"Numerical Recipes in C" code

[Dann Corbit]

Hi All,
I am a begineer in all these topics and was never aware of so much
danger in copying IP.
Please be assured that I wont use this code for developing any big C
application for any business. Its just that I wanted to run some code
of that book for I was trying to solve matrix equations using Gauss
Jordan method. I have run the code and I have made some modifications
to it to suit my requirements. After reading the above discussion and
after having shown it to my friend who runs a big business of his own,
he explained to me how important are these facts.So, I have completely
disposed off that material except for two functions which I have used
for matrix equations.
Now, can I keep atleast those two functions? My friend says keeping
just the two functions since I have already modified it and since I
wouldn't be indulging in such activities henceforth shouldn't do much
of a problem.
He has agreed to buy the product for me and I have already placed an
order for it. Must be getting it shipped in a day or two.
Thank you and sorry for taking it lightly.Infact, I took it a bit too
lightly itself.

If you buy a copy of the book and use the code as described under the
license agreement, then I can't see any problems.

I suggest instead use of some of the excellent codes that are better
constructed and far more lenient in license.

For instance the cephes collection for scientific functions that you lack:
http://www.moshier.net/#Cephes

Atlas for linear algebra:
http://math-atlas.sourceforge.net/

etc. There is no need to violate anyone's intellectual property rights.
Generally speaking, for any sort of application you might like, someone will
have written a package that does the job with very little restrictions on
use. When an author creates a tool set, it is only right that you follow
the author's explicit instructions (and any implicit instructions implied by
law). Imagine if you wrote a package. You might want to get paid for its
use. You might want to get credit, but let it be used freely. If someone
has spent a few hundred or a few thousand hours working on something, it is
only just and right that we abide by their wishes in exactly the same manner
that we would like to be treated ourselves.

Further, if you have specific questions about numerical software, probably
is a better location to ask because it is focused
on doing mathematics with computers.

I own a copy of all of the Numerical Recipies books for C and C++ and I
think that the text is above average but the code is below average. The C++
code is much better than the C code, which is sub-par.

I think that you should familiarize yourself with the meaning of software
licenses. You should know what GPL, LGPL, Berkeley, Apache, etc. sort of
licenses mean. If someday you will work for a sofware company it is crucial
that you understand this. You could expose your company to a billion dollar
lawsuit trying to save $50 if you act incompetently.

Sourceforge:
http://sourceforge.net/

is my home away from home. Learn to use the resources that are at your
fingertips.

And if you have questions about the C programming language come here.

 

[Walter Roberson]

(e-mail address removed)-cnrc.gc.ca (Walter Roberson) writes:

The earliest publication date I can quickly find for the
Numerical Recipes books on amazon.com is 1986. Because patents
only last 17 to 20 years, it is unlikely that there are any
remaining material in the original Numerical Recipes that could
infringe a patent.

I'm a bit fuzzy on patent renewals. I did find a clear statement
that in the USA patents can only be renewed up to the original term,
but in my chasing I wasn't able to figure out what the whole
"patent maintenance fee" was all about, and found one uspto reference
that I interpreted as hinting that patent maintenance fees are no longer
in use in the USA.

The copyright holders appear to be Cambridge University Press, in the UK;
I didn't try to follow up to examine UK patent law.

The OP was asking specifically about Numerical Recipies in C, which was
published in 1988, 18 years ago, which was before the USA changed the
way it determined patent longevity. If I found the correct summary
sections, in that case any applicable patent would last either
17 years after patent approval or 20 years after patent application,
but I misremember now whether that was "the earlier" or "the later"
of the two dates. As the C programs might in theory use different
patents than the programs in the original book, there is a legal possibility
that there is a remaining patent or three involved in the NR in C book.
I don't really think there is [especially as I see no patent numbers
listed in the preface], but I cannot currently rule the possibility out.
The 2nd Edition (more likely to be what the OP was looking at) was 1992
so changed programs could plausibly have a few years of patents yet.


Of course, this whole patent discussion is predicated upon the
unproven premise that some court somewhere might rule that a
translatation or rewrite of the programs was sufficiently distinct
from the original to no longer be a "derived work" (since, after all,
copyright is explicitly not able to protect -ideas-, only expression of
them, so there must be -some- boundaries as to what might be
deemed "derived".) One of the other posters named what is probably
the easiest solution: use one of the other already existing GPL or
license-free numerical libraries.

 

[Richard Bos]

If your case ever goes to trial, then the judge and/or jury should
find in your favor.

I would not put money on this. I would not put money on the opposite,
either, but I see no reason to be sure of this.

And you can keep those two functions you modified.
And you can someday possibly work for Jack Klein.

I would not put money on this, either. I, for one, would not employ him.
Regardless of whether he is or is not in the right, a predictable and
very expensive lawsuit is not something you want as an employer.

Richard

 

[Your Uncle]

Abhi:

If your case ever goes to trial, then the judge and/or jury should
find in your favor. And you can keep those two functions you modified.
And you can someday possibly work for Jack Klein.

All of this is predicated on, of course, you showing us a receipt for
that supposed purchase of yours :^)

I'm having deja vu all over again. I think one is well-advised to have a
conjectural judge judy on his shoulder to consult when one conducts himself.
Period. Do you know what Judge Judy would think of ':^)' as lying
elaboration fills her courtroom? cheers, furunculus