[Ocaml-lib-devel] Re: [Caml-list] User library license

[Brian H. <bri...@ql...>]

Changing lists...

On Mon, 24 Feb 2003, Blair Zajac wrote:
> 
> I see a difference between a kernel which definitely can be made
> proprietary from language or core library extensions, which I see are
> just tools to make proprietary code.

I agree.  The kernel and compiler are GPL, the library is LGPL.  Exactly 
the situation that is on Linux- Linux and GCC are GPL, glibc is LGPL.

You can use the GPL'd GCC compiler to produce propreitary code just fine.  
You just can't reuse, or link with, the GPL'd code- so you can't include 
parts of the compiler in your proprietary project.

> I wouldn't have an issue with this.  I see getting improving the core
> language to be more important than license issues.  After all, these
> libraries are not

I agree- we have to work with the language as it is.

> 
> I would use the C++ Boost library as an example of people contributing to
> a high quality library.  Parts of the Boost library may be accepted by the
> C++ standards body for future standardization.  I would guess that many
> people would be proud to get their code into such a library

And this governed the license requirements.  They greatly increased the 
possibility of being included as a standard with the restictions they put 
down.  The C++ libraries have too many reimplementations- Microsoft, Sun, 
HP, IBM, GNU, etc.  All of whom have votes on the standards committee, and 
are disinclined to have to do more work than necessary to implement the 
standards.  Forcing them to rewrite large amounts of code is unplatable, 
and would likely cause them to veto the proposed library extensions.  On 
the other hand, if you give them an implementation (preferrably one that 
already works on their compiler), so that all they have to do is package 
up the library with the next release of their compiler, they'll be more 
likely to be in favor of the proposal.

Ocaml is a different situation.  With the possible exception of F#, there 
is only one implementation.

Brian