Re: The History of the ALGOL Effort

In article 06-09-050, Henry Spencer
<henry@spsystems.net> wrote:


> Peter Ludemann <p_ludemann@yahoo.com> wrote:
> >One of the nice properties of the PL360 compiler is that it could
> >compile a program, produce a listing, and cross-references faster than
> >IBM's IEBPTPCH utility could produce only a listing.
>
> This experience had wider ramifications, too: PL360 compiled so much
> faster than IBM's glacially-slow linker could link, that it soured Wirth
> on separate compilation. The result was several influential languages,
> most notably Pascal, with no provision for separate compilation.
>
> (Some of the early papers describing Pascal say fairly explicitly that
> this feature has been omitted partly because a well-written compiler
> is so much faster than linkers that you don't need separate
> compilation.)
>
> And that, in turn, contributed to the rise of C, because the fact was that
> people needed and wanted separate compilation.
> --
> [Back in the 1970s at Dartmouth, the DTSS compilers were also so fast that
> they didn't have a linker and for most purposes, didn't even bother to
> save object code. They finally wrote a linker in about 1976 when they
> added PL/I. -John]


      In the 1960's there was watfor [WATerloo FORtran], a fast fortran IV
compiler that compiled directly into executable code as well. Not good
for a huge problem, but it was a boom to all the 'student like'
programs that were encountered by universities at the time. sys360's
original E level fortran compiler was of the same philosophy.

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