"Tom Breton" <tehom at REMOVEpanNOSPAMix.com> wrote in message
news:m33cukk4kb.fsf at panix.com...
>angilion at ypical.fsnet.co.uk (Angilion) writes:
>> > [,alt.religion.wicca,alt.education,alt.religion cut to reduce spamming]
> >
> > On Mon, 15 Jul 2002 07:46:11 -0400, "Parse Tree" <parsetree at hotmail.com>
> > wrote:
> >
> > [..]
> >
> > >If Grace Murray Hopper really did invent the compiler, then her impact
has
> > >been far beyond that of Marie Curie, and most Nobel Laureates in
general.
> > >http://www.sdsc.edu/ScienceWomen/hopper.html> >
> > It appears that she did invent the compiler, but there does seem to be
> > some uncertainty. There is a tendency to exaggerate the achievements
> > of any woman who becomes famous in a field, and the exaggerations
> > become assumed as fact without the checking that would usually
> > occur. For example, I have often read that Grace Hopper invented
> > COBOL, which is not true.
>>> The concept of the "first compiler" is shaky. The very first
> computers were programmed directly in machine code. (toggles, then
> punch cards) Later they were programmed in Assembler, which is a sort
> of 1-to-1 mapping of text to machine code. (I oversimplify a lot
> here)
>> C compilers, which are clearly compilers, also mapped fairly simply to
> machine code, moreso before the modern C standards and without
> optimizations. That's why C has the {pre-increment, post-increment,
> pre-decrement, post-decrement} instructions - those instructions were
> in the PDP instruction set. There are even simpler compilers than
> that, eg "Small C", which leaves out the more complex C constructs.
>> Which is to say, both an assembler and a compiler generate machine
> code (Generally speaking. Sometimes compilers generate other stuff,
> eg Java virtual machine code). Both input text and output machine
> code object files. The difference is all in the mapping they use.
>> So how complex does the mapping have to be to be a "real compiler"?
> To call something the first compiler, then, you have to draw an
> arbitrary line as to when it's source-to-machine-code mapping is
> complex enough to count.
>> But one conclusion I can draw clearly is that whoever stated that
> Grace Hopper invented the first compiler couldn't have looked hard at
> the question.
>> Caveat: Everything I've said here is simplified. But the point still
> holds.
>
Well said, Tom!
We also know that those who claim Hopper "invented" a compiler, don't know
what they're talking about, because of their use of the term "invented". A
more appropriate term would have been "developed", since there were many
people involved in developing software which other people used as a template
for their own contributions.
It might also have been a more credible claim if they'd said that a woman
who worked in industry "developed" "one of the first compilers", rather than
claiming that a public servant "invented compilers". It's inevitible that
many other programmers at IBM, and Sperry, and Honeywell were using
"compilers" that they developed individually, who never considered them to
be any kind of a major breakthrough. Several months ago, the head of the
computer center at my university told me that the mainframe computer I
learned Cobol and Algol on is still in operation--along with the exact same
compiler I developed, without any changes. Was that a "breakthrough" of
some kind? Absolutely not. But if a woman had done it, she'd probably be
president of the university now.
John Knight
