On Thu, Jan 17, 2013 at 09:19:25PM +0000, Bernhard Reiter wrote:
> 
> Bernhard Reiter added the comment:
> 
> I think we should split out the problem that the history is not kept
> consistent with parallel changes of a node. It is separate from the problem 
> of recognising where a change comes from. And it is an important issue, 
> because data is lost. Some in-between states may not recorded and they 
> may hold important data for the history, depending on how the tracker is 
> used.
> 
> Ralf, what do you think?

You mean we should put that into its own issue? Yes, please, good idea.

Thanks,
Ralf