Re: Identify auto tick locations

[John W. Eaton]

On  1-Nov-2007, Francesco Potorti` wrote:

| I am reasonably sure, yes, but I am not a lawyer.  However, that's where
| copyright is different from patents and why we can have compatible code.
| If it is not copying nor modifying, but reimplementing, it does not
| infringe copyright.

Yes, an independent implementation cannot infringe on a copyright.
But one might interpret your comments as advocating that it would be
OK to copy the code then tweak it in ways that make it different
enough so as to appear non-infringing.  I hope that no one working on
Octave uses a method like this, especially if the code that is being
copied is something from Matlab.  I know we were originally discussing
gnuplot code, and I am concerned about that as well since its license
is not clear when it comes to what is allowed if you wish to
incorporate parts of it in your own program, but I am obviously much
more concerned about other things like .m files distributed with
Matlab.  My recommendation to people working on Octave has alwayas
been that they shouldn't neven even look at that code.

| Anyway, if you want to be really sure, the proven foolproof way is to
| describe what the algorithm should do in human words to someone who has
| never seen the original code and have that person implement it.

Yes.

| Moreover, for a piece of code that small, it is even questionable if it
| is worth worrying about copyright infringement at all.  For example, the
| Emacs team requests copyright assignments only if code exceeds 10-20
| lines of original code (not comments, blank lines or repeated snippets).

Yes, but that doesn't mean that it is OK to copy something 10-20 lines
at a time.

Also, I think their practice of not requiring a copyright disclaimer or
assignment for small changes is not intended to mean that it is OK to
copy 10-20 lines of code from some other source and pass it off as
one's one work.  The idea is that it is not necessary to disclaim or
assign copyright for such a small contribution.

The people working on Emacs also do not have to worry as much about
the problem of infringing to copyright of another well-known
proprietary system as they write a compatible alternative to it.

In any case, for ideas about implementing plotting features, maybe it
would be better to look to plplot, which I believe does have a
compatible license.

jwe