[cap-talk] Limiting cases for capabilities

[Jonathan S. Shapiro]

It occurs to me (belatedly) that the case *against* capabilities
consists of a fairly short list of "use cases": problems people think
they want to solve but believe cannot be solved.

There is a dual set of bad use cases for ACLs. Perhaps we should
generate that list.

The problem I foresee is that in the ACL case the problems are all real,
whereas in the capability case many of the alleged problems result from
a desire to achieve something impossible. That being said, it is not bad
to acknowledge that there are real issues to be addressed in the
capability world. It hurts our credibility if we claim to solve all of
the worlds problems but obviously don't do so.


shap