Licenses and subterfuge

[Alex Rousskov]

On Thu, 26 Feb 2004, Arnoud Engelfriet wrote:

> Absolutely right. Still, it could be problematic if there is only a
> GPL-licensed library available to perform the functions you need. In
> such a case you may be forced to have a similar library developed to
> avoid having to link to such GPL-licensed software.

I do not see a problem from copyright/licensing point of view, as long
as the resulting binary mix does not need to be distributed (by the
company or its clients). The software in question simply implements
one "side" of a published interface, which has nothing to do with GPL.
Documentation may say:

  "Our software can be linked with any library supporting Foo
  API. Users report success with FooLib on Linux. Other Foo API
  libraries may be available in your environment. Known compatibility
  problems with Foo libraries are available by searching our bug
  database at ..."

The only potentially gray area is availability of a published API. If
the only API documentation are viral headers, the company can publish
the interface on their own. In fact, they would have to do something
like that anyway, when developing their own non-viral headers to use
during compilation of the binary they intend to distribute.
Documenting/publishing such interface as an RFC may be a good idea for
some interfaces.

Alex.
--
license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3