[OSM-dev] RFC: map call to return deleted elements

Frederik Ramm frederik at remote.org
Sat May 23 15:12:28 BST 2009


Stefan de Konink wrote:
> Isn't the problem here actually the reuse of a node? For example it was 
> moved. If show a deleted way it would still depend on the latest data, 
> hence not the data when a change was made?

You're right; a deleted way using non-deleted nodes is still subject to 
change of geometry when these nodes are moved.

It would probably be (a) difficult and (b) undesirable to return, along 
with a deleted way, the versions of nodes that were in force at the time 
the way was deleted. Remember we're not talking about a "time machine" 
function here (which would also be interesting, but entirely different).

There will be cases where this leads to strange results. Assume you have 
a simple junction of two roads. Now you delete one of them, and it 
remains deleted for a while. Someone comes along and refines the 
geometry of the remaining road, moving nodes about and inserting new 
ones. He might move the former intersection node for a few hundred 
metres and insert another node in its place (I often have little regard 
for the original position of a node when refining geometries, I just add 
a bunch of nodes and then pull them all about to make it look right.) - 
If someone later views the deleted road, it will have a major dent where 
the intersection is, because it still uses the old intersection node.

The same if I should delete the former intersection node. Restoring the 
deleted way will show you that deleted node as well, but the information 
that this node was, at the time, also part of the other way is not restored.

These are things that people would have to keep in mind when playing 
with deleted objects.


Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"

More information about the dev mailing list