[Foundation-l] deleting old versions of fair-use files

Michael Snow wikipedia at frontier.com
Mon May 30 22:17:48 UTC 2011

On 5/30/2011 2:32 PM, Amir E. Aharoni wrote:
> Hello,
> There's a bit of discussion about deleting old versions of fair-use
> files in the Hebrew Wikipedia and it may be interesting to other
> projects as well.
> The main questions is: Should old versions of fair-use files be
> deleted? The two main points that support the deletion are that it
> saves space on the server and that keeping a version of a non-free
> file violates the fair use policy, because the old version of the
> image can be viewed, but is not actually used in any relevant article.
I just want to see if I understand this correctly. Is this a reference 
to files where the current revision is included in some article based on 
a fair use rationale, but the file also has earlier versions that are 
not so used? As opposed to a file that was previously used in an 
article, but is no longer used at all. In the latter situation, it's 
hard to make a fair use argument because there is no actual use case to 
point to, and I think the consensus has been that those files should be 
deleted, consistent with our policies limiting the exemptions for 
non-free content.

For the first case, there's at least a plausible rationale that 
consistent with how MediaWiki operates, its public record of how the 
file was derived could be acceptable, including modifications that may 
be indicated by the file history. It might, for example, indicate how 
the current revision is a sufficiently "transformative" use to qualify 
as fair use, enough to justify the limited "use" of the otherwise unused 
old revision. Of course, if you start an analysis based on revision 
histories like that, you might also conclude that "unused" fair use 
files can be kept because they were used in previous revisions of articles.

That's primarily an abstract theoretical response, and without some 
concrete examples to look at I don't really have much of an opinion on 
which direction we should resolve the question. Nor have I tried to 
consider how persuasive a court would find potential arguments on either 

--Michael Snow

More information about the foundation-l mailing list