Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!merlin.usc.edu!usc.edu!raulmill From: raulmill@usc.edu (Raul Deluth Rockwell) Newsgroups: comp.lang.apl Subject: APL Machines Message-ID: Date: 3 Sep 89 19:34:34 GMT Sender: news@merlin.usc.edu Distribution: comp Organization: University of Southern California, Los Angeles, CA Lines: 28 I am interested in the idea of an APL machine. It is my understanding that some machines have been built along these lines, but their performance has been degraded with the overhead of checking type and rank on generic variables. I think that this difficulty could be overcome with some attention to this in hardware. What I am currently working on is "what is a good intermediate form for expressing apl". In other words, APL is a generalized array language, is there a vector language which can efficiently implement APL? (It seems to me that some intermediate form would be required for expressions, to reduce the cost of a number of operations, such as optimization, idiom recognition and execution). By the way, although I say APL, I mean something like APL0 (no state variables, such as []IO or []CT or []PP, nested arrays present, etc.) I am also thinking of including some sort of hierarchical storage structure, to give named storage (variables/functions) "operating system functionality" to the language (no--I haven't considered all the implications of this yet). Any comments or ideas?? (on APL as non-Von Neumann (sp?) hardware or as an operating system.) -- Raul Rockwell ! INTERNET: raulmill@usc.edu ! UUCP: ...uunet!usc!raulmill ! 55 mph = 82 nc U.S.SNAIL: 721 E Windsor #4, GLENDALE CA 91205 !