Sigma Designs Accused of Copyright Infringement 417
Cygnus v1 writes "The XVID team has ceased development of the XVID video codec for the time being because they say that Sigma Designs' REALmagic MPEG-4 Video Codec software includes their code and has claimed it as Sigma Designs' own work. The current XVID homepage includes some binary-level comparisons." Update: 08/23 03:14 GMT by T : Apparently the folks at Sigma have seen that no good is likely to come from this; an anonymous reader submits a link to this release on Yahoo! which says "complete source code will be available for download starting August 23, free of charge, through Sigma's website."
press release text (Score:5, Informative)
ERLANGEN, GERMANY -- August 22nd, 2002 -- The XVID development team, author of the popular XVID MPEG-4 video codec, claims that Sigma Designs' REALmagic MPEG-4 Video Codec is an illegal copy of the XVID software and publicly requests the company to stop violating their software license and copyrights.
XVID is a leading open source MPEG-4 video research project: software distributed by XVID is covered by a Free Software license, the GNU General Public License (GNU GPL). The XVID team announced that Sigma Designs' REALmagic MPEG-4 Video Codec includes wide portions of XVIDcodec software. By not offering a corresponding source code distribution and by claiming sole authorship on the product, Sigma Designs' Inc. is violating the GNU General Public License and the copyrights of the XVID authors.
XVID learned about the license violation in early July, soon after the initial release of the REALmagic software (version 1.0). Sigma Designs' were immediately contacted, and replied confirming the violation and promising to replace all violating code.
Version 1.1 of the REALmagic software was released on the 9th of August. After examining the new version, XVID developers concluded that the violating code was not replaced, but disguised by programming and compiling tricks. Sigma Designs' were again contacted and asked to remove the REALmagic download link from their website. Thus far, they have not shown any sign of cooperation.
In a statement to the XVID development team, project founder Michael Militzer showed his disappointment regarding Sigma Designs' behaviour: "We have been quite reasonable and have given Sigma Designs' ample opportunity to resolve this issue. Apparently none of our demands have been taken seriously. Nearly two months after the initial release of the REALmagic MPEG-4 Video Codec, Sigma Designs' is still knowingly infringing the GNU General Public License."
Militzer believes this infringement might be of high general interest: "This is an unfortunate event, not only for us but for the whole Free Software movement. Therefore we hope to receive wide support from the Free Software community in our efforts to convince Sigma Designs' to respect the terms of the GPL."
Evidence supporting the claim has been published on the XVID website.
http://www.xvid.org/v1_0_comparison.pdf
http:/
About XVID (http://www.xvid.org/)
XVID is a leading open source MPEG-4 video research project, founded by the German student Michael Militzer in August 2001 to continue the efforts of DivXNetworks' former OpenDivX project. Today, the XVID project consists of users and developers from all over the world. XVID publishes all its software under the GNU General Public License (GNU GPL).
About Sigma Designs Inc. (http://www.sigmadesigns.com/)
Sigma Designs' headquarters are located in Milpitas, California. The company specializes in MPEG based video hardware for encoding and decoding. Recently Sigma Designs' introduced the Xcard, the first consumer hardware MPEG-4 decoder in the form of a personal computer add-on card.
About GNU GPL (http://www.gnu.org/copyleft/gpl.html)
The GNU General Public License is the most frequently used software license for Free Software development and is supported by the Free Software Foundation (FSF). Software distributed under the GNU GPL grants everyone modification and redistribution rights, on the condition that derived or redistributed software carries the same license.
Contacts
For contacting the XVID team please use the e-mail addresses contact@xvid.org or contact@xvid.de
Please address your request to one of the following persons:
Daniel Smith (USA)
Michael Militzer (Germany and international)
Christoph Lampert (Germany and international)
Edouard Gomez (France)
As someone who once worked for Sigma Designs... (Score:5, Informative)
Any Questions? (Score:5, Informative)
MILPITAS, CA, (August 12, 2002) - Sigma Designs (NASDAQ: SIGM), a leader in digital decoder solutions, announced today that the Company will be discussing second quarter results during a conference call on Tuesday, August 27, 2002, at 4:45 p.m. Eastern time. The dial-in number is (612) 332-0226. A question and answer period will take place at the end of the discussion. The earnings release will cross the wire at the close of market on the same day.
The call will be webcast live from www.vcall.com. An audio replay of the call will be available shortly thereafter the same day and will remain on-line for 30 days. For further information, please see the link to this site on our website at www.sigmadesigns.com or email investor relations at IR@sdesigns.com.
Re:I think... (Score:3, Informative)
I'd say it's certainly possible but what XVID has demostrated has me in awe. That's damn near impossible.
As long as they can demostrate large and complex blocks of assembly which are identicle then you're right: there's almost no way that could happen with two independant code blocks.
However, it is possible, just not to this extent. An exmaple would be if two blocks of code independantly implemented quick sort or Euler's algorithm. In that case I would expect those two peices of assembly (provided that they were compiled with the same compiler) to most likely be identicle beacause those are algorithms that are so popular and refined that almost every developer implements them in the same way.
But in this case XVID has certainly demonstrated the large and complex blocks of identicle assembly to prove that they were ripped off. And not only that but that the pieces of identicle code occur at almost the same offsets in the dll!! That's a very good indication that Sigma has extended on, or simply modified, XVID's original code base.
I sure hope they can somehow find the resources to sue Sigma's pants off. I will certainly make a donation if they go that route.
--
Garett
Comment from the FSF (Score:5, Informative)
Re:GPL Powerless (Score:5, Informative)
If someone plagarizes your work, sue them. The only information the judge is going to need is a copy of their source, and yours. Are they the same? Judgement for plaintiff.
Finding a couple of pro-bono expert witnesses in this case should be a snap, if that's even necessary. Hell, ask Stallman, he's always looking for a pulpit.
Over the years, I've had three or four clients who didn't want to pay for work I had performed. A couple of them even said "We're ready to be the 800 pound gorilla on this matter." (That's a direct quote from one, BTW).
Ok, you be an 800 pound gorilla. I have all my notes, all the specifications, all the correspondence related to the project ready to go. I have notes on every phone call, every meeting, every conversation. It costs me $40 to file, and all I have is time. If you want to tie up your $150 an hour lawyers for six months fighting an angry badger about a $20K project, go right ahead.
Funny, the check always shows up after that.
Don't let people push you around because you're a small operation, or because they think having more money guarantees them victory through intimidation.
K.
Re:I'm not surprised (Score:2, Informative)
So, yeah, the bully is in the playground, but the XVID has a big brother he can call on.
Dirty programmers (Score:3, Informative)
In our case a competitor heard we were working on a program with some features similar to theirs. So to try and create the Dirty Programmer situation our competitor sent copies of their program to our developers trying to get them to look at it. Lucky for us the developers went to management and they went to legal department. Legal collected all the copies of the program and had a hell of a chat with our competitor.
Statutory damages for willful infringement (Score:3, Informative)
So suing Sigma Designs is not necessarily a futile effort, though of course it will take some resources and I don't know if it will happen. As for me, when I use the GPL on something interesting, I generally assign it to the FSF, so the FSF can then take action if necessary against infringers.
YES! I read it carefully - it's a smoking gun!! (Score:5, Informative)
Standard optimization trick (Score:3, Informative)
So compilers will often unroll a loop and add a jump to the correct starting point at the start. Even unrolling a loop once (duplicating the code once) can have a noticeable impact, and I think 4x and 8x are most common. 16x seems unusual, but it could be a high optimization level.
DivXNetworks also sad to hear about this (Score:2, Informative)
Re:As someone who once worked for Sigma Designs... (Score:5, Informative)
This was in 1998. Well after software-based codecs were freely available (Microsoft shipped ActiveMovie with MPEG-1 playback support back when it released Internet Explorer 3.0 somewhere around 1995? 96?)
I did a demo for the deparment head showing two systems side-by-side, one with Signa's REALMagic card and their codec, and the other with Windows Media Player installed. Ironically, the guy picked the software based solution as the "higher quality" solution and said it was worth the extra $200.
Needless to say, when the labels were revealed, they immediately cancelled the pending order for 6000 REALMagic cards, a savings of $300,000.
Sigma Designs seems like a bunch of snakeoil salesmen to me.
- JoeShmoe
.
Re:ehmmm... why the shouting? (Score:3, Informative)
-- Bob
Re:GPL Powerless (Score:2, Informative)
Re:unanswered question (Score:4, Informative)
"why have they ceased development?"
If you were painting someone's house, and you
found out for certain that they had no intention
of ever paying you, would you keep painting the
house? Or would you take your ladder, your paint,
and your brush, and go elsewhere?
What if you were painting the house, and you found
out that not only were you not going to get paid,
but that the person was going to then take your ladder,
your paint, your brush, your truck, and call the police
to have you arrested for trespassing?
Would you stay and finish the job?
How would you feel about this story if the Sigma folks
had finished their product, and then turned around
and accused XVID of stealing THEIR work? Who's to
say they STILL won't try that?
Sigma has released to code (Score:3, Informative)
MILPITAS, Calif.--(BUSINESS WIRE)--Aug. 22, 2002--Sigma Designs, Inc. (Nasdaq:SIGM - News), a leader in IP video streaming solutions, today announced the release of the source code behind its free MPEG-4 video CODEC that works as a plug-in under Windows and encodes digitized video content into fully compatible ISO MPEG-4 video files. The complete source code will be available for download starting August 23, free of charge, through Sigma's website (www.sigmadesigns.com), to support developers wishing to enhance the MPEG-4 encoding.
"We are pleased to provide the development community with an open source MPEG-4 CODEC, and anticipate that this will accelerate technical improvements and enhance the proliferation of MPEG-4 content," stated Ken Lowe, Sigma Designs' vice president of business development.
About Sigma Designs, Inc.
Sigma Designs specializes in silicon-based MPEG decoding for streaming video, progressive DVD playback, and advanced digital set-top boxes. The company's award-winning REALmagic Video Streaming Technology is used in both commercial and consumer applications providing highly integrated solutions for high-quality decoding of MPEG-1, MPEG-2, and MPEG-4. Headquartered in Milpitas, Calif., the company also has sales offices in China, Europe, Hong Kong, Japan, Korea and Taiwan. For more information, please visit the company's web site at www.sigmadesigns.com/.
Re:Understanding GPL is not trivial. (Score:3, Informative)
IANAL, but I believe the answer is: when your software is considered a derived work of GPL'ed code [gnu.org], according to the copyright law of the land.
There is, as far as I can tell, no "bright line" between when this does and does not happen, nor can one be made to exist, short of "everything infringes" or, more practically, "nothing infringes".
I've written about this principle in length on USENET's gnu.misc.discuss group [google.com] in the past, if you want to search for my posts from years ago.
But the main thing to remember is: just because we're dealing with technology here doesn't mean we can expect to, or expect the law to, draw us a nice, technologically clean "line" between infringing copyright and not infringing it.
So, the question being "does my program derived from GPL'ed code?", two things, at least, must be answered:
-
-
As you should be able to infer from the above, these questions cannot be trivially answered by resorting to redefining "work" as "single linked executable", since a court might reasonable rule that the work actually consists of two or more executables (or binaries generally) cooperating so closely as to consistute a single work, making that collective work a derivation of GPL'ed software if any one of its components is.What actually constitutes your program, which, in copyright terms, is the "work"?
Does that work contain a substantial portion of someone else's GPL'ed code?
Anyway, since the GPL "protects" code no further than copyright law defines "derived work", it cannot definitely answer the question.
Instead, all it can do is limit the degree to which copyright law's view of a "derived work" might extend beyond what the GPL intended to protect.
Hence things like the "mere aggregation" clause.
Re:Sue (for money) for GPL violations? (Score:3, Informative)
Specifically, see 17 USC 504 (b): you choose between actual damages and statutory damages, and profits of the infringer due to the infringement count as actuals.
Re:Sigma still doesn't get it. (Score:2, Informative)