[Imports] Fixing large number of mistagged buildings near Boulder Colorado, US
Frederik Ramm
frederik at remote.org
Tue Sep 8 18:58:06 UTC 2015
Chris,
On 09/08/2015 06:49 AM, Chris Riddoch wrote:
> Do you think the original mapper would be very motivated to fix this
> number of errors? That sounds like a pretty demotivating task, if you
> ask me. Certainly the error is worth mentioning to avoid being
> repeated, but... I already *have* the list of ways that need changing,
> and they all need the same specific change.
It is worth talking to the original mapper for several reasons.
First of all, they might not be aware of the mistake, and you will help
them to avoid repeating it.
Secondly, the might well in the position to fix the problem themselves,
which is always preferable to have someone else quickly run a script -
not only for educational purposes, but also because the person might
genuinely feel ashamed and might be happy to be given a chance to fix
their mistake.
You can of course still *offer* your help in a friendly way and when
they accept, you can fix it for them - but you should really talk first.
> So I ran josm with -Djavax.net.debug=all so that I could analyze the
> format properly. It looks like it isn't *quite* as simple as just
> providing a file you can just send to the server, because that
> incorporates a changeset ID which is ephemeral. I also wouldn't want
> to try such a thing without properly testing it first. So I knew I'd
> need to talk to someone here about how best to do this -- thus, my
> original message. I came here because I read those wiki pages.
One of the reasons we ask potential mass editors to talk to experienced
users first is that a lot can go wrong in a mass edit.
> Yes, it's *possible* to fix errors by hand individually using JOSM, in
> the same way it's possible to count grains of rice by hand, but... no,
> I'm sorry, I'm just not motivated to do that.
How did you come up with the list of buildings that require change. Have
you clicked on each individually and noted down the OSM ID, or simply
selected everything in an area where you know for sure that there's not
a single apartment anywhere, or ...?
JOSM can download a list of objects from the server (i.e. you could
instruct it to download exactly those you have identified), and JOSM can
then apply a change to all of them and upload. That bit is fairly
trivial. The bit that is less trivial is making sure that you don't
accidentally make a bad situation worse by applying bad selection
criteria, hence my previous question, and the most taxing but for us
technically minded people is always talking to other humans, hence my
initial paragraph.
> It sounds like the most useful thing I could do, though, would be to
> help you all develop a mechanism for reverting changesets.
Reverting chnangsets is already possible, and if I understand you
correctly, this is about buildings that were mis-tagged from version 1,
so reverting that changeset would mean deleting the buildings, surely
not the desired outcome.
Bye
Frederik
--
Frederik Ramm ## eMail frederik at remote.org ## N49°00'09" E008°23'33"
More information about the Imports
mailing list