Re: Trademark scope (just for the record)
Quoting Steve Langasek (firstname.lastname@example.org):
> On Thu, Sep 06, 2007 at 11:50:21PM -0700, Rick Moen wrote:
>> Trademark law never requires any such thing.
> When you are distributing a product which is similar, but not identical, to
> the product of the trademark holder, using the trademark to identify it? I
> beg to differ.
You can beg to differ all you want: Trademark law imposes no such
requirement. As I said, the test is a clearly established one that you
needn't take my word for. You can check Bitlaw; you can check other
standard information resources; you can read the Lanham Act or other
applicable statute for your jurisdiction.
Nominative fair use expressly _permits_ use of names and other mark elements,
as long as the net effect that customers are not likely to believe that
the product with the allegedly confusing mark is produced / endorsed by
the organization that holds the registered mark. A judicious distance
in brand identity might make it prudent that Bob name his offering
"Bob's Earthbadger", but isn't going to bar him from making reference to
the earth and badgers. Bob would not, for example, have to substitute
references to ice and weasels.
>> Our understanding of trademark law is that we're obliged to ensure
>> that users of our derivative work understand that Alice Company
>> does not produce or endorse our version of the browser.
> So, here's where I believe the problem lies. I'm not convinced (and
> won't be, without competent legal advice to that effect) that calling the
> browser "Earthbadger", but posting a notice within "About Earthbadger",
> satisfies the legal requirement to ensure that the users understand it's not
> a product of Alice Company; and I don't believe that Alice would be sanguine
> about this potential for confusion.
Mind you, that is not the only thing Bob could do to ward off Alice's
legal claim. E.g., he could name his package Bob Earthbadger (or do you
think Coca-Cola Company is able to bar sales of Royal Crown Cola?). The
About screen could flash at startup as a splash screen, OpenOffice.org-style.
Bob even could have it be part of the bottom page border of every window
unless disabled by the user in Preferences. Additional or different
measures may be better. But the point is, to prevail, Alice would need
to convince a judge that her customers were likely to think she produced
or endorsed Bob's browser. When she tried that, Bob would point out to
the judge that, no, it says the exact opposite, right below the name and
copyright notice, and cite other measures taken to prevent that impression.
As http://chillingeffects.org/trademark/faq.cgi#QID52 says:
Nominative fair use
This is when a potential infringer (or defendant) uses the registered
trademark to identify the registrant's product or service in
conjunction with his or her own. To invoke this defense, the defendant
must prove the following elements:
* his/her product or service cannot be readily identified without
pointing to the registrant's mark
* he/she only uses as much of the mark as is necessary to identify
the goods or services
* he/she does nothing with the mark to suggest that the registrant
has given his approval to the defendant
> The nature of our disagreement seems to be in our differing
> assessments of what's going to cause confusion in the minds of the
I'd call that progress -- because _all of the prior writings_ from the
Debian Project on this and similar subjects have started and stopped
with "Well, gosh, we need their permission to use their name / logo /
images", which is of course nonsense.
So, yes, the question you _should_ have asked is: What is sufficient to
convince a reasonable man that a claim lacks credibility that a
trademark owner's customers are likely to be confused into thinking a
third-party product is not produced or endorsed by that trademark owner.
Personally, I would expect a prominent, clear disclaimer to the opposite
effect is exactly what meets that need.
_Linux Gazette_ did that, and it was exactly what was needed, even
though SSC's Phil Hughes was blowing $330 on a US Federal trademark
registration and waving his lawyers at us. You open an issue of the
_Gazette_ and it says to the lower right of the front page, and below
the copyright notice on every article, that SSC
doesn't produce or endorse our magazine. Now, does that mean all
readers of the magazine are unfailingly going to see that notice? No.
That's not the standard the law imposes.
(If you were extremely skittish about nominative fair use, you would
also name your browser Bob Earthbadger or maybe even BobBadger -- but
you wouldn't, say, do something as pathetic as retreating all the way to
"Iceweasel", right? ;-> )
Just for completeness's sake: In Hughes's case, additionally, there was
the difference that his trademark claim was very weak, since the
founding editor had not transferred any commercial rights to SSC, during
the period of the magazine's hosting, there.
> > (Your surmise that non-commercial use is "not subject to trademark law"
> > is overbroad, by the way. Non-commercial use can still, for one thing,
> > commit the separate tort of trademark disparagement aka tarnishment.)
> Well, it was you that brought up "commercial" in your earlier post...
I was _accurately_ describing what the actual legal requirement under
trademark infringement tort theory is, something that has been woefully
lacking over the years in the Debian Project as a whole, and apparently
completely unknown in this mailing list.
Meanwhile, had I _meant_ to make the extremely ludicrous claim that no
usage of Debian GNU/Linux is ever commercial, I would have said so. I
did not. Please do not attribute to me views that I do not hold and
> > "We didn't understand the law, so we caved in and gave the trademark
> > owner what he said he wanted, and now don't want to talk about it any
> > more."
> I think it's rather insulting for you to claim this is a lack of
> understanding on our part.
First of all, for someone who just got through attributing to me an
utterly ludicrous assertion of fact that I clearly never made, you take
personal offence rather too easily. I rather suspect cheap theatrics.
Second, I've been reading this mailing list for a very long time, and
have seen absolutely zero discussion, ever, of actual trademark
infringement obligations. Discussion has always started and stopped
with a trademark owner (or alleged trademark owner) dictating what
he/she requires for _permission_ to use a mark.
And that is what I wish to bring to an end.
> You haven't exactly cited any case law to support your position, and
> it's perfectly reasonable for Debian to not be interested in becoming
> case law here.
And how much caselaw do you want, Steve? There's a ton of analysis of
these matters over at http://chillingeffects.org/
But certain, yes, it is very much always in the Debian Project's
interest to avoid having to be a test case. I certainly would not
propose anything likely to bring that about.
> > I can only suggest that you read what I said again, because it is
> > _exactly_ what one does when one's third-party offering is within the
> > same trade or industry.
> That's why Microsoft trademark suit against Lindows.com was dismissed and
> Lindows.com kept that name instead of changing it to Linspire, is it?
Lindows, Inc. had inadvertantly compromised its legal position by
carelessly not using appropriate notices to users. Also, it found it
much more profitable to be bought off by access to some proprietary
Microsoft programming interfaces, and co-marketing concessions.
> > I referred to "falling for" the entirely false but nearly ubiquitous
> > assertion that one must secure a trademark owner's permission before one
> > may use the mark for a third-party offerings. It's just not so.
> > Pepsico doesn't ask the Coca-Cola Company's permission to publish claims
> > that its sugar-water is better tasting than is Coca-Cola. That ought to
> > be a big, fat clue, but far too many people have been successfully
> > conned and don't think about the implications.
> You think that calling a non-Firefox product "Firefox" is equivalent to
> *referencing* a competitor's product by name in one's advertising?
I was not suggesting that was the case. Read what I _said_, please. I
was pointing out one huge clue, from the realm of everyday commerce,
that should have alerted Debian users to the fact that, no, it is not
true that one must ask a trademark owner's permission to use that mark.
My point is that I've never seen any sign of awareness from Debian
Project people or on this illustrious mailing list of that basic truth.