Re: [Fedora-legal-list] "Must-read" licenses and their ethical implications. Possible workarounds.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
[FORWARDED: I noticed my reply to richard didnt get sent to the list, so im sending it here also]
Richard, thanks for your reply! It was helpful, but I think you missunderstood me - my primary concern was the ethical aspect, not only the legal one in this question. I provided some comments below.
This is a very interesting issue, about which much could be said. Actually, clauses similar to this one are not uncommon in FOSS licenses, though they differ in details and degree of annoyingness. They are invariably fossils from proprietary commercial
I went over a lot of the licenses on the Fedora licensing page, from what I can see (and contrary to what I believed at first), clauses like this actually seem pretty uncommon. I have only seen it, if I remember right, in two really small licenses for bundled software in OpenOffice and a portion of OpenSolaris (think it was gPhoto).
software licensing culture where they originated as (arguably quixotic) efforts to establish contractual assent. I myself would
say that, in theory, as a matter of caution, users should 'pre- screen', but we cannot deny that, in typical FOSS distribution contexts, no
one generally does this, and it is usually impractical to do so.
Exactly! And it is this issue I was concerned about.
Note that I was not primarily thinking about the LEGAL consequence here - my general understanding regarding FOSS licenses are that they are very mild - the chance that you will get into a harmful contract-relationship with a licensor via an OSI-certified license seems so miniscule that it could be disregarded alltogether, and therefore there is no real need or obligation to read the license if you are just an end-user (whereas, off course, programmers who want to create derivative works should be checking up on the terms).
My real concern, therefore, was the ethical aspect - you basically HAVE to read and understand the license in order to even USE the software...this is not the case with any other popular FOSS license I know (GPL, ASL, BSD...), and I think it is unacceptable, due to the group of people who will be shut out from using the software because they feel they cannot fulfill this requirement. People should have peace of mind dowloading packages, knowing that they have no obligations towards the licensor/author apart from what is customary in FOSS-licenses.
If this was only a semi-important license, then this problem in itself would not be so big. The problem is that this license has nestled its way into important packages like OpenJDK (via the old JDK) and OpenOffice.org. Man, I wonder if it is not even in the Firefox codebase.
I would say that the W3C type of clause is generally assumed within FOSS culture to have no interesting legal meaning and to be purely ceremonial. I believe this explains why, for example, the W3C license is considered by the FSF to be not only free but GPL-compatible, despite the existence of the clause. However, any such clause has
to be examined carefully; if it appears to be a real enforceable condition on the grant of rights, or if the licensor acts in such a way as to suggest that this is the licensor's view, we might well classify the license as non-free; indeed we may have done so in some past decisions.
I have addressed a mail to W3C (last friday I think, no answer as of yet) that might give an answer to if that is their intention. I might post the reply to this list when I get it.
this, but does not combining code under a GPL-compatible license with code under the GPL cause the combined work to fall under
the
GPL?
Often, yes.
If this is the case, does that mean that any terms of the GPL- compatible license (such as the W3C) can be ignored after such combination?
No, not usually, in the sense that the terms generally are assumed
to persist and burden users of the GPL-licensed larger work downstream. This is the whole assumption behind GPL compatibility theory.
I understood this after researching GPL-compatibility after I posted this mail...I think it is very unfortunate. It is maybe not doable, but it is a shame that any licenses being "devoured" by the GPL cannot simply be turned into copyright-acknowledgements in the main-GPL license of the combined work, and the whole work be "purely" only under the GPL (for example, John writes a parser under the BSD, Doey writes a compiler under the MIT, Sue incorporates them both into her GPLed IDE. Both the BSD and MIT licenses dissapear, but instead copyright acknowledgements for each component are incorporated into the GPL covering the whole work, so any end user never has to worry about the terms of anything but the GPL).
- Ban any
licenses that would fall into this category (possible, since for
now
the only license of this type I know off is the W3C).
No, for the reason stated above; we generally don't take such clauses seriously.
What if the W3C (in this particular case) were to say that they really mean it?
Richard Fontana Red Hat, Inc.
-
Christopher Svanefalk