For Approval: TrueCrypt Collective License Version 2.0

[Ben Tilly]

On 7/12/06, Tomas Novak <counsel at truecrypt.org> wrote:
> Matthew Seth Flaschen wrote:
>
> > Obviously, you could not claim specifically that, but you could claim
> > the program was distributed under OSI-certified licenses, and thus in
> > whole is OSI-certified.
>
> Hi Matthew,
>
> In our opinion, we cannot safely assume that several OSI-certified
> licenses merged into a single compound license will automatically
> create a OSI-certified license. Theoretically, interaction between the
> licenses could cause the compound license to be non-compliant in some
> respect. Therefore, we believe we need OSI approval of the compound
> license as a whole, so that the interaction between the component
> licenses and the definition of the phrase "component license" are
> approved as well.

A license that contains several licenses is substantially more
complicated than one that doesn't.  That complexity will dissuade
people from being willing to look at your license, and therefore
reduces your chances of approval.  Furthermore there is a desire to
only go through the effort of approving licenses if they are reusable,
and your license structure is very much not reusable.

My suggestion is that you should write a single license which in your
opinion is compatible with the requirements of the component licenses.
 Make that license as generic as you can and submit it.  The
components may be available under other licenses, but that is
irrelevant.  The OSI will only have to look at one license.

If you're unwilling to go that route, then my second suggestion is to
submit the individual licenses first.  That way if there are problems,
they'll get sorted out more easily.  Then the combined license will be
much easier to approve.

Cheers,
Ben

PS Given the structure of the open source definition, if a series of
licenses individually is OSI approved, it is highly unlikely that
there would be a problem in approving the aggregate.