Re: Context and privacy

On Wed, Jun 24, 2009 at 12:52 PM, Ronald P. Reck<rreck@rrecktek.com> wrote:
> I have been thinking about the use cases and how one might exert any control
> over situations where information is moved from its original context and put
> into a new one.
>
> 1. what bearing does the old context have on the new one. For instance, if I
> provided some information about myself on twitter, if this same information
> is then placed on Facebook, how does the fact that it originated on twitter
> really matter?
>
> To me, it seems that information I provide would be broken down into
> assertions (maybe its just the RDF in me). These assertions would have
> accompanying metadata. It would be up to the rules or stipulations of the
> new context how that metadata was brought to bare.
>

Could you clarify what you mean by "assertions" here please?
Given some piece of information "X":
Is it like "X is mumble"?
or, is it the fact that "X is generated by Ronald"?
or, is it more like "X was generated subjected to rules A and B (in
context C)", "Rule A is blah" and "Rule B is blah blah"
Or have I gotten it completely wrong? :)

It would be great if you can give an example.
(side note: I am toying with the idea of attaching proofs to RDF
triples on how they were derived. So, this stuff really interests me!
:) )


> I conceive of people as instantiating a role when they provide information.
> So another way of asking the previous question is how does the role I am
> instantiating have a bearing on the new role?
>
> In secure networks, information can go from a lesser secure environment to a
> more secure environment, but going from a more secure environment to a less
> secure environment is difficult. If one were to extend this to our situation
> one might postulate that information can go from a less restricted situation
> to a more restricted situation easier than the converse. I think the rules
> of information promiscuity in the new context need interpretation in terms
> of the previous information promiscuity rules.
>
> Maybe something like, only provide this information to people I explicitly
> allow, or provide this information to others unless I explicitly denied them
> access. Does anyone else agree or disagree?
>

Yes, changing the context gives rise to a privacy concern.

FWIW, it has been shown that it is possible to make use of the rules
governing the information in the original context (for example the
information category, the purposes which it can be used for, the
people or agents who have access to that, etc) and model these rules
to suit the new context by what is known as "Data Purpose Algebra"
[1]. The premise here is that you can derive a new "purpose function"
to suit the new context. While you cannot really exert any control on
the information that is being moved around, this data purpose algebra
method can be used to *attach* a suitable privacy policy which relates
the original context to the new one.

[1] Data Purpose Algebra -
http://dig.csail.mit.edu/2006/Papers/Policy07/data-purpose-algebra.pdf

-- 
Oshani

Received on Wednesday, 24 June 2009 19:48:43 UTC