[Rebuild] We're better off detecting splits and merges

Richard Fairhurst richard at systemeD.net
Sun Feb 5 12:50:15 GMT 2012

ant wrote:
> Whether existing tags were checked against better sources can,
> unfortunately, not be detected by algorithmic means, so we must assume
> they are the work of the mapper who added them first.

I am really, really uneasy about that.

If we try and guess splits and merges - and that's all it can ever be, a 
guess - then we will end up giving all manner of false negatives. It is 
impossible to tell whether a new way with coincident (clean) nodes and a 
coincident (accurate, factual) tag was actually derived from a previous, 
now deleted way. If the nodes are clean and the tags are correct, you 
really cannot assume that it is an infringement.

I would far rather we have no special treatment of splits or merges; and 
that we make extra efforts in the first two weeks after the licence 
change to accommodate any objections from people who claim that their IP 
has been infringed by unintentionally carrying over a tag or two as a 
result of a split.


More information about the Rebuild mailing list