Subject: Re: [open-source] [Fwd: [icecast-dev] Xiph.org announces Vorbis Beta 4 and the Xiph.org Foundation]
From: "Stephen J. Turnbull" <turnbull@sk.tsukuba.ac.jp>
Date: Wed, 28 Feb 2001 11:26:32 +0900

>>>>> "kms" == Karsten M Self <kmself@ix.netcom.com> writes:

    >> The ASF's main focus is to keep the web "free" and avoid
    >> proprietary lock in. We do this for the HTTP protocol, and we
    >> do this for the XML language, for XSLT, FO, SVG and all a bunch
    >> of other open standards.

    kms> Different tack.  AFS does for a *protocol* what FSF does for
    kms> *software*.  Good aim, good goal, good work.  Different
    kms> objective.

What is the difference?  Seems to me that a protocol is a
specification for behavior of a software implementation, whereas a
program is a specification for behavior of a hardware implementation.
When you remember that most of the popular high-level languages today
(Lisp, Java, Perl, Python) actually all target virtual machines, so
that really a program is a specification for behavior of a software
implementation, well....

    >> Why am I saying this to you? well, your opinion on the BSD
    >> license is notorious and normally indicates that the BSD
    >> license is, somewhat, less-free than the GPL. While this might
    >> be the case for lots of software, there are cases where BSD
    >> means more freedom than GPL.

This can only be true "tactically" as I understand RMS's argument.
GPL supporters tend to see freedom as a matter of reducing the amount
of non-free software they cannot avoid using.  BSD supporters tend to
see freedom as a matter of increasing the capabilities that can be
achieved by using free software.

While these goals coincide when the market share of a given
implementation is small (Ogg-Vorbis), they evidently do not when the
market share is large.  For example, for free operating systems today,
AFAIK GCC has essentially 100% market share.  However, would there be
basically one C compiler suite available if GCC were BSD-licensed?

I doubt it.  Maybe the C/C++ compiler would stay unified.  But I
rather suspect that several of the "poor relation" front ends would
split off into competing proprietary implementations, possibly using
proprietary versions of the back end.  I think this would be
especially true for say parallelizing FORTRAN versions, and maybe even
Java.

Now, I think that the chance of having The Best parallelizing FORTRAN
compiler available anywhere running on Beowulf/Linux clusters would be
worth having a proprietary implementation of the GCC compiler suite,
assuming that one niche application were as far as it went.  I rather
suspect RMS would find that tradeoff unacceptable.  

There is also a difference in analytical style to remember.  The point
may be moot for RMS, because he may simply deny that the hypothetical
limitation to a single niche is too implausible to discuss.  But BSD
advocates often do think in those terms.

    kms> The BSD-Apache license is incompatible with the GNU GPL.  I
    kms> believe this is the primary objection RMS has to it.

No.  That is his primary _tactical_ objection (as I interpret what he
has said).  That is, in the short run it splits the community.  He has
correctly emphasized this objection because division of the community
is morally unacceptable (freedom of software is freedom to share),
division is the greatest hindrance to the growth of free software, and
because cooperation is something all free software and open source
software advocates see as an essential strength.

However, _strategically_ he objects to it as "lax".  In the long run
he believes that lax licensing will lead to proliferation of non-free
software.

    >> But people should not be scared-away from the BSD license just
    >> because the concept was not invented by the FSF.

    kms> You're misrepresenting the position of the FSF.  It cites
    kms> licenses which are acceptable within its own objectives.
    kms> BSD-Apache isn't.  MIT and BSD-revised are.

No, he's not mis-representing the position of the FSF as I read him,
because he's not representing the position of the FSF.  He is pointing
out that the FSF presents its own position, that freedom of software
is a basic necessity for human beings and that copyleft protects that
freedom most effectively, very strongly.

On the other hand, RMS has discouraged investigation into alternative
interpretations of freedom (eg, "maximize volume of free software," as
I have proposed) in very strong terms.  This leads many who do not
make the effort to think as accurately about the issues as he and you
do to believe that MIT-style licenses are somehow "less free" qua
license than copyleft.

I'm sure RMS deplores this situation, but he would be crazy to try to
dispel this illusion, since (a) the mere effort would create great
confusion about what he believes is good for the community (KISS), and
(b) the illusion tends to increase support for copyleft, which is a
good side-effect of an otherwise unfortunate situation (ignorance on
the part of many in the community of what he really means).

-- 
University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
_________________  _________________  _________________  _________________
What are those straight lines for?  "XEmacs rules."