On Sep 18, 12:22 pm, glen herrmannsfeldt <g...@[EMAIL PROTECTED]
>
wrote:
> Frank McCoy wrote:
> > In alt.folklore.computers "John H. Lindsay"
>
> (snip)
>
> >>NotPL360. The latter was conceived by Niklaus Wirth
[snip]
> One could probably have used the PL/I (F) preprocessor with PL/360.
> I think you can get it to write out the processed file.
> (If not, it writes it to a temporary file where you should be
> able to get it out.)
No you wouldn't want to do this.
One of the wonderful features of PL360 was that you could compile and
produce a cross-referenced listing faster than IEBPTPCH could produce
just a paginated listing. You certainly wouldn't want to add the
overhead of the PL/I preprocessor to that. And on MTS, which didn't
have a link-editor, you just ran the resulting object file, so PL360
was almost like an interpreted language in terms of the edit-compile-
run loop.
Anyway, there was a general-purpose macro language available at the
time: GPM (whose descendant, m4, still lives in Un*x systems). The MTS
operating system had UMIST (and, later, Eureka). Some grad students at
the University of British Columbia added macros to PL360. That's how
they discovered that MVCL was slower than a loop of MVCs. (BTW, PL/I
"optimizing" compiler also used a loop of MVCs rather than MVCL.)
- peter
|
