The disease of standards

Posted 23 Aug 2001 at 12:01 UTC by muks Share This

JPEG2000 is considered an important upcoming standard, which adds to what JPEG already provides. But should we use and support it? Are we safe in using codecs such as OpenDivX and DivX ;-)? More importantly, do we need use these technologies?

I call you all to boycott the JPEG2000 and MPEG-4 standards. These standards are plagued by patents, which disallow their free use. Please do not use, purchase or sponsor any work based on these standards. They are only "open" in the sense that their specification is available for view for all upon payment. They are not open allowing freedom of use by all. Standards such as JPEG, MPEG-1, MPEG-2 also include patented work. (The Independent JPEG group's JPEG sofware, which is used for JPEG compression in products such as Mozilla, Netscape Communicator, Microsoft Internet Explorer, GIMP, etc. manages to avoid patents on arithmetic-coding held apparenly by IBM, AT&T and Mitsubishi, by disabling the coding processes involving it.)

Video and audio coding are full of patented work. However, we expect standards, especially those created and recommended by organizations such as ISO to be free of this disease. I cannot help but think to myself about an open conspiracy behind such standards.

It has been done before. Ogg Vorbis claims to be patent-free audio. Xiphophorus, an independent open research group, far smaller and with lesser resources than the MPEG consortium could create a high-quality audio coding format free of patents. Why couldn't the MPEG committee do so? Why did they choose to incorporate *SO MANY* patented methods which involved a multitude of companies?

In fact, a few patents on standards such as JPEG2000 were applied for *AFTER* the decision was done to include the method in the standard.
What does this suggest to you?

Why? Why can't the ISO enforce policy stating that no patented work should be included in standards released by it? And that no work released in the standards should be patented after inclusion? Why should we support such an organization?

We can live without these standards, rather than suffer later. Think and ponder if you want to contribute to projects such as OpenDivX, MP3 encoders, etc. and other projects derived from them. Most of the underlying technologies are patented which require a per-copy license. An application conforming to the Open Source Definition will spawn many many source code and personal binary copies for which the author cannot pay license fees. Companies owning these patents will enforce their rights sooner or later. Did anyone forget Unisys and GIF yet?

We don't need to infringe on patents and "get away with it", even for personal use. There are free alternatives available. We know the kind of quality, our community can offer. Support them. Boycott patented technology. This is not for so called geeks of the community alone. Even if you are a large corporation or company, you should help the cause.


Potent Patents, posted 23 Aug 2001 at 12:35 UTC by robster » (Journeyer)

In the UK and many other European Union countries (with Germany being the exception) patents are much further restricted, in that you cannot patent software and ideas. However the possibility of the establishment of the Hague Treaty, threatens this patent veracity.

The Hague Treaty is not directly related to patents however it allows in theory any jurisdiction applicable in one country to applied in another, for instance a court ruling stating the infringement of patents could then be taken and applied in any other Hague Treaty country.

We should repel all attempts to adopt the Hague Treaty without special exclusions for copyright/patent laws, as these differ wildly between countries.

Open Standards from organisations such as ISO should be patent free, however charging for the commercial use of trademarks, standards certification should be encouraged, as the organization has spent time developing the standard. Standards should not be within the grasp of commercial interests.

Lots of carts, too few horses, posted 23 Aug 2001 at 18:04 UTC by Bram » (Master)

Standards are often adopted out of the belief that if it's a new standard it must be better, rather than a compelling engineering reason.

Far too many projects get bogged down in 'standards-compliance' following guidelines set by a self-appointed body which have no noticeable benefit.

why couldn't they do it?, posted 23 Aug 2001 at 22:20 UTC by splork » (Master)

They could.

But the standard committees are always composed of corporations who ultimately do things to serve their interest. So given a choice between two algorithms, the one that is related to a patent or three that a committee member has a stake in gets "chosen." Technical merit of the algorithms no longer enters into the picture at that point.

I also wonder how patent holders can get away with the current tactics of waiting for years specifically not enforcing a patent until enough people are using it without a license they they say, oh, by the way, now that you're dependant on that you need to pay us. "Come on, try one, the first one's free!"

Isn't there a free license for JPEG 2000?, posted 24 Aug 2001 at 02:56 UTC by wmf » (Master)

My understanding is that the patents necessary for the basic part of JPEG 2000 are licensed to everyone for free. That's not as good as if there were no patents covering it, but at least it shouldn't prevent implementations.

insufficient improvement, posted 25 Aug 2001 at 04:04 UTC by rillian » (Master)

Regardless of the patent situation, I've never been convinced jpeg2000 offers a sufficient improvement to warrant adoption. You really want something like a factor of 2 in compression before switching, otherwise the curatorial issues dominate. You're better off keeping the old format to simplify things for folks trying to read your old data--especially with a lossy format where you can't just transcode. That would be reason enough to ignore the format even if there aren't patent issues.

Contrast this with bzip2, which is a similar improvement in compression ratio for text-oriented data. In this case I think it is worth offering, not because we shouldn't care about keeping old source archives readable, but because it's a lossless format so data can be trivially migrated. The usage patterns in this case are quite different as well; source is more "alive" than images tend to be. Distribution sites often offer both .gz and .bz2 (and .zip, etc.) One is also more often interesting in copying the latest and greatest over the network, and I tend to be more annoyed by the extra 30% in transferring 10MB of kernel source than in 100kB of image.

As far as mpeg-4 goes, I think we should encourage them to publish the darn thing and get their business in order. I've heard persistent rumours that H363+ (or whatever the low-bitrate video component is called) might be patent free. If so that would be great, but we have know way of telling until the dust settles.

I have some sympathy with users of mpeg-4. It really it almost a factor of 2 better than mpeg-1/2 (which have patent problems of their own) and at least it's trying to be an interoperable standard. Furthermore, there are no free alternatives, and several ones more proprietary. Certainly we ought to keep working to remedy that situation, but in the meantime it's looking like the lesser evil. OTOH, I think mpeg-1 is still a perfectly reasonable format for distributing short clips, again because the fractional increase in download time isn't onerous, and the format has extremely broad support including multiple open source decoder implementations. So I would urge everyone to make it an option.

Different approaches, not that many solutions, posted 25 Aug 2001 at 11:44 UTC by DV » (Master)

It really depends on the standard body, they set the rules for admission in the Working Groups who are defining those standards. I can't remember for IETF, at W3C at least now it is clear that if you want to participate in a standardization group, your company has to make an IPR statement upfront. Seems that ECMA is a bit more lax (was raised for the C#/Mono case) and they accept partitcipation without clear IPR statement until the last moment (I'm afraid this gives Microsoft plenty of time to try to lock it, we will see).

There is also exit rules which may be different, at ISO once the group produced a last version I think there is a public challenge and if someone comes with another spec proved better then that one is selected as the standard, even if locked with patents :-\ . At W3C and IETF I think the Patent holder must accept free usage of their patented technology but just for the implementation of that specification. I don't know about ECMA for this.

In general ISO just don't care about standardizing when Patents actually covers the work produced. This led a lot of people to believe in the standard community (it's really a world of it's own) that it was just fine, and also take the bad habbit to try to patent anything which is submitted for standards (or lobbying for stuff they have Pending Patents the best case because it's not public yet).

Those people are evil bastards, but hey that's what happens in a world without laws to keep people behaving. The first thing would be for the US to fix (i.e. close until they fix it) their Patent Office, and second would be to lobby all governement to get ISO and other Standard bodies to fix their practices, ISO status depends for a part on the fact that governements requires ISO status for a lot of things, governements are the only ones who could change the rules of the game in that area.

Conclusion: the only solution is unfortunately political, asking for veto here is very unlikely to have any effects, annoying your political representative *and vote* are far more likely to have effects, even if chances are minimal :-\

W3C, posted 26 Aug 2001 at 02:42 UTC by mnot » (Journeyer)

If you care about these issues, now is the perfect time to give feedback on the W3C's proposed patent policy; see

Potential free alternative to MPEG, posted 26 Aug 2001 at 02:49 UTC by wmf » (Master)

On2 claims that they'll be releasing VP3 as open source soon. Their proposed license has a clause requiring derivatives to be compatible with the original, which may or may not turn out to be a problem.

Ogg Tarkin, posted 27 Aug 2001 at 10:09 UTC by Radagast » (Journeyer)

There's also Ogg Tarkin, the video companion to Ogg Vorbis (both of which use the Vorbis stream format). I've been lurking on the mailing lists for a while, and although development (or rather research and prototyping at the moment) is slow, there are some very dedicated and talented people working on it. They've also considered looking at the On2 codec mentioned above, I know.

As for whether or not this is needed, I think it definitely is. While JPEG2000 might not be a big enough improvement to warrant upgrading (although the samples I've seen seem rather promising, in particular in terms of subjective image quality), there's a definite need for a free, unencumbered video codec that covers a reasonable range of bandwidths.

I strongly believe that the next frontier in free software is going to be media-oriented. More than just MP3 players, this means we need all the pieces of a complete toolkit to work with sound, images and video. Additionally, the codec market is lucrative; many people are paying very high prices to be able to stream their data (either in licenses for the streaming server, or licenses for the encoding software, or other schemes). A set of high-quality, free codecs could potentially take this market by storm.

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!

Share this page