Re: Capability Authorization-enabled Decentralized Object Model [was RE: The "Verifiable" Economy [was RE: a few thoughts about zcaps]]

Christopher Allen <ChristopherA@lifewithalacrity.com> wrote:

>
> I’d love to see just that narrow scenario, as a series of narrated slides
> saved as a video. I really don’t like the car keys example, and this one is
> practical example of the problem with an example of failure & an example of
> success leveraging ocap.
>

I have a slide deck.  Now all I have to do is find it.

--------------
Alan Karp


On Wed, Apr 7, 2021 at 11:03 AM Christopher Allen <
ChristopherA@lifewithalacrity.com> wrote:

> On Wed, Apr 7, 2021 at 10:55 AM Alan Karp <alanhkarp@gmail.com> wrote:
>
>> I believe that any system that can handle the example in
>> https://www.hpl.hp.com/techreports/2008/HPL-2008-204R1.pdf is general
>> enough.  (It won Best Paper at ARES 2010, so you don't think I'm blowing
>> smoke.)
>>
>> In that scenario, Alice wishes to use Bob's backup service, which relies
>> on Carol's copy service.  We show that authentication-based schemes can
>> lead to bad results.  For example, Alice specifies a file she is not
>> allowed to read, and Bob specifies a file he does not have permission to
>> write. Nevertheless, Carol reads the input and writes it to the
>> output, potentially destroying one of her files.  These problems cannot
>> arise with capabilities.
>>
>
> I’d love to see just that narrow scenario, as a series of narrated slides
> saved as a video. I really don’t like the car keys example, and this one is
> practical example of the problem with an example of failure & an example of
> success leveraging ocap.
>
> — Christopher Allen [via iPhone]
>
>>

Received on Wednesday, 7 April 2021 21:01:39 UTC