Licenses and subterfuge

Alex Rousskov rousskov at
Sun Feb 29 19:09:57 UTC 2004

On Sat, 28 Feb 2004, Arnoud Engelfriet wrote:

> If there's only one library in existence that implements the API,
> then you _must_ have used that library.

Technically, no. I do not _have to_ use any library to make a
dynamically linked executable. For example, I develop open source
software that links with proprietary modules. Usually, I have no idea
what those modules are. I just add hooks for them. Some hooks are
probably used by a single module only. That does not make my software
a derivative of the module or vice versa (I hope). The only thing we
really share is the interface, which is not copyrightable.

But let's assume that software developers did, in fact, use a viral
library to test their software. That fact should not make the
resulting software (that contains no pieces of the viral library!) a
derivative, IMO.

> Then I cannot see how your program can be anything other than a
> derivative of that library.

Using something does not automatically makes the result a derivative.
For example, benchmark results are not the derivative of the
benchmarking software although they can only be produced by that
software. Most programs are not derived from the software used by Mac
chips, even though many of them can use nothing but the instruction
command set supported by Apple software. Use is not derivation. Only
certain kinds of uses make the software a derivative.

license-discuss archive is at

More information about the License-discuss mailing list