[FRIAM] Why not ACTORs?

[Marcus G. Daniels]

Dale Schumacher wrote:
> Around the same time that FORTRAN and LISP were developed, another 
> model, the ACTOR model, was proposed.  Hewitt's "Viewing Control 
> Structures as Patterns of Passing Messages" started the ball rolling,
> If that is the case, then why don't we have well-developed actor-based 
> languages?  What fatal flaw in the model am I missing? 
I don't think it really requires a language.  All that's needed are 
spawn/synchronize primitives, and then the rest is following the 
programming model.   In a number of those papers, they basically use 
Lisp to demonstrate the model. 

I'd say it's just a problem so many programmers having deeply ingrained 
habits for assigning stuff to mutable variables, and thus there being so 
much (C/Java/Fortran) legacy code that does that.    Even in C with gcc, 
there's now a notion of vector variables.   Just have to decide to use 
that instead of a `for' loop. 

A recent take on "Hey, transactions facilitate parallelism!", is 
Software Transactional Memory... Meanwhile Lisp-like languages like R 
work very neatly with MPI with simple, but high level operators like 
`apply'.

Marcus