ietf-openproxy
[Top] [All Lists]

RE: OPES Ownership

2001-02-03 04:55:41
Hilarie Orman wrote,
I think there's a legitimate question here, about what 
integrity guarantees can be expected.  And I think it's 
difficult, but may worthwhile, to draw up some guidelines.  
Things such as

 By default, content integrity is assured.  

Sadly this clause is useless without a definition of content 
integrity. And that's going to be hard to come up with unless we 
have some idea of content equivalence, ie. given a tranformation, 
f(), content integrity is assured iff,

  r is-equivalent-to f(r)

But what's content equivalence? Clearly it's content type
dependent and application type dependent: a GIF might be
consisdered equivalent to a JPEG in some but not all 
circumstances; an HTML document with inserted banner ads might be 
considered equivalent to an HTML document without in some but not 
all circumstances; an HTML document in French might be considered 
equivalent to a document in English in some but not all 
circumstances.

Given that these example transforms only scratch the surface, and 
are already a bit of a computationally intractable miscellany 
this might be a bit worse than 'difficult'.

 Refusal to deliver content is not modification

Can you motivate that?

 Publishers and users should have clear ways of 
 specifying acceptable policies for content 
 modification.

This depends on having a way of specifying transforms. But what's 
a transform? This looks like being no easier to regiment than 
content-equivalence.

 Users should have clear ways opting out of content
 modification services.

If you mean on an all or nothing basis, then this might be 
feasible. But to be useful we probably need something more fine 
grained.

 The content should have an audit trail of modification
 services applied end-to-end

Hmm ... but doesn't the audit-trail itself become part of the 
content? Peeling off a modification audit-trail seems like a 
clear cut case of an unacceptable, integrity-violating transform, 
but I can conceive of cases where it might not be. Equally, I'm 
unsure how a prohibition against such behaviour could be enforced 
(or how such behaviour could be guaranteed to be visible to the 
end recipient(s)) consistently with further transformations being 
applied.

 Content services should not move data between unrelated 
 transactions

For this constraint to be meaningful we'd need a definition of 
'related transaction'. Again, this looks like being no easier to 
specify that either content equivalence or transform.

Cheers,


Miles

-- 
Miles Sabin                               InterX
Internet Systems Architect                5/6 Glenthorne Mews
+44 (0)20 8817 4030                       London, W6 0LJ, England
msabin(_at_)interx(_dot_)com                         http://www.interx.com/



<Prev in Thread] Current Thread [Next in Thread>