Re: GNAT availability

[mfeldman@seas.gwu.edu (Michael Feldman)]

In article <D5zp52.5A5@thomsoft.com>, Keith Thompson <kst@thomsoft.com> wrote:

[snip]

>As some else will probably have posted already, GNAT is *not* public
>domain; it's covered by the GNU "copyleft".

>For many purposes they're about the same (you can freely copy it, read
>the sources, etc.), but it's an important distinction.

The real distinction, for better or worse, is that a Public Domain
product can be picked up by anyone and improved, and the result can
then be commercialized with its source code kept proprietary. A
"copylefted" product can be improved and commercialized, but the
sources of the result must be released when the binaries are.

This is a two-eded sword, of course, but clever developers can and
(I hope) will capitalize on GNU software - including GNAT - anyway.
You can build tools on top as proprietary stuff, as long as you
don;t commingle your own sources with anything copylefted. 

For example, I am happy that Bill Yow's OS/2 PM binding is also copylefted, 
but if Bill wanted to, he could have made it a proprietary product.

We could also have developed our DOS GNAT editor as proprietary, but
chose to copyleft it instead. Actually, we _had_ to copyleft it,
because it is derived from our Ada/Ed editor, which is copylefted.:-)

One more thing: you can write code with GNAT or any GCC tool and keep it
proprietary, as long as you are careful to follow the GNU Library
License rules for linking in GNU libraries. The best known case of this
is NextStep, which was written with GCC but is _certainly_ proprietary.

Mike Feldman