[tlug] Yes! Another argument about the GPL! You knew you wanted it....

[Curt Sampson <cjs@example.com>]

On 2009-08-08 19:46 +0100 (Sat), Lars Kotthoff wrote:

> I think in academia the GPL is the best license to use for most things.

I find this a fascinating opinion. I have had some contact with academia
(three PhDs in my immediate family, and both parents tenured professors
in universities), and I've been enjoying free software since the days of
CP/M, long before GNU.

When I'm in "academic mode," or even just "free software mode," I rather
despise the GPL, though more for how it's sold than the particular
limitations it places on the freedom of those who use software licensed
under it.

But more recently, as a business guy, I've come to appreciate it quite
a lot. I think that it's a great way to open something up somewhat, get
great (if undeserved) publicity, and yet cut the feet out from under
any potential competition. At this point, I feel it's probably the best
license for a business to use if they want to make something look open,
but make sure it's not open enough that they'll actually have to work
hard to innovate faster than other businesses. For the copyright owner,
it's got a wonderful way of stifling competition. The thirteen year life
followed by the billion dollar sale of MySQL A.B. offers a pretty good
example of what the GPL can do for you.


On 2009-08-09 12:53 +0100 (Sun), Lars Kotthoff wrote:

> If you came up with [a] software product and released it under a
> do-what-you-want-with-it license, anybody could continue to develop
> it, make it proprietary, and sell it.

Well, two out of three. Certainly others can continue to develop it
and sell it, but they can't make it proprietary. Once you've made
*your* code freely available, unless everybody in the world refuses to
distribute it, it remains freely available.

It's certainly possible under, say, the BSD license, for someone to
write his own code and not distribute it freely, but regardless of
how you license your software, he's still capable of writing his own
independent code and keeping it proprietary.

All you can do with the GPL over other licenses is, under certain
specific circumstances, prevent others from selling their own code
without also giving it away for free. (Thus, the appeal of the GPL to
the businessman--stop the competition.)

This is really the key point about the GPL, when it compared with
licences such as the BSD license: it's all about what other people can
do with the code that they write.

> 1) any additional innovation by the profit-oriented entity would not
> be available to the research community the original innovation came
> from....

This is pretty typical confusion in the free software community; you
assume that, since the GPL places some great strictures on what you can
do with the code, other licenses are placing the opposite strictures on
it. But there's nothing stopping the commercial developer of code under
licenses that give them more freedom than the GPL from giving their
changes back to the community.

In fact, there's actually a reasonable amount of competitive pressure to
do so, since the cost of maintaining a branch is fairly high. Unless you
have some very significant additions to the software (enough essentially
to call them a product in their own right), it is in the general case
considerably cheaper to return the changes to the community than to keep
merging them with community-developed code over the course of years.

> 1) any additional innovation by the profit-oriented entity would not
> be available to the research community the original innovation came
> from, which entirely defeats the point of publishing it in the first
> place...

I requote in full here because I want to be clear. You really are saying
that, if you had an interesting research result, you would not publish
it unless you were reasonably certain that someone else would build on
those results and return that additional work back to you?

> Note that I'm not saying that you should always use such a license. If a
> profit-oriented entity wants to pay the researcher money to re-release it
> under a license they like, by all means!

Well, I presume by this statement you're not working for a
"profit-oriented entity." (Though it's hard to say that with a straight
face when you come out directly demanding profit for doing your work.)
But, as Stephen points out, does this mean that the public has to pay
twice for your work in order to be able to do whatever they want with it?

On 2009-08-10 01:52 +0900 (Mon), Stephen J. Turnbull wrote:

> BTW, the BSD and MIT/X Consortium licenses were written with the
> advice of some of the best lawyers in the business.  Do you really
> think that U.C. and MIT didn't know what they were doing?

Err, "clause 3." :-)

On 2009-08-09 20:03 +0100 (Sun), Lars Kotthoff wrote:

> Not at all. Again, I'm not saying that the GPL should be used for
> everything. In these cases it makes IMHO perfect sense to release the
> software under a more premissible license. The situation is different
> when you're working on software which is highly specialised and the
> algorithms used in it are the actual research.

Hm. I'd call the BSD TCP/IP stack a piece of software which is "highly
specialized and the algorithms used in it are the actual research."
Would you rather we experience more network congestion because we force
all those corporations building network devices to try and hack their
own stack rather than use one that has the best available algorithms
for avoiding network congestion and promoting fairness between nodes
generating traffic?

> > Where do you think that "public" money comes from?  Not taxes paid by
> > professors and graduate students!
> 
> They're paid by the general population (which includes professors and
> graduate students) and companies.

I don't think it's fair to say that someone who receives tax money from
the goverment and then gives some of it back is really funding himself.

> Assuming that a particular piece of research will only benefit one or
> a small number of companies....

But if it does, that's because other companies made the choice not to
use it. Keep in mind, when we're discussing the GPL here, we're not
discussing the original research; that remains freely available for all
under any open source license. We're talking about *additional* research
or other work done *beyond* the original research by a third party.

> ...it seems unfair that everybody else (including competitors of those
> companies) should essentially subsidise their R&D.

It would indeed be unfair, but that additional research is funded by the
entity in question, not anybody else.

> So it's not really taking and not giving anything back. There's a
> difference between standing on the shoulders of giants to see further
> (and maybe not seeing anything), and standing there to be able to deal
> out your merchandise more efficiently.

"Deal[ing] out your merchandise more efficiently" is often a public
good. Would you prefer that we all have to pay twice as much as we
currently do to buy a computer? I'm guessing that even you would agree
that being able to bring the price of a notebook computer down to $100
so that we can afford to buy them even for kids in Africa is a public
good, even though bringing the price down to that level was done pretty
much entirely by corporations seeking profit.

Note that I'm not saying corporations are always, or even most of the
time, serving the public good. But if a corporation extends open source
software and thus provides us with $50 home routers instead of $100 home
routers, that still leaves us better off whether or not they contribute
their changes back. Further, we're also better off if ten corporations
extend the software and only one returns its changes to the community
than if none of those corporations extend it at all.

The GPL, come to think of it, has a particularly American point of view,
concentrating on punishing bad behaviour over giving the opportunity for
good behaviour to happen.


cjs
-- 
Curt Sampson       <cjs@example.com>        +81 90 7737 2974
           Functional programming in all senses of the word:
                   http://www.starling-software.com