On Saturday, November 10, 2001, at 11:14 am, Daniel Joyce wrote:
>> If you provide software with a restriction, the restriction
>> restricts the actions of the users of the software (i.e., it
>> removes a freedom from the consumer).  GPL seems to be trying to
>> restrict what restrictions the user may *add* (for a subsequent
>> end-user), which can probably be shown to be an internally
>> inconsistent restriction (rough sketch: try applying it twice,
>> 180 degrees out of phase, to produce a contradiction).  A
>> contradiction in this sense simply means that somewhere down the
>> road a user is required to (1) do X, and (2) not do X.
>
> 	Uhm, howso. The GPL never gets 'looser' or 'stricter'. It's the same.
> Everyone plays by the same rules, so how can it get 
> inconsistent? The scope
> is never narrowed or broadened.
>
> 	And how would you apply it "out of phase?"
[argument about licences of composition of programs snipped]

Before going on, I'd like to point out that I haven't actually 
read the GPL license.  It sounds a lot like an idea I had (and 
abandoned) in university in the late '80s.

Never mind.  Let's try a completely different approach.  Suppose 
I copy the GPL license in its entirety and change every place 
"GPL" occurs with "MvGPL" (see "from" field).  I conjecture that 
the resulting license is actually incompatible with GPL!  
Therefore, GPL is not only virally restrictive, it even has its 
own identity.  That particular property is what led me to 
believe there would be an inversion of restriction that could 
produce a contradiction.  The phase refers to how many steps 
along the "user A distributes X to user B who uses it in Y, 
which is distributed to...".

As for my actual *opinions* about the licenses (which I have not 
yet expressed):  If I work on commercial product, I will license 
it fairly restrictively in an attempt to profit from my work.  
If I work on code which I expect to provide particularly 
intangible value for someone, or value which I do not expect to 
be able to reasonably profit from, I release it with a license 
that simply states anyone can do whatever they want with it 
except claim they wrote it.  I might even waive that clause if 
someone asks me nicely enough (and if they can show good reason, 
like wanting to release their own code without having to list 
everyone who ever patched the assembler with which was built the 
C compiler which was used to construct the database in which was 
stored the to-do list used during product development...).

My impression of GPL (from discussions on this list and other 
(secondary) sources) is that it is unsuitable for commercial use 
and unsuitable for uncommercial use.

The sandbox analogy is terrible.  If someone enters the GPL 
sandbox owning a toy, they have it taken from them when they 
leave!  (I.e., they aren't allowed exclusive ownership of 
anything that has touched the quarantined toys in the GPL 
sandbox).  That's the exclusive-ownership side of things.  From 
an information-authorization point of view, it's more like 
saying that if kids snap together commercial Lego and GPLego, 
they're not allowed to show anyone what they made.


> 	GPL when you get down to it, is just a 'play nice clause'.

Ha ha ha ha ha ha ha ha ha ha!  You're joking, right?


> 	It keeps someone from squirreling away code.

Which is a problem how?  By forcing people to publish any 
experimental garbage they happen to ever write (and we all do), 
how does GPL benefit *anyone*?


> 	Uhm, the GPL is pretty simple, and I don't understand the 
> hatred for it.

I don't hate any inanimate objects or concepts, even the 
disgusting concept of religion.  I hate that anyone would ever 
happen (or choose) to be harmed by the application of such 
concepts.


> 	"I made this software, and I'm sharing it with you."
> 	"The only thing I demand is that you share it with others"
>
> 	That's the gist of it. I can charge for it, but I have to 
> share the source
> code.

That sounds like the opposite of GPL.  GPL demands that anything 
you ever use it for must also be shared with others.  It's not 
ensuring free beer for all, it's more like outlawing money, 
glasses, and breweries!


> 	Has "Sharing" become such a dirty word since the 80s?

It has its place, but it's not the only concept in the 
universe.  Besides, if I add GPL code to MvGPL code I can't 
share it with *a single person* (if my previous conjecture is 
correct).


>> Hm.  I never really congealed an analogy with co/contravariant
>> types, but I think (and hope) the gist of the argument survived.
> 	
> 	Well, accounting it has been applied to THOUSANDS of 
> software packages, and
> no such contradictions have become apparent...
>
> 	Also the GPL is quite simple.
>
> 	"If you choose to use the GPL:
> 	If you sell the program, you must give the customer the source code.
> 	You may not prevent your cstomers from modifying the source code or
> reselling it.
> 	The source code given to the customers must be under the GPL as well".
>
> 	If you can get all the copyright holders of the software to 
> agree, you can
> release different versions under different license. Copyright, 
> and licensing
> are strictly different. There are propietary licensed vers of Alladin
> Ghostscript, and the year old versions are relicensed as GPL. 
> ( Which is
> great, if you want to see how GS rendering works I imagine....)
>
> 	Pretty simple rules I think, and I don't understand why 
> everyone wants to
> shoot it down?
>
> 	I mean, it's JUST a licensing idea. No different than BSD, etc.

That's how I treat it.  It's just an idea, nothing more, nothing 
less.  Whether it's appropriate for a particular purpose or not 
is where it gets interesting.


> 	But when I release software, and I want people to play 
> nice, I'll use the
> GPL. BSD, they could make a propietary version, and keep all the fun to
> themselves. And where is the fun in that.... ;)

Maybe actually being employed in the IT field?  If you feel the 
IT field should be shut down, you must do so very carefully (and 
very slowly).  Maybe GPL is the right way to do this after all.  
Then again, it simply fractures software into multiple "camps" 
which are forever incompatible.  I believe someone mentioned 
proposed US legislation like DMCA that would ban today's 
hardware just like DMCA banned some kinds of thought.