[cap-talk] Wikipedia: Object-capability model - reference vs. capability?

David Hopwood david.nospam.hopwood at blueyonder.co.uk
Wed Jan 17 12:20:03 CST 2007

Charles Landau wrote:
> At 5:38 PM +0000 1/17/07, David Hopwood wrote:
>>Charles Landau wrote:
>> > I need to know whether a reference to the KeyKOS number object is a
>> > capability or not.
>>It is not, in the obj-cap terminology.
>>> [...] In KeyKOS, I can hold this reference in a capability register
>>In the obj-cap terminology, those are registers that hold references.
> In your view, is the number 7 on which the "add" instruction 
> operates, the same data object as the number 7 in a number 
> capability, or a different object?

Mu. The "add" instruction operates at a different level of abstraction to
operations on references. In particular, 'the number 7 on which the "add"
instruction operates' is not an object at all.

> In the context of a language such as E, variables can hold references 
> to either data objects or non-data objects. So defining a capability 
> as a reference to a non-data object seems to me to be the wrong 
> concept in either context. In E, you don't deal with capabilities, 
> you deal with references.

Yes, and that's why MarkM's thesis uses the term "reference" far more
than it uses "capability" (in this sense).

David Hopwood <david.nospam.hopwood at blueyonder.co.uk>

More information about the cap-talk mailing list