[cap-talk] capabilities for semidocuments? Capabilities for OOmethods?

[John Carlson]

>Have you looked at E in a Walnut
>
>http://www.skyhunter.com/marcs/ewalnut.html
>  
>

Okay, I read throught some of this.    I'm not sure I
understand what a partial capabilities is yet, in E, although it is
likely it is possible.  The problem I am trying to solve
is when the active object trying to gain access must
exercise two facets on separate objects before gaining
access to an third object.   I think this can be done by having
three objects, one facet on object Alice that contacts
object Dave and lets Dave know that Bob used a facet,
if Dave discovers that Bob has already contacted Carol,
he can let Alice know that, or introduce himself directly
to Bob.  Swap Alice and Carol to handle the other case.

I am somewhat concerned that Dave must keep track
of what Bob has been doing.  This seems like a lot of
overhead.

If I draw the parallel to keys in the real world, there
are four doors into a safe.  Two are only one hallway
and on a parallel hallway there are two doors.
The first door on hallway one is opened by the Alice key.
The first door on hallway two is opened by the Carol key.
The second door on hallway one is opened by the Carol key.
The second door on hallway two is opened by the Alice key.

The problem I see with this is that you could have a whole series of
doors to go through, and they could be opened in any order.
An example of this I have is engineering approvals, where a design
has to go through 5 different people before a design is released.
Each person is assigned a task to approve the design.  The tasks
can be approved in any order.

In this example, there is a route object which keeps track of the
tasks.  In the the case above, the route would be Dave.

So I think this is doable.   I guess the system I work on has these
kinds of capabilities.  I just never called them capabilities before.

It would be interesting to see workflow or objectflow integrally
tied to capabilities.  I think promises may do that.

John