Subject: Re: Sun to free Solaris.
From: "Stephen J. Turnbull" <turnbull@sk.tsukuba.ac.jp>
Date: Wed, 26 Jan 2000 13:47:14 +0900 (JST)

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

    kms> Scott "Free Beer" McNealy is taking half measures.

Sure.  Top of Mt. Fuji is only halfway to heaven, but it's a grand
view all the same.

The important fact is _not free beer_.  This is _semi_-_free speech_.

I'd expect you, Karsten, to remember that this is still new to many of
the developers, definitely the managers, and especially the lawyers
inside corporations.  They don't have the benefit of decades of rms's
sermons, flamewars on gnu.misc.discuss, and so on to draw on.  You
know the story, I needn't go on.

And demands from the free software community that these stirrings of
opening be immediately extended to fully freeing the software reek of
self-interest.  I don't see why anybody with demands on their time
would pay much attention to them, unless they were accompanied by
irrefutable accounts, preferably with anecdotes about success stories,
of how freeing the software benefits Sun now.

Sun is not in the position of Netscape; they can afford to follow the
strategy Brian Bartholomew proposed of releasing now without making
promises and then following through with full freedom later.

    kms> I'm given to understand that Sun will be releasing Solaris,
    kms> with source code, and free of charge, under at least certain
    kms> conditions.

This means that lusers get all the _direct_ benefits they could ever
expect from open source.  They've cut open the hood, yeah, baby!  The
indirect benefits of encouraging development, that they lose.  (But
maybe not all of them, q.v.)  Still, Sun's attempt to capture as much
of the returns from future development as possible is natural.

And I believe misguided.  But that doesn't mean it's simply marketing
hype; there are real benefits to Sun's action.

    kms> Worse than all that, Sun's actions just don't make much
    kms> sense.  Offering a free (beer) OS really doesn't do much for
    kms> either end of the server OS market.

No, but opening the source for free does.  I can tell you what some of
the Solaris users on the XEmacs lists will think about this.  They
will be able to participate more actively in fixing OS-related bugs,
since they'll have access to the OS source.  Oh, joy!  And some of
those bugs do turn out to be OS bugs.

It empowers the user/developer.  How often does a vendor say "sorry,
your code is buggy, the OS is fine?"  How often do you think they'll
say that if there's a risk of getting a reply "here's the patch,
fool"?  And if they do say it and get that reply, the chances are
infinitely greater that they'll apologize and implement the patch than
if source is not available!

This is really not all that different from using a FSF-assigned GPL
application with a cathedral style of development, as long as Sun
maintains it well enough to prevent any thought of a fork.  (Except
that presumably code cannot be borrowed for unrelated or competing
projects, a big loss, but not a total loss.)

And it stiffens the competition.  I would bet that even if you can't
take the source directly, you can learn stuff from the Solaris sources
that would be applicable to Linux development, etc.  How about device
drivers?  Even if Sun owns the Linux port, they may very well be
willing to allow them unrestricted redistribution as modules (a first
step to truly freeing the Solaris code iself).  Or they might be
convinced to GPL them, since they are probably going to be doing some
of that kind of porting themselves for Linux anyway.  (And if they
were stupid enough to write a broken version for Linux to make Solaris
look good, that would be immediately obvious.)

A risk for the Linux developer, that her code could never be released
as part of Linux?  Sure.  But as rms tells us, programmers code
because the project is there, not because of the benefits.  This will
discourage, but not prevent, such "transgenic" activity.  And maybe
Sun _will_ allow distribution of modules, or even free that code.

Compared to the status quo, it's all improvement, and not negligible,
as far as I can see.

    kms> ... may well alienate the free software movement at worst.

<FLAME MODE="sterilize virus">
So much for those folks who forgot to get their rabies vaccination and
are now paying the price.
</FLAME>

This is the FSB list, can we look at the costs and the benefits, now?

Am I for free software?  Yes!

Do I think Sun should release Solaris under GPL?  Yes!

Do I look this gift horse in the mouth?  Yes!

And when I see it's missing a few teeth, do I kick in the ones that
remain?

No!  I say "Thanks, Scott!" and "I'll take a look at it, but I already
know that I'm going to have the following problems:  (a) my own[1]
portfolio of kernel drivers are already GPL and so can't be ported,
(b) ..., (c) ...".

And several somebodies with more prominence in the field than me
should hold a press conference and say "We welcome Sun's announcement.
Opening the Solaris technology to the industry's and public's view is
a great thing.  Etc, etc.  BUT ...

... independent developers are not going to come flocking; people who
are currently using Solaris will be less likely to jump ship for Linux
or *BSD but that drain will continue for (x), (y), and (z) reasons...

... and so this action probably isn't going to do much for Sun.  We
hope that Sun will continue down this path and release a Free Software
version of Solaris, preferably making it their flagship OS offering.

Here's what that stategy does for Sun...."


Footnotes: 
[1]  Nonexistent, of course.  :-)

-- 
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."