[seqfan] Advice for creators of b-files

[Neil Sloane]

I discovered a Pari program from 2014/2018 that would not run on my 2015
Pari complier.
Antti helped me solve the problem.  Now read on:

On 3/9/19, Neil Sloane <njasloane at gmail.com> wrote:
> Antti,
>
> Of course I considered that possibility.  But 2015 is not very old!
> OK, I will upgrade my Pari.  Thank you very much for helping.
> But imagine if you got into a 2015 car, and it would not start

Or wouldn't work with 2019 brand gasoline!

But that's the problem with most of the programming languages! Many of
them also have habit of "deprecating" big parts of themselves a few
years after.

This is one of the reason I add to each b-file I submit this "standard
preamble":

# Computed 2019-MM-DD in GP/PARI CALCULATOR Version 2.9.4 (released)
amd64 running linux (x86-64/GMP-6.1.2 kernel) 64-bit version,
compiled: Dec 19 2017, gcc version 7.3.0 (Ubuntu 7.3.0-1ubuntu1),
threading engine: pthread
# with the following program:
#
# Axxxxxx(n) = {my program code};
# and the terms:
1 1
2 5
etc.

So, if somebody later finds the data incorrect in some terms, and it
is not a simple mistake in my code, but something intricate in some of
the PARI-functions, then it might even be possible to find all the
same mistakes from other b-files also (provided they have information
about the version number of PARI, or mutatis mutandis, Mathematica or
Maple).

Me:  I think this is an excellent idea if you are using anything more than
the most basic Pari (or Maple, MMA, etc) commands.

Start the b-file with a comment saying what version of what software it was
created with.