On Sun, Dec 23, 2012 at 5:37 PM, Henry Baker <hbaker1 at pipeline.com> wrote:
> There are variants of Lisp in which cons cells are
> _functional_/_immutable_, in which case whether the list structure is
> copied or not can't be determined....
>
Certainly. In which case, as you say, the OP's problem goes away.
More to your point, there are lots of massively parallel computers in which
> arguments for functions are "marshalled" and copied to another processor's
> address space using, e.g., "MPI" (Message Passing Interface); these systems
> obvious "deep" copy their arguments.
>
> There are some Lisps that were aimed at computers like this....
>
Do these Lisps have deep copy semantics for *all* argument passing and
assignment? Do they also have mutating operations (rplacX)?
> And then there's the whole "MapReduce" system developed by Google, which
> also incorporates all kinds of redundancy to enable recovery from failures
> of various kinds....
Not sure how that's relevant here.
-s