[cap-talk] ... enforcement - hope? Capabilities as clumsy, not

[Ben Laurie]

David Wagner wrote:
> By the way (this is an unrelated tangent), I find it interesting that
> object capabilities don't seem to require the full set of features that
> most programmers today seem to associate with object-oriented languages.
> For instance, object capabilities don't need subclassing, inheritance,
> virtual methods, interfaces, etc.  I think the notion you need is an
> 'object' which binds together code and state, along with static or
> runtime type-safety to ensure that object references cannot be forged,
> that the object can only be accessed through its exported interface (e.g.,
> the methods specified by its code), and that the state (data) associated
> with an 'object' can only be accessed by its code and not by other code.
> This is a much smaller substrate to build upon.  Did I get that right?

This sounds pretty much exactly like what I was trying to do with Python 
(before it became obvious that this was futile), so certainly I think it 
is right.

You also need no globals, btw. Or, at least, no writable globals.

-- 
ApacheCon! 13-17 November! http://www.apachecon.com/

http://www.apache-ssl.org/ben.html       http://www.thebunker.net/

"There is no limit to what a man can do or how far he can go if he
doesn't mind who gets the credit." - Robert Woodruff