>Well, the runs counter to the many sites that moved their applications >to PL/1, including State Farm (who trained 1,500 non-programmers to use >PL/1 in a 9 week course - those with musical ability, mathematical degrees, >or knowledge of two or more languages accounted for 90% of the folks who >were retained after that class!). We benchmarked SAS, PL/1, and ASM >programs >in about 1973, and saw ASM's execution was faster than PL/1, which was only >slightly faster than SAS, but the writing of the ASM program took 7 units of >time, >the PL/1 took 4 units of time, and SAS took 1 unit of time.
I was a PL/I developer around 1972. I loved it with BASED variables and STRING types. But up in Canada it really fell out of favor after about 1975. A lot of people were trying to do apps that were pretty marginal on the 370/158's of the day. So a lot of customers did COBOL or ASM, not PL/I. PL/I was reputed to have a 30% disadvantage compared to PL/I. My old employer the Bank of Montreal did its retail banking system in CICS/ASM and its Mortgage system in COBOL. When I visited the IBM Hursley Lab in 1976 I was told that PL/I had only really caught on in Europe. No doubt once they got past the "F" compiler and onto the many Optimizing Compilers things must have improved. I wonder if there is a present-day benchmark comparison? Also, is it PL/1 or PL/I? John ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
