loc@yrloc.ipsa.reuter.COM (Leigh Clayton) (02/09/91)
-------Message from Bob Bernecky, ex I.P. Sharp Associates, fresh------------- -------from a birthday and not at all embarassed about remembering------------- >no. 4915571 filed 3.12.12 fri 8 feb 1991 >from rbe@ipsa > >@transferred from ipsa no. 4706653 filed 3.12.09 fri 8 feb 1991 > >The 360/40 did indeed (assuming I am properly guessing the >antecedent message which I did not receive...) have a >microcode RPQ (Request Price Quotation -- IBMese for non-standard >addon) for APL. I think most of what it did was syntax analysis, >which typically accounts for about 15-30% of APL interpreter CPU >time. > >A similar model of the 370/145 had a similar facility. In spite of >its good performance( I recall it ran about 30 times faster than >a naked interpreter for scalar-bound functions [which of course >aren't that common...]), IBM never pushed it on other models. > >It was basically a product of IBM research, where many good ideas >are developed and discarded. I think it was Hassitt and Lyon who >did the work. Lemme check some old ACM APL conference proceedings. > >Hmm. Can't find any references. I'd check the following in your >library: ACM SIGAPL Quote Quad Conference Proceedings, early 1970's >to mid-1980's. Also check IBM Journal of Research and Development >for the same era. > >What was the original question? > >Bob Bernecky >Snake Island Research Inc. > ----------------------------------------------------------- loc@tmsoft.UUCP uunet!mnetor!tmsoft!loc loc@ipsa.reuter.COM (Leigh Clayton)