Promotion of derived products, the Microsoft case (was: Scope of copyright on derivative works)

[Philippe Verdy]

Alexander Terekhov [mailto:alexander.terekhov at gmail.com] wrote:
> - Neither the name of the <ORGANIZATION> nor the names of its
> contributors may be used to endorse or promote products derived from
> this software without specific prior written permission.

This kind of statement is exactly the typical restriction that I was
speaking about the Microsoft's possible attempt to limit the use of its
licence name, because it would indirectly reference Microsoft itself without
permission for promoting a product distributed under the terms of one of the
proposed Microsoft licences.

So if this statement is enforceable, this means that such restrictions are
valid and enforceable, and so, unless such permission is explicitly provided
in the Microsoft proposed licences, the names of the licence will cause a
problem, given that the trademark "Microsoft" is effectively protected, and
obeys to its general usage policy explicited in the Microsoft web site or in
the EULA, or simply by law and in by official trademark registrations in the
various countries where it has been protected.

I can't safely assume a "fair use" legal clause, simply because the "fair
use" exception does not apply to the case of promoting competing products,
and this is not even a citation (if software are considered as artistic
creations, in countries where this is applicable). Don't forget that "fair
use" is mostly a US-only exception, this legal right may not even exist in
any other country or applicable legislation.

The applicable right is the one applicable to references to identifiable
parties, and this right is typically not included within those granted by a
licence (some licences are even forbidding their licensees to reveal the
identity of the licensor, but generally this applies to exclusive licences
only, where the rights ownership are completely transferred to another
party, i.e. sublicenced).

We can see for example many websites that FORBID even any kind of reference
to their public website without prior written permission. To enforce it,
they include "robots" exclusion rules within all their pages, and are
creating the external references themselves (they choose the partner
websites or search engine where these external references to their site will
be accepted).

If a Microsoft licence is effectively open, it should not limit who are
authorized to use it. So any kind of derived products could be done,
including:
* open-sourced products competing with commercial proprietary Microsoft
products.
* products made by people actively campaigning against Microsoft positions.
* products whose application is in domains that are NOT within the
commercial interests of Microsoft (for example the pornographic domain), and
with which Microsoft would probably not like to be associated, even
indirectly.
* products containing some political or religious interests (possibly
breaking the Microsoft's own policy of neutrality in these domains). 
* and so on...

It's acceptable to limit the form under which the licence name can be cited
(the licence can specify a required sentence to use so that no confusion
will be made between the name of the company and the name of the licence),
but not its context of use (so referral exclusions, even those that are
implicit, should not be accepted in a true open licence, because one cannot
respect both the legally enforceable exclusion and the obligation of citing
correctly the terms under which the content is licensed).

Am I now clearer about the reasons why I asked this to Microsoft, so that
they give explicit statements solving the possible conflict between the need
to reference the licence and the need of not infringing the highly protected
Microsoft trademark?

Some have asked me to find a past legal action or trial initiated by
Microsoft for such trademark infringement. It is not relevant. What is more
relevant is:

** Will Microsoft be able to initiate such legal action in trademark
infringement, given the proposed licence name and the associated licence
text, because of some uses of the licence in some contexts that should have
been covered by an "open" licence? **

I consider this being a good test for other open licences (including those
that have been already approved by OSI) and whose names include a registered
trademark associated to severe protections (so this concerns also any other
open licence used by IBM, Sun, Apple, AT&T and Bell companies or other
telcos, Google, Yahoo, ...).

The licences published by the FSF cause much less problems: the FSF has
restricted its own name and registered the "GNU" project name only to avoid
agressive registrations by someone else, but then published many statements
about how to cite the licences. Its licence names do not even include "GNU"
or "FSF", so there's no possible trademark infringement when speaking about
the "GPL" or "General Public Licence", even when advertizing products
delivered under this licence.

And even if we refer to the Free Software Foundation, this is still a
*non-commercial* organization, so it cannot legitimately make any profits
from its own registered name (unlike Microsoft, Sun, Apple, the MIT...), and
is not a physical person (so the moral rights exceptions don't apply to it
in countries where moral rights of physical persons are recognized as valid
and legally protected IP).

Unfortunately this is not the case with the *exact* reference names of "BSD"
and "MIT" licences, that are citing protected trademarks or organization
names in the name of their licences. And in fact, this seems to be a
possible problem with most of the existing OSI-approved licences! A better
way to solve the issue would be, for authors of licences, to give them a
distinctive title and register it (like the titles of books or films), and
then explicitly grant to all licensees, in the license text itself, an
universal, unlimited, and non-exclusive right of referencing this name, with
a statement also making this name sufficient for referring to the licence
text itself that explicit the granted rights and mandatory obligations.

This has another advantage: no need to constantly change an existing open
licence such as in the proliferation of BSD-like or MIT-like licenses, and
then simplifying the resolution of conflicts between multiple licences that
should have been theoretically "compatible"; to make this possible, we
clearly separate the license text itself (which has a distinctive name and
can be freely referenced by anyone), and the copyright assignment. Both are
linked together within the published or advertised products by a simple
copyright notice like:

"<Product Name> - Copyright (c) 2007, <Author Name>, <Country>. This product
is licensed under the terms of the <Licence name>".

That's exactly this kind of very simple statement (which offers no
possibility of ambiguous or conflicting interpretation) that has been used
since ever within GPL-licensed products and in their sources.

And this is another reason why I still think that the FSF-published "free"
licenses are still FAR better than most of the OSI-approved "open" licenses
(not so open because some needed rights are *NOT EXPLICITLY* granted by most
of them).

----

In fact, if I was a bit extremist, I would request to all "open" licences to
be:

* first published like a book, with a registered ISBN (which could be an
alternate way to refer to the licence text): the ISBN registration would
clearly identify who wrote it, what is the law applicable to it and
protecting the text itself.
* and then softwares could be licenced by making an unambiguous reference to
this book or ISBN number.

Note that old ISBN numbers are now part of a more general EAN registration
number scheme, applicable to any kind of product including licences; for
information about the migration that occurred since last year, look at the
ISBN web site; a software can also have its own EAN number if it is
distributed independantly, or a ISBN number if it is published as part of a
book, or ISSN number if it is part of a serial publication or part of a
website.

Many kinds of EAN number registrations are (still) free of charge in many
countries (this depends on the local registration authority you choose and
the kind of products or services they accept to register, depending on the
initial form of distribution you choose for the software seen as a product).
The ISBN or EAN registration is open to individual persons (this does not
necessarily require an "publisher", but if a "publisher" must first be
assigned a group of numbers, before assigning a number to a book from a
specific "author", a non-profit association could play this role).