Re: OpenTTD infringes TTD
[snipped lots of stuff where we agree]
On 5/31/05, Nathanael Nerode <firstname.lastname@example.org> wrote:
> I care less about OpenTTD in particular (I wouldn't touch it because of the
> dependence anyway, even though I have TTD) than about the application of
> the principles to other games. That's why I gave the Transport Empire
> example: do you think *it* has any chance of being copyright infringement?
Probably not, as long as the people working on it don't have a history
of using others' copyright material without license to bootstrap their
development efforts. Sometimes you get away with that -- as in Sony
v. Connectix and Sega v. Accolade -- and sometimes you don't -- as in
Atari v. Nintendo. The degree of prejudice with which a court views
possible non-literal infringement is in practice often determined by
the parties' larger history of dealing in good or bad faith.
> >There are large swathes of decompiled code from the original.
> Yuck. That's obviously outright copyright infringement, no doubt about it,
> no defense, and if this is the case, Debian must not distribute it.
http://encycl.opentopia.com/term/OpenTTD -- and note that Atari
Interactive (now a brand of Hasbro Interactive) certainly has resumed
legal pursuit of clone distributors.
> >a patch to the original that's been merged into OpenTTD.
> Really? I thought they'd just implemented all the "functionality" of the
> patch. Did they actually take the code? If so, not good.
Even implementing its "functionality" puts one in the "unauthorized
sequel" boat -- although the authors of TTDPatch would of course be
unable to obtain a copyright on it, since it was created without
license from the TTD copyright holder. Compare Palladium Music v.
> > For a case right on the borderline, look at Pearl Systems v.
> > Competition Electronics as cited at
> > http://www.law.gwu.edu/facweb/claw/ch5a.htm .
> Apparently they failed to assert a user interface defense. Unlucky for
> them. The professor quoting the case clearly doesn't agree with the court
> (and probably for that reason).
There may have been additional evidence, not apparent from that
excerpt, that Competition Electronics had plagiarized code from Pearl
Systems. It would not be the first time (or the last) that a
programmer labored under the illusion that restating a program "in his
own words" (perhaps in a different implementation language) made it
> How about FreeCiv, though?... that's a much more interesting case, since it
> has no history of code copying and the internal implementation is certainly
> completely different (given the client-server arrangement). Some of the
> elements are clearly subject to the doctrine of merger (cities, counting
> how much food is produced, turns, squares, buildings, wonders, etc.),
> others to scenes a faire (terrain, units, attack and defense ratings), and
> yet others to being uncopyrightable user interface (interface for directing
> units, city production list) -- but some of it is harder to defend. For
> instance, the use of "shields" to represent production (seemingly arbitrary
> -- why not ingots?); the exact structure of the tech tree; and the name and
> function corresponding of the technologies, buildings, and Wonders. (Most
> of this, interestingly, is implemented in *text files* which could be
> completely replaced.)
The tech tree and Wonders are bad news, as is the history of using the
"civ2gfx" tool to convert the original Civ II graphics for use with
Freeciv. And the "doctrine of merger" exempts ideas at a broad-brush
level and facts about the real world in much greater detail, but
cannot be used to justify copying all of the little numerical
properties (unit costs, movement and attack/defense ratings, and all
that) which make for balanced gameplay.
Freeciv may not contain literally copied code from Civ II/III but it
remains an unauthorized, copyright-infringing sequel. Expunge the
remaining literal similarities (tech, Wonders, little shields on the
unit icons) and it's a little better, but it's going to be an uphill
battle to convince a judge that it has substantial non-infringing uses
if everybody still uses the Civ II/III rule set.
> How about the putative Transport Empire? That's certainly not a ripoff in
> any way.
Not unless and until it gets fleshed out with copyrightable expression
plagiarized from somewhere. :-)
> >From the Microstar vs. Formgen case, footnote 5:
> > We note that the N/I MAP files can only be used with D/N-3D. If another
> >could use the MAP files to tell the story of a mousy fellow who travels
> >through a beige maze, killing vicious saltshakers with paperclips, then the
> >MAP files would not incorporate the protected expression of D/N-3D because
> >they would not be telling a D/N-3D story.
> Interestingly, this is somewhat analagous to the often-stated-here idea that
> "if it works with multiple libraries, it's not a derivative of any one of
> them, but if it only works with one, it may be a derivative of it".
Not really. Judge Kozinski was only addressing the "characters and
mise en scene" theory of non-literal infringement, which is not the
issue with library APIs. As I have argued ad nauseam, it doesn't take
multiple implementations to make a library API uncopyrightable (or, in
some circuits, so closely tied to a method of operation as to be prima
facie defensible as fair use). In any case, the
mousy-beige-saltshaker-paperclip alternative would have to pass
scrutiny as a substantial non-infringing use, not just a fig leaf for
use with Duke Nukem.
> >How much storyline do you think Duke Nukem has?
> Way more than TTD. Oh, and it's less generic. The judge specifically notes
> that it's clearly a fantasy, and apparently the lack of realism contributes
> to its strength as a protectable story.
That's true; but Freeciv's "realism" (howitzers and paratroopers
instead of jetpacks and "Pig Cops") doesn't mean that the gameplay
details aren't protectable expression. Ditto TTD's tile costs,
terrain types, or whatever (I haven't played it). Judges are not
stupid and may well listen to expert witnesses who explain that
gameplay balance isn't just "sweat of the brow", it's as much a part
of the game's expressive content as the character/tile names and
> User investment in keystroke patterns is just as important. It's possible
> that courts will rule the wrong way on future cases similar to the Lotus
> case, but if they do so consistently, we will have problems on the order of
> the software patent problems. Apple didn't do too well with its
> extravagant "look and feel" claims either. User interfaces -- not counting
> the nonfunctional decorations and artwork -- really need to be freely
> implementable, due to the "used to it" effect, and in fact they are
> routinely treated this way by *all* computer software companies,
> proprietary or not, because otherwise it would be impossible to make any
> usable user interfaces without a license.
Apple v. Microsoft 35 F.3d 1435 (9th Cir. 1994) is an interesting
case. The district court ruled that there were elements of
protectable expression in Apple's interface paradigm, but that they
had been licensed to Microsoft under terms which (inadvertently, from
Apple's point of view) permitted their transplantation to later
Windows versions. The appeals court upheld the district court's
conclusion that Apple was therefore only entitled to the sort of
"thin" copyright protection that subsists in compilations of facts:
"Having correctly found that almost all the similarities spring either
from the license or from basic ideas and their obvious expression, it
correctly concluded that illicit copying could occur only if the works
as a whole are virtually identical."
The implication for Freeciv and similar games is that copyright in
gameplay details may be limited to this sort of "thin" protection.
There's a long continuum between "virtually identical" unit properties
and more or less inadvertent overlap of Fibonacci-series unit prices.
Where on that continuum the border lies between defensible and
indefensible similarity may depend critically on the sincerity of
efforts to avoid plagiarism. Again, compare Sony v. Connectix and
Atari v. Nintendo.
> > Taking a stand on
> > MPEG patents and crypto laws is one thing; providing a distribution
> > channel for truckloads of e-toys of dubious provenance is another.
> Indeed, Debian should not provide anything of dubious provenance. If you're
> right about the decompiled code, then it is clearly infringing.
> At this point, I'm convinced that it's probably unsafe to distribute
> something which depends on proprietary artwork, either.
In general, it's unsafe to distribute things that were created by
plagiarism even if people have subsequently made more sincere efforts
to clean up their act. Other forms of unfair exploitation of a
pre-existing work can also make it riskier to skate the bounds of
copyright infringement. It's a good thing for the GNU TLS folks that
the public archives of the gnutls-dev mailing list demonstrate that
the OpenSSL shim is a relatively late development; otherwise the FSF's
efforts to suppress use of OpenSSL from GPL applications would lend
color to charges of plagiarism as well as unfair competition.
Let me say that again carefully: I think the record demonstrates that
GNU TLS was not originally plagiarized from OpenSSL. But I think
that, given the FSF's bogus claims about GPL / OpenSSL
"incompatibility", putting an OpenSSL compatibility shim into the GNU
TLS code base was anti-competitive and obnoxious. Doubly so when it's
inserted into the GPL "libgnutls-extra" instead of the LGPL
"libgnutls", providing a low-risk migration path to GPL projects and
no one else. (Thread at
I haven't checked whether the shim is still GPL.) The FSF had better
hope that it can successfully argue "APIs are uncopyrightable" if that
one ever winds up in court.
(Hmm, looks like I veered off into my idee fixe again.)
> My main argument was against giving excessive credence to broad claims of
> "non-literal copying", which are not only hard to prove, but consistently
> hit the limits of copyright, and routinely involve free speech issues.
> (Broad claims, as opposed to "non-literal copying" cases where the copy is
> essentially a translation into another format.)
My approach tends to be to look at the available facts and try to
think like a judge. What protectable expression might the two works
have in common? What evidence is there that the text of one work was
available to, and examined by, the creator(s) of the other? What
evidence is there of conscious measures to avoid plagiarism, and how
sincerely and competently do those measures appear to have been
carried out? If the defendant had to fall back on "fair use", how
would the four "fair use" factors stack up?
Once these questions are answered dispassionately, I think it's
usually pretty clear who's in the right -- if you've read enough case
law to have a sense of where the boundaries are. IANAL and a fortiori
am no judge, but for what it's worth I have found that rulings that
originally bothered me (such as Apple v. Microsoft and Sun v.
Microsoft -- and even Heritiers Huston) have made a lot more sense
once I have read the whole series of cases.