[Simh] Pontus asks Is [the] BSD [license] liberal enough?

[Clem Cole]

​I fear you might be missunderstanding/confusing the "usage" rights and the
"ownership" rights.​



On Sun, Jun 7, 2015 at 4:00 AM, Pontus Pihlgren <pontus at update.uu.se> wrote:

> If A gives B a software product under GPL then A must provide
> source code to B upon request. B has no obligations to anyone,
> not even if B modifies the code.
>
​Sadly - that is not correct.  If B modifes the code, then B is also
obligated to make it available - period.  That's the virus and what make
its powerful as well as difficult.



>
> A third party, lets call them C, has no rights to the product or
> source, it's just when B gives the product to C that the
> restrictions apply. And even then B has no obligations to A. Any
> modifications made by B is only available to C upon request.
>
​I wonder if you are getting stuck on the "upon request" part.​  If B
modifies the code, the truth is anyone can ask B for the sources to the
"dervitive work."  Note the virus does not say how the source be made
available.  When the GNU project started, it was "good enough" to "charge"
some electronic media that contained the sources, provided the "fee" for
the media was purely to recoop the writers cost to create it.

The internet of course existed since before the GNU project started, so ftp
sites became that de rigor way to met that requirement of the license;
although I remember "buying"media from the GNU project to help support
rms's work.

BTW: forcing B to make her/his sources available without any charge is
exactly the behavior RMS intended with the license.

I'm not sure it protects the user in any way other than tries to guarrentee
that a user of an invention/work, can get the sources for it if they want
it and B still exists to ask.   Which (if you believe in a completely
shared commons) is a fine thing.

Since the user rights are in front of the developers right, a problem with
this licenses is that the developer of a work, loses many/most of their
rights because they are obigated to give the sources away of the work also,
which allows anyone to make their own copy/replica of it.   And it also
means it is much more difficult to make a commerical product that depends
on it.   That is not to say you can not have successful commerical products
based on GPL'ed code - but experience has show that it becomes more
difficult to do so.  And in fact, for a pure software product, it can be a
limit on the developers success, although that is not always true.



The dead-fish stye license guarrentees that original developer of a
technology gets the credit for her/his creation; but it also means that
developer of  a derivitive work does have full protection of her/his work.
Because no one can "demand" access to the sources.   Which means that
developer of technology and its derivitives, are protect from the user (or
another developer) taking the work and competing with you using your own
technology.


You need to go back the Symbolics vs. Lisp Machines, Inc wars to truly
understand the whys and hows of the GPL virus.  At the time, rms was very
upset that people were leaving the MIT AI lab, and creating commerical
products (the Symbolics lisp machine in this case) and not making their
changes/additions/extensions available to community at large.   [When
Symbolics brought out a new feature, rms would often work night and day for
the LMI guys to develop the equivilent, but he hated that users like him,
could not just obtain the code from Symbolics.  So set about to recreate it
an gave his implementation to the LMI folks].

Hence went the GNU project started, rms wanted a license for his work that
would guarrentee that he could get to any change/extention etc that people
who followed him would want to make [and obviously the GPL worked in this
regard].  One other thing to rmember about the GPL, it helps defend against
is companies that go our of business.  If the work was made available,
there is a good chance some one has it and new projects be born from them
[good example is now dead Mandrake Linux vs. the current Mageia].

However the GPL "virus" does not stop a lot of developers, firms etc.. to
be warry of its "infection" because once you get it, like herpes, it stay
with you for life and unlike herpes, stay with you even after you are dead
and gone.

A non-viral license is much easier to import code into a project because
you do not chnage the behavior of the project you already have.  You have
just added a new feature too it.  The developer choses if they will make
their technlogy open source or not, they are not required too.  Hence the
more "liberal" nature for the project/developer.

A couple of "for instance" here might help.  Tru64 is a derivited work from
a lot of places, a big one being the OSF/1 kernel which was based on the
CMU Mach uKernel.   CMU had made rhe Mach ukernel "open source" and gave it
away - but the license did says to hd say you work was based on something
you got from CMU.   OSF and DEC used that technology and now DEC is long
gone, but getting the DEC sources is difficult [unless the owners of that
IP - in this case HP -  some how released it].    Ultrix is in the same
camp, or for that matter the DG rewrite of System V [which was one of the
best pure Unix implementations for a SMP system].   Similarly, OSx (darwin
etc) is also based on Mach, BSD etc..   At one time, Apple did chose to
make darwin available, but not all of the latest enhancements from Apple
have been made available.

Clem
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.trailing-edge.com/pipermail/simh/attachments/20150607/20037156/attachment.html>