Rich Kulawiec rsk at
Sat Jan 21 12:11:50 UTC 2012

On Fri, Jan 20, 2012 at 03:06:04PM -0500, Ricky Beam wrote:
> Upon receiving notice a file is infinging, they know that *file*
> is illegal, and must now remove all the links to it, not just the
> one that was reported.  

But what -- *exactly* -- is an "illegal file"?

As Leo Bicknell astutely pointed out in this thread:

	"Also, when using a hashed file store, it's possible that
	some uses are infringing and some are not."

His example goes on to explain how this is so.  (And I'll point out
that his example applies, for example, to Amazon.  There are coprighted
files there -- e.g., books, music  -- which may be used legally
by those who have purchased them.  Do they become infringing
if someone finds a way to access them without authorization/payment
because Amazon's programmers made an error and left a backdoor open
that allow them to be retrieved via static links?  No, they don't.
Should Amazon delete them in this instance?  No.  Amazon should fix
the backdoors, i.e., remove the spurious links.)

Suppose that Joe and Jane are photographers.  Joe has produced image X
(to which he holds copyright) and Jane has produced image Y (similarly).
Digital images X and Y are used as inputs to program P which produces
output Z that is visually unrecognizable -- that is, anyone who looks
at it sees what appears to be random noise.

Does Z infringe on Joe or Jane's copyrights?  How?  Why?

How does this change (or does it change) if program P' which can
reverse the actions of P exists?

Let me give another example, this time using content that is intrinsically
illegal -- and to avoid triggering hot-button responses, I'm going to
posit a hypothetical: marshmallow peep dioramas.  Let's suppose that
these are illegal in every country on the planet, that those responsible
for them are universally reviled, that it's a crime to photograph them,
possess photographs of them, etc.

We thus conclude that a file consisting of a picture of one of these
is always illegal: that is, it's illegal no matter where it's found.

Now what happens if that picture is decomposed into individual files, each
consisting of one row of pixels from the original?  None of those files
contain anything recognizable as a marshmallow peep diorama.  The original
cannot be reconstructed from any one of them.

Is any one of them illegal?

Further: reassembling these will require something: an index, an algorithm,
some construct that allows the individual files to be recombined.
(This construct contains no content of any kind, marshmallow peep or
otherwise.  It's merely a recipe for putting together files.)
Is that construct illegal? 

If those individual files are spread across a multitude of hosts,
are any of those hosts holding an illegal file?  How would they know?

(If you're going to argue that those individual rows of pixels are illegal
because the original is illegal, then replace the above with "individual
pixels".  I trust nobody will argue that a single pixel is illegal.  Ever.)

One more scenario: a photo of a marshmalllow peep diorama
is encrypted and uploaded onto server A.  Does server A hold an illegal
file?  How would the operators of server A know?  How would anyone
(other than the uploader) know?  Now suppose that the uploader,
the only person on the planet with the decryption key for that file,
dies; therefore, the file is reduced to -- for all practical purposes --
a random collection of bits.  Is that file still illegal?  Why?  How?
Who will be able to determine this?  (Schrodinger's cat paradox in 1...2...)

I posit these thought experiments (and I'll stop here, although
many others suggest themselves) to highlight some serious
problems with terminology, and with the law: it's an attempt to apply
the principles of the physical world to the digital one, and it's a
total failure.  The putative sharp dividing line between "legal file"
and "illegal file" doesn't really exist -- although many people would
like it to exist, hope it exists, etc., because it serves their agendas
or would make things easier for them.  That doesn't make it so.

Sometimes the world changes, and sometimes when it does, it's time to
discard outdated philosophy that no longer applies to current reality --
because stubborn attempts to hang onto it at all costs, especially by
warping it into something completely unrecognizable from the original
framework, really DO cost, often dearly.  (It's 2012, and there are
still inferior people living on this planet who assign more credibility
to astrology and ghosts than to evolution or anthropocentric global warming.
This isn't funny or quaint any more.  It's stupid and dangerous.)

Schneier famously said "Trying to make bits uncopyable is like trying
to make water not wet".  What we are witnessing is precisely an
attempt to do that, via a combination of anti-security technology
(e.g., DRM) and purchased legislation, orchestrated by failing,
legacy companies run by insatiably greedy people.  These people simply
don't care how much damage they do, how many lives they destroy,
how much they hold back civilization, how much they twist the law,
-- as long as they get paid.  They are *exactly* like one of their
own famous characters:

	"It can't be bargained with.  It can't be reasoned with.
	It doesn't feel pity, or remorse, or fear.  And it absolutely
	will not stop, ever, until you are dead."

See, for example:

which points to an excellent exposition by Clay Shirky on this very point.

So: {Internet, Hollywood}: choose one.


More information about the NANOG mailing list