[OSM-talk] Splitting a way may completely hide a taint

[Ian]

On Thursday, December 22, 2011 12:34:30 AM UTC-6, Toby Murray wrote:
>
> On Thu, Dec 22, 2011 at 12:13 AM, Russ Nelson <nel... at crynwr.com> wrote:
> > Oh, and, could we convince r_coastlines to accept the license? I've
> > put a LOT of work into fixing the coastlines (which were utter CRAP in
> > the first place), and I'm unenthusiastic about having them
> > lost. OSMInspector is reporting the wrong thing, by the way. Look at
> > http://osm.mapki.com/history/way.php?id=3753612 and you'll see a bunch
> > of nodes, only two supposedly touched by me. Yet if you click on any
> > node, say
> > http://osm.mapki.com/history/node.php?id=18666667 or
> > http://osm.mapki.com/history/node.php?id=18666670 you'll see that I've
> > touched them.
>
> When displaying a way, mapki.com only shows changes that affect the
> way version. This includes node addition/removal but not moving
> existing nodes. I brought this up with Ian on IRC and the way it
> works, it really isn't practical to do deep node inspection as part of
> the way history.
>
Indeed. This is a feature of the API as it stands: moving a node does not 
create a new way revision, only a new node revision. When viewing the 
history of a way, you won't also get the history of constituent nodes 
unless you ask for it one-by-one. I don't think the sysadmins would like me 
doing that :).

PS: The mapki.com history viewer is nothing more than a pretty view into 
the OSM API's history calls.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20111222/b6c74d18/attachment.html>