope!).
It’s just that I wonder why a bunch of seemingly unused nodes would cause a reaction of “help!?” or “I’ve got to clean this up”. I sincerely do.
Thanks for your encouraging words. I think that there would not have been confusion (or ‘panic’) about the ‘POI flooding’ when the nodes being part of the upload process could appear in another shape compared to real POI’s, for example as red triangles. Real POI’s are represented by little green solid circles as were the nodes being uploaded at that time. Anyway, I learned someting new! What’s next…?
Nodes without tags (as these were) and nodes with tags (as a POI would have) appear slightly differently in JOSM. What they appear as as default these days, I don’t know, as I’ve tweaked those settings. If you had clicked on them, you would’ve seen they had no tags, and thus weren’t “real POIs”.
The basic advice in OSM usually is, or should be (:)) : if you don’t understand something, or don’t see the point of something, someone else may think differently. It’s always better to try to find out the meaning, before touching the data.
I hope you find OSM fun to work on!
That said, 3DShapes should reconsider their import method. Anything that leaves unwayed, untagged nodes around for hours is clearly going to lead to situations like this. It would be better to upload a series of comparatively small changesets that include both nodes and ways.
Richard, as BikePC said, when he cleaned up some, he saw others appear. So the upload was still ongoing. I know the person that did that region, and he’s very careful with things like this.
Personally, I’m using Upload.py (with its smarter-sort.py script in the pipeline) to upload nodes and associated ways (and relations) close together. Others use JOSM, which is still dearly missing such a sorting method on upload.
You’re quite right that when using JOSM, uploading changesets of 30k+ objects is tricky these days. It takes a lot of time to finish the upload, and in the 30-50 minutes that it runs, there will be some node clouds in that area.
I switched from JOSM to Upload.py exactly because I ran into the situation twice that during an upload someone deleted some nodes and I received conflicts. Fortunately, I know a bit of the API, so I could resurrect those deleted nodes and resume the upload. That took a bit of time, so I went looking for a better way. This turned out to be Upload.py.
/me nods. I guess either JOSM should be patched or people discouraged from using it to upload big changesets.
(Personally I wish people would stop uploading such large changesets altogether and would just go out and do some real mapping, but that’s a different story… 
)