Deceptive comment in toom33_mul.c

Martin Guy martinwguy at
Fri Sep 13 22:51:51 CEST 2013

On 13/09/2013, Niels Möller <nisse at> wrote:
> Itai Bar-Natan <itaibn3141 at> writes:
>> This isn't really a bug

That depends whether you regard programming primarily as a one-way act
of command from the lone programmer to the machine to achieve a
desired effect

or primarily as a form of communication from one human being to
another, to explain the problem and the algorithm proposed to solve
it, so that they can become collaborators in checking that the
following code does indeed perform the desired transformation and
achieve the intended results.

Hot air, I know. And the specific case is minimal. But I do appreciate
the attention to detail.

In my own long-tempered conception of programming, the commentary says
what the following code should do, then the code (hopefully!) does
what the commentary said, something close to Knuth's "Literate
programming" paradigm. The practical result is that if the code
*doesn't* do what the commentary says, then some years later the
difference will stand out to an attentive reader - OK, the code was
wrong but the theory was right - enabling them to fix the code bugs.

so very much +1 for fixing mistaken commentary!


More information about the gmp-bugs mailing list