[cap-talk] Capability-based Projects - theory vs. practice

Jed Donnelley capability at webstart.com
Fri Aug 3 04:53:28 EDT 2007


At 07:32 PM 8/2/2007, Mark Miller wrote:
>On 8/2/07, Jed Donnelley <jed at nersc.gov> wrote:
> >  However, I note that if we take that tact then we significantly
> >  thin the ranks of 'capability' systems.  By that criteria I think
> >  we must eliminate all the language based systems, Emerald,
> >  Network Objects, and E just as examples that come to mind?
>
> > 1973: Actors - MIT *******
> > 1985: Emerald - University of Washington *********
> > 1990: Joule - Los Altos ????? (Los Altos??)
> > 1995: Original-E - Electric Communities *****
> > 1996: W7 - MIT *****
> > 1997: SLK/J-Kernel Cornel ?????
> > 1999: E Language *****
>
>For at least Joule, Original-E, and E, their persistent protection
>state was also simply capabilities, and so should be marked with a
><<<<. For Actors this is also true in theory, but I don't know if this
>was ever embodied in persistent practice. For the others on your list
>above, I don't know what their persistence story was, if anything.
>It's a neglected but important question.

I was hoping you'd challenge the lack of persistence (better term
than durability I guess?) for E MarkM.  That allows me to ask
(or try to ask) what persistence means in the language context.

For an operating system I well understand the "persistence" of
files and directories (even processes, but that's another story),
and the ability to associate a 'user' with a base or home
directory.  Once you have that starting point, any movement
of capabilities (e.g. insertion of a capability into a
home directory) will still be there after a system restart.

For the language context (e.g. E) is there a similar
notion?  Maybe this question isn't well enough formed
to be answered, but please give it a try if you understand
what I am getting at.  Can you show me, for example, how
actions in E can result in shared access of an object
between two "users" that persists after a system restart?
Is something required at the OS level to make such
persistent object sharing work?

--Jed  http://www.webstart.com/jed-signature.html 




More information about the cap-talk mailing list