Thoughts on Keybase

[Mark H. Wood]

On Mon, Jan 05, 2015 at 08:22:47PM -0500, Robert J. Hansen wrote:
> > We only *suspect* that: we saw him holding a smoking gun but did not 
> > actually see him fire it.

True.  But we have established an identity between him and a person of
interest in the case.  Investigation of that interest is going to
require some more identities ("where were you on the night of the 13th?")

> Yes, which is plenty sufficient to soothe my conscience about invasive
> measures.  If there's a homicide, ought it go uninvestigated and the
> shooter undiscovered just because we're concerned we might be invading
> the privacy of a possibly-innocent person?  I would suspect I was
> grossly misunderstanding you were it not for what you said below:
> 
> > I'm not fine with invasive anything whilst they are *only* a
> > suspect. And once you have proven guilt or innocence it matters not a
> > jot who they are.

I suspect that imprecise language such as "who they are" lies at the
root of the disagreement here.  I think there may be some disagreement
about the meaning of "invasive" as well.

> "Until you prove guilt I won't approve of any serious investigation into
> who did it or how.  And if somehow you prove guilt anyway then you don't
> need to ask these questions any more, so I still won't approve."
> 
> Okay.  Thanks.  I'm really glad you're in the minority: if I were to
> wind up murdered on a city street, I'd really hope the police would care
> enough to find out who did it and how it was done and why -- even if
> those questions might offend people's sensibilities.

Well, if a person is suspected of a crime, many of his various
identities are irrelevant.  Others may be critical to establishing
guilt or innocence.  ("But this photo of me in the Boston Globe shows
that I was nowhere near the scene at the time you say the crime was
committed.  Look at that clock behind me.")

Now, if guilt is established, that new identity matters a great deal,
since it tells us who to discipline.  If guilt is disproven then that
should be made clear to anyone who might reasonably have learned of
the suspicion.  So:

o  if guilt is proven, that is the only identity we care about
   w.r.t. the crime;

o  if guilt is disproven, then the suspect's public identities are
   relevant to publishing his innocence.

Things get murky when you consider established procedures.  If the
suspect is released, but ordered to remain available ("don't leave
town") then the police need to record and distribute established
identities sufficient to detect whether the suspect is disobeying the
order.  Later there may be a need to identify a person who is no
longer to be especially watched.

(This is why I tend to think of identification as the establishment
and maintenance of sets of mappings or labels.  I have a lot of labels
("identities") stuck on me by family, friends, enemies, employers,
trading partners, etc., each of which is more or less independent.
Various sets of these labels make up how my associates retrieve their
concepts of me.)

-- 
Mark H. Wood
Lead Technology Analyst

University Library
Indiana University - Purdue University Indianapolis
755 W. Michigan Street
Indianapolis, IN 46202
317-274-0749
www.ulib.iupui.edu
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: </pipermail/attachments/20150106/dbcb9c46/attachment.sig>