Mike Miller wrote:
> On Fri, 23 Feb 2007, rwh wrote:
> 
>> Mike Miller wrote:
>>> On Fri, 23 Feb 2007, Justin Krejci wrote:
>>>
>>>>> The only guy I know and talked to uses BSD as the base for their
>>>>> products (and this is not a small embedded company) *exactly* because
>>>>> of the GPL and license requirements to distribute source code.
>>>> Yup, BSD based is the way to go.
>>>
>>> Not for the individual licensing his code, but for the individual who
>>> wishes to use code written by someone else.  If you want to contribute
>>> something to the world, you definitely should not choose BSD for your
>>> license.  The GPL will do much better in promoting the software and
>>> leading to further development.
>>
>>
>> That seems a bit simplistic. How many places/projects do you think
>> might want to use GPL but don't care to participate in the legal
>> vagarities revolving around GPL?  If I license a commercial library I
>> pay cash and agree to restrictions on distribution. If I go the GPL
>> route I don't pay cash but I do have to buy into the social goals of
>> the GPL and accept some uncertaintly about IP liability.  If I go the
>> BSD route I only have to accept some uncertainty about IP liability.
> 
> You are writing from the perspective of the individual or group using
> the code written by someone else.  I agree with you (see above), but the
> individual who wrote the code is the one who should prefer the GPL to
> the BSD license.

I'm afraid that I don't accept that GPL is socially optimal to BSD. If
the GPL license inhibits the use of the code I'd rather use BSD, and if
someone grabs my code and finds a way to make a buck off it, more power
to them. The only time I get irked is when someone takes someone else's
free code, uses it in a proprietary manner, and then attempts to claim
ownership of the original concept. I think MS was recently caught trying
to file a patent for a particular interface used to teach object
oriented programming that was essentially copied from someone else's work.

[snip]

>> I'm a bit surprised by your comment. I doubt there are any
>> restrictions on your published papers that would prevent me from
>> extending what you've done (in principle anyway, practice is a
>> different matter) and using it in a strictly commercial setting. That
>> is essentially what the BSD folks are doing.
> 
> Every published paper has a copyright held by the publisher, but I don't
> think that is relevant.  The thing you might not be getting is the
> difference between using code to accomplish something versus using code
> as part of another program.  It is fine to use GPL code to, say, design
> a car and then sell the car and give no credit whatsoever to the
> developer of the GPL code.  It is not fine to use GPL code within a
> program that you then distribute under a different license.  So people
> might use some of my papers to discover genes that affect diabetes of
> blood pressure or whatever, and they owe me nothing (but they might cite
> my papers).  On the other hand, they are not free to use a paragraph
> from my paper inside of their paper (not without quotation marks and
> correct attribution).

I guess I don't distinguish between code and words - they're both means
of conveying IP.

Suppose you identify a genetic sequence that serves to identify
diabetics who are at particular risk from the effects of high blood
pressure. Unless you are able to protect that IP through a patent, there
is nothing to prevent me from using that information to market a test
for that sequence.

The philosophy behind GPL would allow you to require that anyone making
use of your discovery would need to reveal the associated IP that they
merged with your information. There would be nothing to stop you from
getting a patent and imposing those conditions on anyone using your
discovery (the Biological Open Source movement is one example,
http://www.bios.net/daisy/bios/home.html)

I guess the question is would you publish (like BSD) or go BiOS (like GPL)?

> 
> 
>> Two small corrections to some of the stuff I've seen flying by in this
>> thread. The code most people are referring to when they talk about MS
>> using 'BSD' code is the network stack. As I recall this predates
>> either GPL or BSD licenses. I believe it was put into the public
>> domain as a result of AT&T's failed suit against the UCal Board of
>> Regents distribution of the original BSD code which was based on
>> AT&T's Unix code.
> 
> It might predate the most recent version of the BSD license that removed
> the requirement to repeat the "Regents" line all the time.  Do you have
> a source for that?  I couldn't find anyting about it.
> 
> Another possibility is that many of us are misunderstanding the BSD
> license.  Look at this:
> 
> http://www.groklaw.net/article.php?story=20070114093427179
> http://bsd.slashdot.org/article.pl?sid=07/01/15/1757235&from=rss
> 
> I'm not so sure he has that right.  Distribution of a copyright notice
> is a bit different than distribution of a license, but IANAL.
> 

O'Reilly has a detailed history of BSD at
http://www.oreilly.com/catalog/opensources/book/kirkmck.html It looks
like the spirit of the BSD license goes back to the early 80's.

> 
>> It was also my impression that Apple's OS X was based on the same BSD
>> licensed by NeXT from Berkeley rather than FreeBSD, although parts of
>> OS X were ported from FreeBSD. I'm also pretty sure that Apple had
>> been posting back enhancements to FreeBSD - Wikipedia mentions the
>> Base Security Module being ported back from Apple's implimentation.
>> They were also contributing patches and features from Safari back to
>> KDE's Konqueror - at least until there was some conflict in priorities
>> between Apple and the KDE folks.
> 
> That's nice of them, but the license is still such that OS X source is
> not available to us.  I guess the message is that the BSD license
> doesn't force people to stop contributing code, but it doesn't force
> them to contribute either.
> 

The O'Reilly article mentions that after the lawsuit was settled BSD
split into a -Lite version and an -Encumbered version, the latter
requiring a commercial license. One of the reasons that folks like Sun
and Apple didn't worry too much about the whole SCO drama was that they
have commercial source licenses. Apple may not have any requirement to
release any source code.

But my impression is that a good chunk of the OS is available from
Apple, but the code for the interface isn't. I thought that Darwin was
the underlying operating system with Cocoa sitting on top of that.

http://www.opensource.apple.com/darwinsource/

> I did not distinguish correctly between the BSD variants and said
> "FreeBSD" when it was something else - but the point is really that the
> license does not force Apple to share their source.  OS X is making
> Linux a little less attractive and it is thereby slowing Linux
> development. That's how I see it, and it makes sense, but that isn't the
> sort of thing you can prove.
> 

I use both and they each have their place. I feel very strongly about
setting open standards and then let the market decide. I don't mind
paying Apple for OS X since I think I get a good value for the money.
I'm less sanguine about paying for Office (even with the academic
discount) or Oracle, and my home is a Microsoft-free zone.

Have a nice weekend.

--rick