[cap-talk] keyrings and bootstrapping capabilities

[Kevin Reid]

On Jun 7, 2007, at 19:30, Pierre THIERRY wrote:
> Scribit Kevin Reid dies 25/05/2007 hora 14:26:
>>> all possible policies need to be known  at compile time.
>> What kind of policy do you imagine being defined at run time? I  
>> doubt that message-restricting facets will be; runtime behavior is  
>> more about *which object* is provided in some position than *how  
>> restricted* that object is.
>
> Just take a look at your Pascal's triangle: not counting facets  
> with all methods or no one, an object with 3 methods has six  
> possible facets, 4 methods give you 14 possible facets, 5 methods  
> 30 facets, and so on...
>
> So there's definitely a reason to define such facets at runtime, it  
> seems.

Those are message-restricting facets, as I meant the term.

I do not claim that all possible message-restricting facet  
definitions should be created at compile time, but that few possible  
message-restricting facets are useful, so there is no necessity to  
have a general system for computing them, either at compile- or run- 
time.

Furthermore, many useful facets are not describable as message- 
restricting; they have nontrivial code, or are fundamental to the  
original component (e.g. read and write ends of a queue/pipe).

Broadly: defining a facet is a kind of programming, usually part of  
some larget program being written, and so there is no reason  
facetization needs to be *more* dynamic than the rest of the system.

-- 
Kevin Reid                            <http://homepage.mac.com/kpreid/>