[cap-talk] Value of 'copy on write' as attenuation mechanism.

[Neal H. Walfield]

At Mon, 26 May 2008 01:41:05 +0200 (CEST),
Rob Meijer wrote:
> So basically would it be fair to state that COW is a composition pattern,
> not an attenuation pattern?

Assuming Unix's semantics, yes, I agree.  Implementing Mach's MAP_COPY
semantics requires interposing on all writes to the underlying object.

> I'm still a bit confused about this, as coming
> from a read/write base, COW would feel like an attenuation to me,

If it is not the user who is implementing it in terms of a read-only
facet, then it requires added functionality on the server.

Neal