[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: GPL'ing Solaris won't save SUN

From: Richard L. Hamilton
Subject: Re: GPL'ing Solaris won't save SUN
Date: Tue, 24 Aug 2004 23:24:34 -0000

In article <>, (Mike Cox) writes:
> (Richard L. Hamilton) wrote in message 
> news:<>...
>> In article <>,
>>      Paul Eggert <> writes:
>> > At Tue, 03 Aug 2004 02:53:39 -0400, Donn Miller <> 
>> > writes:
>> > 
>> >> My guess is that Sun will do like Netscape and Apple did, and
>> >> create their own license
>> > 
>> > Sun has already done that.  Several times.  They have a complicated
>> > licensing repertoire that I'll bet not one in 10,000 professional
>> > programmers understands.  They've got the Sun Community Source Licence
>> > (SCSL).  They've got the Sun Industry Standards Source License
>> > (SISSL).  And they've got the Sun Public License (SPL).  Sometimes it
>> > seems like they use a different license for every product they turn out.
>> > 
>> > And (as far as I know -- it's hard to keep track) all their special
>> > licenses are incompatible with the GPL.  That's Sun's privilege, of
>> > course.  But it's terrible marketing.
>> > 
>> > Say what you like about the GPL -- but it's clear, and lots of
>> > customers understand it, and it does the job.  It's serving Sun well
>> > in larger efforts like OpenOffice and in smaller ones like
>> > evolution-jescs.  If Sun yet again comes up with their own
>> > incompatible-with-GPL license, they'll have yet another uphill battle
>> > convincing the world to pay attention to it, much less understand and
>> > use it.
>> > 
>> > Here's the acid test: can I take a module out of Sun's open-source
>> > distribution, modify it, and redistribute the result freely as part of
>> > (say) my Linux kernel?  If I can do that, then Sun's open-source
>> > license will be useful to me and to lots of other people; if not, then
>> > I'm not sure it's worth Sun's time to come up with Yet Another Sun
>> > Software License.
>> It's useful to me just for troubleshooting, even if I never compile a
>> line of it.
>> It's also useful as a massive collection of examples, of drivers,
>> filesystems, etc.
>> But the bottom line is not that of the ideologue, or even the contributor
>> and participant; it's that of the investor.  Unless there's a cash value
>> answer when one of them asks "what's in it for me?", it shouldn't be done.
> Then why the hell should a linux advocate even care if SUN is going to
> Open Source Solaris?  If a person who is not a SUN engineer/employee
> cannot make changes to the CVS or take code and put it in another OS
> (aka BSD license), then it has no value.

While I disagree that it has "no value" under those restrictions, and am
not necessarily proposing those particular restrictions, I don't
personally give a damn _what_ a "linux advocate" cares about.  And for
turnabout, as long as Linux drivers are GPL'd rather than LGPL'd, I see
the advantage giving away everything in exchange for a little.  I see the
point as improving the value of Solaris to Sun's customers, not improving
all other free software everywhere at the expense of Sun.

> Heck, even MS took a bunch of BSD TCP/IP code and used it in Windows. 
> In terms of commercial value, BSD is the best license for commercial
> companies because they can use BSD code in commercial products without
> revealing their source, the GPL is best for goodies because everyone
> is forced to share.  The Solaris Open Source License will have nothing
> going for it apparently.

So you know what it's going to look like, that you can judge it to be
useless?  You must have some sources that nobody else has, like a
time machine to look into the future.

Personally, I think they ought to do a BSD-like license for drivers,
create new uncommitted (in the sense that it would be understood that the
interface might change in future versions) documentation of some internal
interfaces (like the vfs/vnode interface needed to create new
filesystems) and license the documentation similarly, and for everything
else, create a license that allows only licensees to share changes, and
only when they're going to be used with Solaris.  Also, there should be
a mechanism to move parts from the more restrictive to the more free
if/when it becomes clear that doing so provides more value to Sun by
supporting their customers better.

But that's just my opinion; I have no connection with Sun other than as
someone who uses their products (and has no control or inflence over any
money other than for the very small amount of stuff I might spend for home

> I think this will be a case of "What if they opensourced a product and
> nobody cared" type of thing.   Everyone who wants a true enterprise
> Open Source solution will just continue to go to IBM and use Linux on
> the mainframe.  AIX is  being gutted for Linux, and if Solaris code
> isn't avaliable to be gutted and put into Linux, then SUN won't have
> an enterprise level Linux OS on their bread and butter machines. 
> Customers will see that, and avoid SUN enterprise hardware.  SUN needs
> to see that no one cares about Solaris or any other propriatery UNIX. 
> They have paid for propriatery UNIX for decades, and it is now a
> commodity.

You're putting ideology ahead of practicality again.  Stallman does enough
of that for everyone (although the alternative perspective is useful even
if only by contrast).

> If SUN is going to Open source Solaris, the *only* viable license IS
> the GPL, for the REASON that all the enterprise code could be put INTO
> LINUX.  Yes, I know that sounds counter-intuitive, but that is the
> only way Open sourcing Solaris would work, because then the Linux
> people could port Solaris goodies into Linux and then Linux would run
> on every expensive highend SUN box.

Aha, the bottom line.  "One OS to rule them all."  Sounds like the
talk of the Dark Lord, whether he hangs out in Redmond or in geekdom.
The analogy goes further - if any one with sufficient power succeeded
in dethroning the current occupant of that title, they'd simply end
up becoming just like them, whatever their intentions.  That much influence
is inherently corrupting.

No one OS can be, or ever will be, perfect for all situations (however
much shrinkwrap application vendors [who mostly _aren't_ into Open Source])
might like that.  And if any one thing took over everywhere, it would get
weaker over time simply due to lack of competition.  Nor do separate Linux
distros entirely answer the need for competition - AFAIK, none of 'em do
what enterprise customers want, namely pay decent attention to backwards
compatibility, and be driven by customer needs rather than what attracts
the interest of the programmers.


Lasik/PRK theme music:
    "In the Hall of the Mountain King", from "Peer Gynt"

reply via email to

[Prev in Thread] Current Thread [Next in Thread]