[ih] What research capability-based OS was on the ARPAnet?

[Clem Cole]

On Tue, Dec 6, 2022 at 3:52 AM John Gilmore via Internet-history <
internet-history at elists.isoc.org> wrote:

>
> For a time, I ran across an interesting research machine that was based
> on "capabilities".  It was well documented for its time, and I learned
> about it by using it just to explore, but I never programmed for it.  By
> now I've forgotten its name and who its inventor was.  Do any Internet
> historians remember it?


There were a couple, but did you mean CMU's C.mmp - which ran Hydra, which
was capabilities based and was on the Internet as CMU-D or CMU-Hydra? The
system influenced the Intel 432 BTW, which was also based on capabilities.
  Bill Wulf and his co-authors Roy Levin and Sam Harbison have a book about
it: "HYDRA/C.mmp: An Experimental Computer System" (ISBN 0-07-072120-3).
After all their work on building Cdot and Hydra, the book is the report.
It has my favorite dedication of any book in the research world, which
says:  "To builders and programs of real programming systems."

The follow-on for Hydra was StarOS for CM*, which was also
capability based, but I don't think it was ever directly connected to CMU's
IMP, only on the LAN, so it's unlikely you would have had access to it.
 FWIW:  Henry Levy has a book from Digital Press called '*Capability-Based
Computer Systems*' (ISBN 978-148-3101064), which has a pretty good survey
of most of them, so if Cdot/Hydra was not it, there is a good chance Henry
describes the system in his book, which I think is available via google
books. [I have printed copies of both]


Clem
ᐧ