On Sun, Nov 30, 2025 at 6:19 AM Mark S. Miller <
eri...@gmail.com> wrote:
>
>
>
> On Sat, Nov 29, 2025 at 8:22 PM Matt Rice <
rat...@gmail.com> wrote:
>>
>> On Sun, Nov 30, 2025 at 12:44 AM Alan Karp <
alan...@gmail.com> wrote:
>> >
>> > cap-systems
>>
>> I prefer cap-systems too,
>>
>> > cap-theory sounds like capabilities aren't useful
>>
>> Whenever cap theory is used, I've always preferred the use of
>> capability discipline instead.
>> Not for the 'sounds like they aren't useful' instead because I feel
>> like if something is presented as a theory,
>> there is a formalism that comes along with that and as such it should
>> be presented alongside or referencing the actual theorems. It also
>> seems to imply there is a single unified theory of capabilities, where
>> there really seems to be variation between systems and substrates.
>
>
> I agree that I don't like cap-theory for the website name. That said, I do think there is a single unified theory of *object-capabilities*, which I tried to state precisely in my dissertation. As with other unified theories, there can be multiple concrete formalizations that emphasize or ignore different aspects. So "single unified theory" need not imply a unique formal expression of the theory. The precise theory as stated in my dissertation has indeed been formalized in many different ways.
>
When I said that I was mostly considering differences like password
capabilities, distributed vs single system and whether or not
revocation is supported. These have implications which affect the
power such as forgable or unforgeable, so I have some resistance to
considering them unified in that sense but certainly sharing many
foundations.
I agree that your dissertation gives such, what I meant to convey is
that I feel that the weight of cap theory behooves the levels of rigor
that formalisms such as your dissertation provide. Only rarely have I
seen "cap theory" used when reaching those same levels of rigor.
Simply because the above stated diversity of power, mean you and I
might have different versions of forgeable or revocability when we
both discuss "cap theory".
I didn't intend to say that e.g. your dissertation doesn't provide it.
On the contrary, just that when cap theory is used it often isn't held
that level of formalism that your dissertation does provide.
I hope that clears it up
> There have also been many things called "capabilities", most infamously "posix capabilities", that are not the capabilities we mean. Are there meanings of capability that we do want to include that are outside object-capabilities?
>
Not that I wish to include.
> To view this discussion visit
https://groups.google.com/d/msgid/cap-talk/CAK5yZYjNVQ-Z%3DJtMvGZ6qvBFnOTP4bDT0hDdzPJg_CsD2X%2B%2BTw%40mail.gmail.com.