octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: rcond, condest, and a block 1-norm estimator


From: David Bateman
Subject: Re: rcond, condest, and a block 1-norm estimator
Date: Sat, 24 Nov 2007 22:26:38 +0100
User-agent: Thunderbird 1.5.0.7 (X11/20060921)

Jason Riedy wrote:
> And David Bateman writes:
>> 1) Relicense under the GPL. I kept the original license as a comment,
>> but perhaps that should be removed as well.
> 
>> + ##     * Neither the name of the University of California, Berkeley nor the
>> + ##       names of its contributors may be used to endorse or promote 
>> products
>> + ##       derived from this software without specific prior written 
>> permission.
> 
> I'm pretty sure the above chunk should remain...  It's an
> additional restriction explicitly allowed under the GPLv3.
> Obviously the combined work can be GPL because the licenses are
> compatible, but that's not really a relicensing so much as an
> implication of the individual licenses (inclusion in the build
> and test process is more than mere aggregation).  But I'm only a
> math/CS guy, not a lawyer (thankfully).

I kept the original license in place, so I think this issue is
addressed.. Though perhaps its better to move the original license just
under the GPL. I think that yes it is a relicensing in that any changes
that are then made to the file are under the terms of the GPL and not
the BSD. In fact even my changes might be considered GPL with the patch
as it stands.

> IIRC, the explicit LAPACK license has a similar non-endorsement
> clause.  For a long, long time, LAPACK had a "fuzzy" license...
> We finally made it explicit and simple.  But people still ask how
> LAPACK is licensed.  sigh.  Even more fun explaining that parts
> of the reference BLAS now are in the public domain (pre '78).

Yes well LAPACK license was always a bit fuzzy in the past. Looking at

http://www.netlib.org/lapack/index.html

there is no license page easily available, and you have to go to the
user guide to find it.. A good link to the license from the front page
might help people understand the license better.

>> 4) Change the name of block_onenorm_est to onenormest as something
>> shorter. I hesitated to make this an internal function, but thought that
>> a 1-norm estimator is useful in its own right.
> 
> I mostly use the 1-norm estimator for my work, so I'd hate for it
> to be "internal".

Thats what I thought :-)

>> 5) Replace the block of code [...]
>> [L, U, P, Pc] = splu (A);
> 
> Cool, didn't know about that one.  I mostly use a hacked-up
> TAUCS, but I need out-of-core (and software extended precision).

You can actually use just "lu" above rather than "splu" as "splu"
overloads "lu" for sparse matrices..

>> Are you happy with the proposed changes? Can you suggest any further
>> changes?
> 
> Sounds great.  Thanks for all the work!  Now we just wait for
> users to find bugs. ;)
> 
> Jason
> 



reply via email to

[Prev in Thread] Current Thread [Next in Thread]