Development of Creative Commons licenses

[Gregory Maxwell <gmaxwell AT gmail.com>]

[Limiting this response message to the URI attribution requirement
subject, sorry— I should have split my original response]


On Tue, Sep 17, 2013 at 4:56 PM, Cc <cc AT phizz.demon.co.uk> wrote:
> On 17/09/2013 23:50, drew Roberts wrote:
>> On Tuesday 17 September 2013 17:57:32 Sarah Pearson wrote:
>>> Is there a reason you think indicating a URI via terms of service is
>>> particularly problematic? Why is it worse than indicating a URI in a
>>> license notice, for example?
>> One a quick browse, I think his point is that a hosting provider could 
>> require
>> (via their terms of service) you to use a URI of their choosing on your 
>> work
>> rather than one of your choosing. I think it goes along with the point that
>> terms of service are generally take it or leave it deals and not the 
>> subject
>> of negotiations.
>>
>
> Surely it is up to the content creator whether she uses one particular
> web host or another?

You appear to be assuming that there is a very simple relationship.

Author -> Webhost -> Consumer.

This is one model, but it is not the only one.

What happens when there are many authors and they post works on many
webhosts, and then additional authors, potentially including some but
not all or none of the original authors, make further adaptations
taking from these works and placing them on other hosts?

The freedom to move material from place to place is fundamental in
sharealike, and complex authorship flows involving multiple service
providers are a reality (and has resulted in litigation).

The freedom to choose providers or multi-host is substantially
non-existent in longstanding collaborative works or works which
combine components from many sources. It's also sometimes the case
that the service may have no facility for export, or the service
itself constitutes a unique medium from which the work can't readily
be removed (e.g. the totality of your facebook profile, your
second-life virtual home, or such).

> Additionally, there is nothing to prevent a creator
> placing the same content on multiple hosting sites.

As I note above, there are often social and technical barriers and for
a long term collaboration it may be infeasible to contact all the
authors in order to obtain a free and clear license for the work
without TOS encumbrances from prior providers. In these cases the
freedom over the work has been practically reduced.

> For example if I
> place content on flickr I don't think that Mr Maxwell can come along and
> ignore the URI requirement because he has a beef with Yahoo!, or indeed
> that I should be able to ignore the URI of content on Picassa becuase I
> have a problem with Google. If the URI is a problem we are both free to
> NOT use the relevant hosted content.

I think the language of the license is ambiguous, but I can walk away
with three possible interpretations, all of which can produce
undesirable outcomes.

(1) One possibility is that the any and all URIs ever affixed via a
TOS requirement must be preserved.

You, I, and a few others collaborate on Yahoo! and create some work.
Then Yahoo! sues you when you complain about their poor service (this
is the grim reality for some WikiTravel contributors),  and so you
decide it would be better to move our community over to Picassa.  You
are now required to backlink and advertise the litigious former host
which all the active parts of your community has abandoned (but it is
infeasible to reliably contact all members to receive an alternative
license).

Additionally, when large collaborative works flow through many such
site the end result would— if the licenses were strictly complied
with— require the presentation of potentially dozens of URIs,
advertising a dozen different incidentally involved commercial hosting
services, making the work (or substantial excerpts) more difficulty to
redistribute in conformance with the license.

Effectively this understanding creates a special type of licensee
which enjoys a privileged position over the work simply by being in a
position impose some advertising hook in their terms of service, and
to distribute and display the work without any third party URIs.  This
inequality between the first user under the license and later ones
creates friction in the freedom to share and collaborate over the
work, effectively it constitutes a taking where the service provider
receives perpetual free advertising for which the author(s) may not
have been justly compensated.

In these examples the host has no copyright interest in the work, so
why do we allow them to adhere parasitic marketing via the licensing
terms?

If the URI cannot be demanded by an otherwise uninterested service
provider via their terms of service, then the above problem examples
under this interpretation largely go away.  Having to list some large
number of author mandated URIs is "just" an advertising clause, and
doesn't create issues where an service provider can use their much
more powerful position to exploit authors by affixing.

Preserving URIs as history rather than first class attribution also
would reduce their burden.

[I would note here that if (1) is the correct interpretation, I can
point to my own works moved onto Wikimedia sites by third parties
which Wikimedia and all their downstream reusers would now be
distributing in violation of the license, because they are
distributing the works as credit Wikimedia instead of my original URI]

(2) Another possibility is that any URI requirement on an existing
work can be overridden by terms of service by a downstream service
provider when the work is submitted by another party.

I take your work hosted on Yahoo! which requires a backlink to them,
and excercise my rights and post it on Picassa which then overrides
the backlink requirement.  Worse, I could create a pretextual service
"BobsSpamBlog" to rehost your works just for the purpose of TOS URI
rewriting.

If this is the effect of the license then URI requirement is pointless
since it could be so trivially mooted. Under this reading it doesn't
even sound like BobsSpamBlog itself is actually required to help
people find what the original URIs were.

(3) Another possible interpretation is that works which have been
subjected to distinct URI requirements are incompatibility licensed
and cannot be combined, and that a work under one TOS URI requirement
couldn't be uploaded to a site with another TOS URI requirement.

I believe that the language of the license mostly suggests (1) or (2),
but that (2) will be what people actually practice because no one has
a workflow to track a potentially ever growing list of things to
attribute, they can hardly get basic "attribute the authors" right.  A
problem with that outcome is that dishonest parties will get a free
pass to basically do whatever they want while the few who diligently
try to conform with the license will undertake additional cost
following a (1) approach and/or uncertainty and risk of nuisance
litigation from people insisting that (1) is correct.