[OSM-talk] riverbanks, way, segs, corrruption, user experience, traction.

Dermot McNally dermotm at gmail.com
Fri May 4 14:14:03 BST 2007


On 04/05/07, Thomas Walraet <thomas at walraet.com> wrote:

> If someone else is cleaning coastline imported by the almien_coastline
> script, I suggest this order of actions :
> 1) Select an area and launch validation with the validator plugin
> 2) On warning branch, select "segments with tags" and remove the tags
> 3) On the error branch, select "duplicated segments" to see where they
> are. Delete them with the above drag/select/delete method
> 4) You should only have the "duplicated nodes" errors left. Select and
> correct those one by one. You should have at least one of this error for
> each island and lake, so it's the perfect time to check the orientation
> of all segment (islands needs to be reversed, lakes should be left as
> they are (clockwise))

I'm currently cleaning up the Irish coastline, having imported it
using almien_coastline. I've been using pretty much your approach, but
hadn't realised the connection between duplicate nodes and islands (a
useful tip, since there are thousands of tiny islands I need to handle
and many are easily missed). As a coastline newbie, I wasn't quick to
change the script, but is there any reason it does tag the segments?
Granted, it's very quick to fix (unlike thost bloody duplicate nodes,
grr!), but shouldn't the script be fixed to avoid the problem?

The other big cleanup issue I've faced is that many (but not all)
coastline segments are created facing the wrong way. Is there any
logic to the direction chosen (such that it could be made smarter) or
is it simply down to luck?

Dermot




More information about the talk mailing list