[cap-talk] Is EQ necessary for PetName systems?

[David Hopwood]

Jonathan S. Shapiro wrote:
> In the past, my understanding was that a pet name was simply a
> convenience name for a capability. That is: there is a mapping from pet
> names to capabilities, but no mapping from capabilities to unique pet
> names.

The desirability of having a mapping in that direction is that, if we
have an arbitrary capability for which the user might have a petname, we
can display the petname to the user if it does.

Note that, as I said in my other post, this only requires the mapping
to work for objects that could be a target of a petname (and reliably
fail otherwise). If not all objects can directly be such targets, then we
have not really lost any generality in the petname system -- just added
the slight inconvenience of sometimes having to refer to the intended
target via a "handle" object.

-- 
David Hopwood <david.hopwood at industrial-designers.co.uk>