[Toybox] informed that my mathematics would not be used

Rob Landley rob at landley.net
Mon Mar 12 12:13:21 PDT 2018


On 03/09/2018 03:57 PM, CM Graff wrote:
> I have been working on an arbitrary precision library for the toybox
> project so as to support `bc' for the past half of a year. To expedite
> the project I recruited Gavin Howard, Bao Hexing, Zach van Rijn and
> Adrian Parvin D. Ouano and also received some help from Graham Watt
> and Tuomo Venäläinen.
> 
> Gavin Howard was recruited to implement the actual bc language. After
> many months of complaints from him and dealing with his suggestions
> for API redesign and at the very end stages of the mathematics library
> completion -- Howard has announced that he will no longer be using the
> rest of our work and has in two weeks written his own arbitrary
> precision implementation which he will be using for bc instead.

And given that the submission I just got is over 9000 lines, I'm likely to
change the result until it's unrecognizable in the usual
https://landley.net/toybox/cleanup.html multiple passes way. (Only probably more
of them than usual because 9000 lines.)

> As a student and part time free software volunteer, after having
> implemented algorithm D division, karatsuba multiplication and
> floating point details -- I don't have time to fight this type of
> Machiavellianism any longer or whine about my loss of time so I am
> respectfully stepping down.
> 
> The rest of the team will continue the research branch of our library
> in private (leaving the finished bc type aside and working on
> IEEE-754-2008, fixed-point and rational support).

I pointed somebody doing the bc stuff at https://bellard.org/libbf/ last week
(Fabrice Bellard just wrote an arbitrary precision floating point library, which
is MIT licensed but maybe if I asked I could get permission from him to use it
under 0BSD), and they said it wasn't compatible with what they were doing.

(99.99% of everything seems to be happy with either 64 bit math or double
precision floating point. There's not usually a lot of call for finding digits
of pi.)

> Thank you for your time,
> Christopher M. Graff

I'm sorry things didn't work out. Thanks for getting the process started.

Rob



More information about the Toybox mailing list