CCL: Filters

 Sent to CCL by: Brian Salter-Duke []
 On Thu, Oct 27, 2005 at 06:17:25PM +0200, Eugen Leitl wrote:
 > Sent to CCL by: Eugen Leitl [eugen(-)]
 > On Thu, Oct 27, 2005 at 12:13:30PM -0300, Cory Pye cpye{} wrote:
 > > It is comments such as this that could convince the Gordon group to
 > > to switch from available source to binary distribution (or none at
 > I'm not sure this would change a lot. The code is purely complimentary,
 > as no one can change a line of it and share it but within their group.
 This is techically correct but it is not the whole story. Mike Schmidt
 will accept bug fixes and sometimes he accepts new code. Have you asked
 him ever to accept changes that you want? I have on many occassions and
 he has always been extremely friendly and helpfull. For example, the
 hooks in the latest July Gamess version to our VB2000 code are much better
 because Mike took my changes to three gamess modules and then spent time
 improving my code.
 If you do not like the build process, which I agree is prehistoric,
 write a makefile for your own use and then ask Mike whether he wants to
 distribute it. He may, I do not know.
 There is another point about code release. It may not be "free", as in
 free software foundation definition, but you can look at it and see
 if it really does what it claims to in all circumstances and you can
 check how it does it. That sure beats "binary only" releases.
 > > Why are people complaining about getting something for free? If you
 don't like
 > Because being merely free (as in beer) is not enough. The license must
 > offer the freedom of being able to contribute, and, preferrably, to
 > fork (under a different name).
 > > aspects about the distribution the package of group X, write your own
 > > open-source package, or improve upon someone elses. Any volunteers to
 write an
 > I would gladly start hacking on GAMESS, such as improving the build
 > process (which is frankly prehistoric), and porting to gcc4.0. The license
 > doesn't allow me to. No other package under a less restrictive license
 > plays in the same league. I'm sure there are lots more of users which
 > share this frustration.
 > > efficient integral package? OK. Uh-oh, it works on Suns and IBMs, but
 not on
 > > SGIs or HP. It works with RedHat, but not with Mandrake or Debian. I
 can't find
 > If you want it to work on Debian, change the license.
 > > optimized BLAS libraries. How come it core dumps every time I try to
 do an
 > > f-type integral? Version 4.2.3 of compiler X on machine Y has a bug.
 Are you
 > > willing to provide support for this free program in perpetuity? In
 > The point of open source under a nonrestrictive license is that the user
 > fix the bugs they find, and offer support. The GAMESS mailing list is such
 > a place.
 > > most mutations are non-viable.
 > If somebody fails to support a mutated fork, of what concern is this
 > to the authors maintaining the original code line? If they don't
 > fork, but contribute to the main tree, there are submission privileges
 > and regression unit tests which will quickly weed out the weak.
 > > Sounds to me like downloading from Kazaa and then complaining about
 the audio
 > > quality.
 > It sounds to me as if you don't understand the impact of a license
 > on a fate of a project.
 > The license alone does not grant a blessing. But no open source project
 > will be successful without a good license.
 > This is getting off-topic for the list, so no more in this thread from me.
 > --
 > Eugen* Leitl <a href="";>leitl</a>
 > ______________________________________________________________
 > ICBM: 48.07100, 11.36820  
 > 8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE>
          Brian Salter-Duke (Brian Duke) b_duke]~[
        Post: 626 Melbourne Rd, Spotswood, VIC, 3015, Australia
     Phone 03-93992847.
 Honorary Researcher, Chem., Melbourne Univ. & Med. Chem., Monash Univ.