[cap-talk] Firefox breaks the principle of identifiability

[Ian G]

David Wagner wrote:

>But I think Ben Laurie has gotten at a much more serious problem.
>Suppose I ask my bank how much money they'd be willing to put behind
>their word when they introduce me to Paypal.  I think they're going
>to laugh at me.
>  
>

The example chosen was somewhat stylised.
In practice, we don't get the introductions
we desire, unless we pay for them (and even
then ....).  Nobody owes us that trust, no matter
how much we think it would solve our security
problems if they did.

>Moreover, I think they may be right to laugh at me.  They're in the
>business of managing money, not in the business of introductions (that's
>the Yellow Page's business).  Why would being good at the former imply
>that they are good at the latter?  To put it another way, sure, I trust
>them to do the former, but why should I trust them to do the latter?
>I think we've got an example of a standard fallacy when reasoning about
>trust: "I trust X for purpose P" doesn't imply "I trust X for purpose Q".
>  
>

Right.  In practice we get out introductions
from a variety of sources, and we apply
discretion and due diligence to them as they
come in.  No biggie.  The main thing here is
to identify the source of the introduction, it
is not the agent's job to do/measure trust,
it is the user's job to make a trust judgement.

In that sense, the use of the word trust outside
of "Alice decides whether she and she alone
trusts X" should be treated with suspicion.  I'm
glad this is starting to be understood:

http://www.financialcryptography.com/mt/archives/000333.html

iang

-- 
News and views on what matters in finance+crypto:
        http://financialcryptography.com/