[OSM-legal-talk] [Rebuild] Progress update

[Ian Sergeant]

On 22 June 2012 04:28, Simon Poole <simon at poole.ch> wrote:
> Actually in real life the damage to nodes is substantially less
...
> more importantly,
> 0.53% points (so not quite half of the 1.2%) of the tainted nodes are from
> imports, that, should we so wish, could be reimported (on the case of
> ABS2006 with better data).

I'm not sure that percentages are useful applied globally and in this
way.  As you point out, the percentages are skewed by large imports,
both tainted and untainted.

In the particular case of the ABS2006 data that you mention, all the
unmodified imported data has already been removed.  The remaining data
is all modified in some way.  Rivers, roads, coastline, etc, use and
change this information. I think it is accurate to say that all the
remaining ABS2006 data, when deleted, will be losing some manually
created information that can't be directly created from a reimport.

Of course, the significance of that information is hard to estimate.
The Nile River has many nodes, but infinitely easier to create than a
dense mass of inner-city streets with skyscraper obstructed imagery,
gps canyons, turn restrictions, one way rules.  We have no measure
that I've seen which is useful to assess that - but there is no doubt
that in several cities it will be very significant.

Ian.