Re: Tuning Karatsuba in LibTomMath

From: Tom St Denis (tomstdenis@yahoo.com)
Date: 04/22/03 

From: tomstdenis@yahoo.com (Tom St Denis)
Date: 22 Apr 2003 06:37:17 -0700

Douglas Wikström <douglas@sics.se> wrote in message news:<3EA513B5.C36F1AEE@sics.se>...
> Dear Tom,
>
> I do not intend to start a flame war either, but why don't you try to
> explain to Torbjörn Granlund why their code is bad? or ask them to help
> you? or you could help them fix the problems in GMP? or do it better
> yourself without critizising the effort of others? There are several
> constructive alternatives to writing posts like yours on sci.crypt, some
> of which I believe would let you draw from the experience of some
> talented people.

I'd say my previous post was not overly rude. I really do think the
GMP source code is horrible to look at. Some reasons? Of the top of
my head is the over use of macros and #ifdef's throughout the code.

As to me helping them fix the problems... Um, in case you haven't
noticed I'm writing my own bignum library! I barely have time to work
on it let alone help out other developers with theirs.

As for "drawing from others experience" I'd have to say that's a
loaded statement. Not that there are not a couple of people here who
are nice [Colin Percival and Scott Fluhrer come to mind] but for the
most part people either lurk or reply with trivially useless
information.

You might reason "my bad attitude" is why people don't help but then
why did 743 people visit my site last month producing a total of
27,125 hits? I'd say in the number theory circles [since I doubt many
joe-blow users will download my libraries :-)] I'd say that's not bad.

So obviously some small circle of people use my code and an even
*smaller* circle actually contribute back.

All I wanted was someone with an idea or two to chime up to help fix
the tuning application. Eventually it will probably happen [law of
averages and all]...

In the mean time I'll go back to giving away free software that people
seem to like...

Tom

Relevant Pages

  • Re: Messenger Spam Pop-Ups
    ... I dont think it is even a fix, ... <Tom Della-Latta> ... :: Surely you are liable/negligent if you don't take the time to learn to ... ::: Messenger even when messenger is closed and you don't ...
    (microsoft.public.windowsxp.security_admin)
  • Re: Scanned photo looks "fat"
    ... May be not a fix, but a test to see if the scanner is really expanding the ... It must be a true circle. ... Print the Tiff image at 300 DPI. ... You can compare the original scanned paper against the printed image. ...
    (comp.periphs.scanners)
  • Re: sharing information in a class
    ... but I can't quickly explain how to fix it because there is no easy fix. ... Here you create a circle, it's the variable called circ in the method ... One basic problem is that the object you create in ... I think Victor was right, you need to do some studying and probably you need ...
    (comp.lang.cpp)
  • Re: Delorme SA2006 vs MS Streets and Trips 2006
    ... >>>receiver but adequate for driving the moving map. ... >>either would never get a fix, or it would take half an hour. ... I will say, that once the Delorme got a fix, it wouldn't lose ... >Tom, years ago when I had SA-8 or possibly it was before that the ...
    (rec.outdoors.rv-travel)
  • Re: Unauthorized Laser Illumination Event
    ... just bringing it up on this group has created an awareness that ... I still don't know if I would circle ... getting a fix on his position for later use. ... which was easy given the left hand turn. ...
    (rec.aviation.piloting)
Quantcast