Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!rutgers!cmcl2!acf5!mitsolid
From: mitsolid@acf5.NYU.EDU (Thanasis Mitsolides)
Newsgroups: comp.lang.prolog
Subject: CP, Linda and Multiple Producers.
Keywords: cp, linda, prolog, stream
Message-ID: <755@acf5.NYU.EDU>
Date: 4 May 89 23:30:17 GMT
Organization: New York University
Lines: 30


I believe that the ability for one language to simulate features
	of another easily, is very important and a good indicator
	of the language's relative power.

At the same time I believe that 
	it is important that the implementation should be reasonably efficient.

A simulation of Linda's features in CP, appeared in this group
	which was inefficient.
	In fact the complexity of executing n Linda operations seems
	to be n*n.

It seems to me that an efficient implementation of Linda on any
	concurrent logic programming language would require
	a "linda monitor" but also lots of "merge" invocations.
	Thus, making that Linda Implementation almost impractical.

I believe that one of the problems that CLP languages must face
	in the future is the ability to handle
	multiple producers (of a stream of messages),
	in a simple, efficient and fair way (direct messages to objects?),
	avoiding the ever needed (often unfair) explicit merge invocations
	and the inefficient n*n solution (above).

That should extend the (practical) expressiveness of CLP languages.

Please correct me if I am wrong.

Thanasis