LGPL clarification

[David Johnson]

On Wednesday 01 November 2000 06:04 pm, Ken Arromdee wrote:

> > But the LGPL puts no restrictions on the distribution of P, which is
> > what the proprietary user cares about.
>
> That is not, however, what RMS believes.  If there is only one shared
> library that exists, he considers P to be derivative of it even before it
> is linked; and this triggers all licensing conditions on L even if P is not
> distributed with L.  Remember readline?

I can vaguely recall RMS arguing both ways on different occasions. The last 
that I heard, he said that the (L)GPL fully allows runtime linkage, but that 
dynamic linkage is not runtime linkage. Since I don't have his quotes readily 
available, I won't comment on his words further. I may be remembering them 
totally wrong.

However, there are a lot of people that confuse derivations of copyrighted 
works with derivations of code. They are two different things. Just because 
code A is dependant upon code B does not make A a derivative work of B. One 
thing that "classic" copyright law does not address is referencing. If a code 
base dynamically links to another, is it only referencing that other code 
base, or is it a derivative work? I would argue the former (and find ample 
debating opponents), but I don't believe that there is any case law that 
addresses the issue.

The license is a whole other ball of wax, however! Linking your proprietary code 
improperly to GPL or LGPL code might not be a violation of copyright, but it 
could be a breach of contract. For this reason I am much more amenable to 
licenses that solely grant permissions rather than those that impose 
restrictions along with the permissions.

-- 
David Johnson
___________________
http://www.usermode.org