Which distro for the enterprise now?

On Feb 3, 2004, at 6:22 PM, Craig White wrote:

> On Tue, 2004-02-03 at 14:46, Chris Gehlker wrote:
>> On Feb 3, 2004, at 1:08 PM, Craig White wrote:
>> But the GPL allows exactly that the code can become "the provincial
>> domain of any corporation , large or small, public, private or
>> non-profit" or any other organization.
> ---
> until it passes beyond their borders, this is true. But in a larger
> sense, I think this is motivation for large companies to use the
> software AND if they decide that this wasn't a permanent fork, they
> would certainly contribute to the code base.

---
It's really hard to say how willing companies are to contribute back.
If you read the Ruby list you will get the strong impression that there
is a feeling among coders that our clients are less willing than they
should be to open up the code. They seem to feel without evidence that
revealing the code would reveal too much about their business methods
or simply that having more developed code gives them a competitive
advantage. But of course coders are biased toward openness simply
because we want to show off.
---

>> You, among others, have softly
>> chided me for being naive enough to ever believe otherwise, though you
>> did it with some sympathy.
> ---
> I don't recall chiding you about GPL licensing - I think you are way
> more knowledgable than I am about the licensing programs. I struggle to
> understand the principles behind them as opposed to you, who is
> concerned with the code itself.
> ---

OK, it could have been Hans and Derek.

>> So to set the record straight, the GPL imposes *no* obligations on any
>> corporate entity to make derived programs available for use, study or
>> modification no matter how widely those programs my be deployed. So if
>> you are contemplating an 'open source' release, consider the RPL. It
>> is
>> much better at keeping derived works in conformance with the free
>> software definition.
> ---
> again not true - GPL does impose restrictions if the code is
> transferred
> to someone outside of the company and its internal usage.

I used the term "deployed" advisedly because 'deployment' is the term
that the FSF's FAQ uses to describe what the GPL permits and the RPL
forbids. The term 'distribution' is used to describe releasing the code
outside the organization.
>
> I recognize what you are getting at with RPL and that is provincial in
> another way, provincial for the coders themselves - that is not a bad
> thing and I wouldn't necessarily disagree with this. I only wonder if
> the RPL might cause corporations NOT to get involved with projects that
> involve this license because they are obligated to release any and all
> changes that they make for internal use (at least, that was my
> comprehension of the RPL license as you described it).

Yes the RPL does require corporations to release any changes that they
actually put into production. They are allowed to keep
developmental/testing code secret.

Yes when I say 'community' I pretty much mean the coders themselves. If
you press me I will admit that users are part of the 'community' but
only in the sense that the audience is part of the play.

Yes, I think there is a very valid concern that the RPL is so
restrictive that it will force corporations to bypass RPLed code even
at the cost of reinventing the wheel, or at least recoding it. I have
the same concerns about the GPL but to a lessor degree simply because
fewer companies are inconvenienced by it.

I'm about to put forward a line of argument that is totally untestable
but which I nonetheless tend to believe. Consider it an exercise in
pure speculation if you will. It goes like this:

The TCP/IP stack was developed as part of BSD and released under the
BSD license. Other networking systems were being developed by DEC,
Novell, Banyan and others. Because the TCP/IP stack was under a BSD
license, it became the fundamental standard of the internet. Everybody
could jump on board including Sun, governments and educational
institutions around the world and, belatedly, Microsoft.

Now imagine that TCP/IP had been developed under the GPL as part of
Linux. Sun, IBM, Microsoft and everybody else would have been scared
off. As the concept of networking was extended to something global
rather than something just for the office, a battle of proprietary
technologies would have ensued. Microsoft would have used the good tech
that they acquired by hiring the smart refugees from DEC and their
desktop monopoly to win the competition and what we have as the
Internet today would be Windows Networking version 4.2. Linux would
have been an interesting student project.

So yeah, there is always a trade off and choosing a license that
prevents your code from being embraced and extended may only mean that
it gets bypassed and ignored. Actually, most of the stuff I do these
days that is in any sense Open Source is released under the AFL which
is a very BSD/MIT style license. This is partly because of the very
concern that you raise.