ech@spuxll.UUCP (Ned Horvath) (05/20/85)
I am at a loss to explain what would bother someone about PL360. At the time I was using it (early 70's) I considered it an excellent alternative to assembler. It was a bit curious in the sense that you had to KNOW the architecture of the 360 (nonsense like multiplies requiring even registers and leaving the result in an even/odd pair, funny stuff about R0, etc.) in order to use it -- the sort of thing only assembler programmers ordinarily know. Given that prerequisite, however, it was MUCH more convenient to define data and code than assembler was. Perhaps it was too nuts and bolts for HOL programmers and too effete-looking (ye gods, block structure!) for died-in-the-wool assembler types... But I suspect that what killed the likes of PL360 was advancing compiler technology. A lot of us learned C using the Ritchie compiler on the 11, and it didn't take too many 'cc -S' runs to convince you the thing wrote HOT code; folks who ought to know tell me that Bliss is alike in this: you can still touch the hardware when you need to, but the rest of the time you can pretty well ignore the underlying architecture. The best PL360 story I know is indicative of a greater lesson, germain to this discussion: Al Demers was unhappy with the 'roff' version then running at Princeton, and the Comp Center folks weren't giving him any satisfaction. So, he coded a roff end-to-end in a week or so, in spitbol -- a great prototyping tool for such a task, albeit too slow for large documents. Having made his design errors in a cheap prototype, he proceeded to recode a (now much better understood task) in PL360 -- again, in a week or so. The result was a much-enhanced roff facility that ran like a bat. =Ned=
mash@mips.UUCP (John Mashey) (05/22/85)
Ned Horvath writes: > But I suspect that what killed the likes of PL360 was advancing compiler > technology. A lot of us learned C using the Ritchie compiler on the 11, > and it didn't take too many 'cc -S' runs to convince you the thing wrote > HOT code;... A supporting example is LIL (Bill Plauger's Little Implementation Language at BTL in the early 70's), which has about the same semantic level as PL/360, and never quite caught on. Bill once gave a talk in which he said that LIL was a language whose time had come... and gone, because every time he thought he had a niche where it beat C enough, Dennis improved the compiler. -- -john mashey UUCP: {decvax,ucbvax,ihnp4}!decwrl!mips!mash DDD: 415-960-1200 USPS: MIPS Computer Systems, 1330 Charleston Rd, Mtn View, CA 94043