> GPL'ed means that it's as free as possible.
No. Please read the GPL, and compare it it makes you do with what
the BSD license makes you do. In fact, the GPL is so restrictive
that Red Hat and other companies appear to me to be violating it
on a regular basis, because it's such a pain in the ass to comply
with. In fact, I'll bet the LUG is going to violate it at its
installfest.
> Earlier you mentioned (was that you or someone else? that some enhancements
> were done to the BSD protocol stack software in the light of the outdated
> mbuffs code which micromanages memory and isn't well suited to modern
> architectures and machine configurations.
No, I never mentioned any such thing. Nor did anyone else, that I recall.
I would appreciate it if you would not put words in my mouth that
imply that the mbuf code is `outdated'; it is, in fact, not, as
far as I can tell at this time. Certainly the concept of buffer
chaining is something that Linux is moving towards, after many
years of suffering without it.
> Unfortunately, these changes were
> done by vendors and are not available to the free software community.
I don't know if these changes you speak of exist. If they do, that's
fine. I believe that everyone should be free to determine what he
wants to do with the code he writes, and everyone should be allowed
to contribute to the free software community to the extent he feels
he can manage. I don't like this `contribute everything or get out'
policy that the GPL promotes.
(Do note that there have been plenty of changes done by commerical
companies on the BSD code that have been contributed back to the
free community. If keeping some of their code proprietary helped
to keep these companies in business so that they could continue to
give us some of their code for free, that seems to me to be a win.)
> > have to do to make it free (and usable by BSD folk) would be to
> > remove the GPL, as the GPL insists that you change the license of
> > other code not owned or written by the coder in questino.
>
> Right; that's because it's not always easy to separate new code from old code.
> What is a modification and what is an add-on?
Why does it matter? The original code is still available to anyone
who wants it. Nothing has been lost to the community if I make some
changes, sell the binaries, and don't give away the source.
> For cases in which it is easy, the LGPL was invented.
I'm aware of the LGPL. Unfortunately, a lot of code isn't under
it, and even so it's still more restrictive than a BSD licence.
> The GPL is in line with international copyright law. If you take some existing
> copyrighted work and change it, you are creating a work that is derived from
> the original work. The original copyright applies even if you made genuine
> additions to the work.
But copyright law doesn't require you to turn over all of your work
to the original writer if you make changes.
Regardless, the analogy breaks down pretty quickly anyway, because
we're talking about licencing and source versus binary distribution,
neither of which exist in written materials.
> If you don't like that something is under the GPL, or you would like to extract
> some portion of a GPL'ed program to use as a component in a proprietary
> program, you can always contact the author to arrange alternate licensing
> terms.
You can. But that alone can be a royal pain; imagine doing this
for the hundreds of different authors that often appear in large
systems. I've seen the nightmare just doing attributions for the
advertising clause of the BSD licence can produce; this would be
ten times worse: so bad that people generally won't get involved
with it.
> If it weren't for the GPL, we would see all kinds of proprietary versions of
> Linux by now.
You've no evidence of that. There's only one proprietary version
of BSD right now, and it's existed since before the NetBSD, FreeBSD
and OpenBSD projects started. And those folks regularly contribute
code to the free community.
Regardless, how would a proprietary version of Linux impede your
ability to make changes to Linux?
> These would all be incompatible, just like the various BSD
> flavors, and all the operating systems derived from BSD, not to mention
> non-UNIX-like operating systems that contain stolen BSD code.
You're engaging in sheer hyperbole here; it looks like you may be
attempting to start a flame war.
The BSD code they use is not `stolen.' It was freely given to them.
Just because you don't appear to believe that code should be freely
given to others for any use doesn't mean that all authors feel the
same.
As for compatability, no, we're not all perfectly compatable (though
we're pretty close). Nor are all Linux installations perfectly
compatable.
However, it certainly is easier for us to be compatable in some
ways. For example, we can put together and distribute a package
that contains the FreeBSD libraries and other code needed for
emulation quite easily, and a simple `pkg_add' command will let a
user add this.
If we want to do the same for Linux, we have to do one of two
things. We must either add the entire source code for the Linux
binaries we're distributing to that package (thus bloating the
package size greatly, increasing download times, and wasting space
on the machines of the majority of the users who have no need or
desire for this) or we must be able to keep that particular version
of the source code available on our FTP server for three years.
Given how often Linux libc changes, we'd be spending a lot of space
(and thus money) on making the Linux source available.
These just aren't worth the hassle, so we don't make binaries of
the Linux libs available. This hurts nobody but the users of free
software.
> Would Windows NT
> have anything resembling reasonable networking if it weren't for the BSD
> implementation? Fat chance.
I expect it would. Regardless, if they used Berkeley code in their
networking implementation, they are welcome to it. If they used
any of my code that's in the NetBSD stack in their impelementation,
they're welcome to it. That's because the code I write is free for
anyone to use for any purpose.
> I believe that the GPL protects the interests of free software developers
> better than any other kind of licensing scheme....
I happen to be a free software developer, and have contributed many
hundreds of hours and thousands of lines of code to the free software
community. The GPL does not protect my interests as well as the
BSD licensing scheme. I personally know several dozen developers
who are in agreement with me. Therefore, your generalisation is
incorrect.
> Making open-source freely available is only half of the battle; the other is
> preventing money-hungry software corporations from stealing the work.
Perhaps for you. I'm not so selfish. That may be because I've done
commerical software developement, too, and I have a better idea of
what costs are like, and what benefit the end user sees from free
software incorporated into commerical work.
> It's a joke when free software is exploited by enemies of free software.
You are quite incorrect in your implied assumption that anyone who
makes money from developing software is an enemy of free software.
Think about this for a bit: if `free' software is not supposed to
provide benefits to commerical software users, why do we have the
LGPL? That appears to have been created to allow the commerical
exploitation of free code in a way that the GPL doesn't. In fact,
why do we not have a licence on gcc that says any binaries compiled
with it must be distributed with source, instead of allowing
commerial software developers to get a free ride on a free compiler?
cjs
Curt Sampson
Info at http://www.portal.ca/
Internet Portal Services, Inc. Through infinite mist, software reverberates
Vancouver, BC (604) 257-9400 In code possess'd of invisible folly.