On 09/ 8/10 03:56 PM, William Stein wrote:
On Tue, Sep 7, 2010 at 12:51 PM, Jason Grout
<jason-s...@creativetrax.com> wrote:
On 9/7/10 2:35 PM, Bill Hart wrote:
Hi all,
After really a lot of people from Sage related projects and elsewhere
wrote to
me and said they'd support a BSD licensed bignum library, I finally
decided to
go ahead with creating such a library.
Could you comment on the relationship of this with GMP/MPIR? Is it a
competitor, but with no shared code?
Thanks,
Jason
It seems that nobody responded to Jason's post. But my understanding
is also that bsdnt is a competitor to GMP/MPIR, with no shared code.
However, note that the second one enters the BSD-licensed space,
e.g., almost the whole world of scientific Python code, EPD, etc.,
then such a library basically doesn't have much competition, and
GMP/MPIR are irrelevant.
Bsdnt will be relevant to Sage, since it (1) distracts developers from
working on GMP/MPIR, and (2) attracts new developers to work on bignum
code, which can get later included in GMP/MPIR/Sage.
-- William
Having two independent code bases, which can do the same thing, does allow for a
degree of verification that's impossible with just GMP/MPIR. Even if you compare
results to Mathematica, they likely share the same code. But another library,
with totally independent developers would permit some useful checks. Any
interesting results can be compared with two systems - it adds weight to a
scientific paper if you can show the results don't depend on one bit of software.
That said, it does dilute efforts. But then plenty of similar projects have
different developers - think of all the databases.
Dave
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
