[Foundation-l] Licensing interim update

[Sam Johnston]

On Tue, Feb 3, 2009 at 8:31 PM, Nikola Smolenski <smolensk at eunet.yu> wrote:

>> Given that full attributions are both largely worthless and onerous to
>> the point of forbidding reuse in many circumstances (e.g. paragraph
>
> Please stop beating the dead horse. No one has ever suggested that full
> attributions are necessary.

Yes they have.

>> quotes, most physical mediums, compilations, etc.) and partial
>> attributions are in many ways worse than no attributions at all,
>
> Could you specify at least some of these many ways?

Ok, so off the top of my head:

 - It is impossible to reliably determine the top contributors in a
mechanical fashion, because:
 - There are no reliable metrics for identifying 'top contributors'
(e.g. edit count vs wikiblame vs creator vs something else?) but:
 - Manual determination of top contributors creates opportunities for
internal conflict where there would otherwise be none yet:
 - Partial attribution creates opportunities for external conflict
(think DMCA, lawsuits, etc.) where those excluded take exception,
which leads us to:
 - Optional attribution which incents those who might otherwise not
care to request attribution and besides:
 - Content changes over time so the article consulted at a random
point during the life of the derivative work will differ from that
when it was incorporated; thus it's meaningless anyway unless:
 - Re-users are forced to copy the entire (already massive and
constantly growing) edit history and identify the specific version
that was used and even then:
 - Extracting signal from the noise is virtually impossible, even for
a small number of authors which takes us back to the start.

Wikis, or 'Massive Multiauthor Collaboration Sites' (as the FSF calls
them) are a relatively new concept. Copyright, attribution, etc. works
well for individuals and extends to relatively small groups (e.g.
bands, tv/film crews, journals, etc.) but many of us believe that it
breaks badly at this scale.

In any case it is clear that Erik/WMF have a good handle on the issue
and Brian's nailed it:

"With a system that can find the authors of any given piece of text no matter
when it existed in any language version:"

"Wikipedia"

Sam