Re: language design implications for variant records in a pascal-like language

From: <compilers@is-not-my.name>
Sent: Thursday, 13 January 2011 10:41 PM


> It's interesting that PL/I, which is a really nice language and
> was available since the early 1960s never caught on, nor did it
> replace any COBOL or FORTRAN,


Well, it did. However, FORTRAN programmers couldn't perceive that the
language was of any benefit to them.


In general, programming was out of their depth.
They failed to see any advantage in the fact that when
their PL/I program crashed they could get the statement number
where it crashed (instead of a hex error number). Or perhaps they were
ignorant of that.)
Three great facilities in PL/I were not seen as improvements
over FORTRAN, namely,
(1) dynamic arrays ;
(2) variable field widths in formatted output; and
(3) character strings.


From published code, it is evident that Fortran programmers went to
extraordinary lengths to make their code portable and flexible, in an
attempt to emulate dynamic arrays, for example, often doubling the
size of the code in the process, and even then the finished product
did not come close to what could be done with PL/I in terms of (1)
portability (2) bullet-proofing, and (3) ability to update.
[PL/I suffered from much less mature compilers than Fortran. Back in the
late 1960s, Fortran H produced great code, PL/I F produced pretty bad code.
By the mid 70s the PL/I optimizing and checkout compilers were a great
improvement, but nobody was interested in switching to PL/I then. -John]

Return to the comp.compilers page.
Search the comp.compilers archives again.