udi%wisdom.weizmann.ac.il@CUNYVM.CUNY.EDU (Udi Shapiro) (04/29/89)
Nicholas Carriero, in his message of April 26, 1989, fails to see the point of embedding one concurrent language in another. This is probably because he is more interested in promoting a particular language, rather than understanding the relations among languages. Language comparison is a difficult task, with no agreed upon methods. There are many "soft" methods for comparing languages, and we have seen them being used in arguments in the sequential world ad nausia. No reason why we should make better progress with such arguments in the concurrent world. However, I know of one objective method for comparing and understanding the relations between languages and computational models: embedding ("compiling") one language in another, and investigating the complexity of the embedding. Carriero implicitly implies that embedding one language in another is always such an easy and trivial task, that showing its possibility for two particular languages is hardly interesting. This is far from being true. Try embedding Linda in Occam, in a parallel functional language, or in an actor language, and see what a mess you end up with. Similarly, try to embed these models (and some concurrent logic language) in Linda. I doubt if the results of this exercise will help you promote Linda, but they will certainly increase your understanding (and ours, if you would share the results with us), of the relationship between Linda and the rest of the concurrent world. Carriero finds concurrent logic programs "impenetrable". I do not think that his intellectual capabilities are inferior to those of the hundreds who can compose and comprehend concurrent logic programs (some subset of which have proved this capability to my satisfaction in exams, exercises, and programming projects). Hence I believe that this is because he has not invested the time and energy required to pass the hurdle of familiarizing oneself with a radically different computational model. To better substantiate his arguments in the future, or simply for the joy of it, he should do just that.