Re: Capabilities interact nicely with Substructural Types and Reactivity
"Mark S. Miller" <[email protected]>
| Newsgroups | gmane.comp.capabilities.general |
|---|---|
| Message-ID | <CABHxS9iRuFXcqMsL9jSDX4av+meOMUkFp4Qid5+bPBadPHV2aw@mail.gmail.com> |
I think the DeepFrozen auditor (or likewise the Immutable auditor of Joe-E) solve both of these quite well -- essentially by ensuring that the value in question has no mutable state. By the conventional meaning of "pure function", DeepFrozen is a looser restriction. Pure functions are deep frozen but not all deep frozen objects are pure functions. Nevertheless, DeepFrozen is adequate (and pleasant) for ensuring determinism and confinement. Likewise for Immutable in Joe-E. On Thu, Sep 12, 2013 at 9:52 AM, David Barbour <[email protected]> wrote: > Compared to languages with pure functions, such as Haskell, I do not > believe that E has "good support" for all forms of determinism and > confinement. > > For example, in E it is difficult to create large behaviors that cannot > remember their inputs... except by re-constructing these behaviors once per > input (which can be very inefficient). This regards, a form of determinism > - where we deterministically obtain the same output for the same input. > > Similarly, in E it can be difficult to enforce that a factory does not > interact with its objects after construction or prevent direct interaction > between objects from the same factory, unless of course it's a factory of > some relatively specialized 'value' types. This regards a form of > confinement - syntactically bounding influence and control over values in > the language. > > On Thu, Sep 12, 2013 at 9:00 AM, Mark S. Miller <[email protected]>wrote: > >> David, I don't think I get your point either. How would these issues >> affect determinism or confinement? E is oo ocaps with good support for both >> determinism and confinement. >> >> > _______________________________________________ > cap-talk mailing list > [email protected] > http://www.eros-os.org/mailman/listinfo/cap-talk > > -- Cheers, --MarkM _______________________________________________ cap-talk mailing list [email protected] http://www.eros-os.org/mailman/listinfo/cap-talk