[e-lang] Origins of Coercion

[Toby Murray]

On Tue, 2009-02-24 at 09:17 -0800, Mark Miller wrote:
> On Tue, Feb 24, 2009 at 7:14 AM, Toby Murray
> <toby.murray at comlab.ox.ac.uk> wrote:
> > I know that coercion originated in Joule. Is there a specific reference
> > to the idea of coercion anywhere in any Joule-releated documentation?
> 
> Hi Toby, have you seen <http://www.erights.org/history/joule/MANUAL.B10.pdf> ?

Yes. Let me make some assumptions that I hope you can confirm or
disprove.

"Verifiers" are what I'm looking for. Verifiers implement a superset of
the functionality of Coercers. (Coercion is a special case of Unsealing
where the thing being unsealed is an authentic capability.)

The Joule manual states
> Joule provides the Verifier abstraction, a mechanism for certification
> built using only encapsulation and message passing.

Do you know if the Verifier implementation was anything like Marc
Stiegler's shared-variable Sealer-Unsealer implementation (e.g. as
depicted in http://www.skyhunter.com/marcs/SecurityPictureBook.ppt )
which I had thought was the first Sealer-Unsealer implementation (not
requiring EQ) that uses only message passing and encapsulation ?

If it is not similar (i.e. doesn't use a closely-held slot in which
boxes can divulge their contents only to the corresponding unsealer)
then the Joule Verifier implementation would represent a second approach
to implementing Sealer-Unsealers using only message passing and
encapsulation. 

Are there any details about the Verifier implementation from which I
could learn these answers?

Cheers heaps

Toby