Comment by WildUtah | Hacker News Reader

Comment by WildUtah | Hacker News Reader

Parent Full thread

Of course you can't use this compression technique without being sued. And that is the key reason software patents are really bad.

You may hear a lot about trolls and grossly invalid obvious patents making millions in licensing and extortion by legal fees and I agree those things are bad. They are very bad. But the real problem is the apparently valid patents.

Monopolizing a programming technique is a long-term barrier to invention that provides little or no incentive to create anything new compared to all the invention it prohibits. It's the same in any field where experimentation doesn't require billion dollar research teams and each new idea is built on thousands of others. No new work you do in video, audio, or image processing or encoding will be legal without a mountain of licensing agreements. Anything beyond common established software applications will be unlicensable and effectively outlawed. Google had to fight for years with a giant legal team to get permission to release WebM and you don't have that kind of power.

> It's the same in any field where experimentation doesn't require billion dollar research teams and each new idea is built on thousands of others.

So how little money do you believe has been spent creating MPEG, H.264, H.265, and all the other codecs? Do you think these were created by a handful of people in a garage over a weekend?

Meetings for these codecs were more than 300 people and those were just representatives of larger groups. This is research that does require billions of dollars and thousands of experts working for years. That's not going to happen as a charity or by weekend hackers. Linus isn't going to get frustrated at the size of video files and take a week off to come up with a better codec.

Software patents created these video codecs by funding the research and development, and patents aren't long-term barriers as many of the early formats have expired or are expiring now. You want to use MPEG-1? Go ahead. Or if you want to leverage hundreds of thousands of man-hours experts put into improving on it for the next generation then for a few more years you'll need to pay for it to cover those costs. That's exactly how patents are supposed to work.

> That's not going to happen as a charity or by weekend hackers. Linus isn't going to get frustrated at the size of video files and take a week off to come up with a better codec.

That's kind of what he did with Git.

> That's not going to happen as a charity or by weekend hackers.

To be honest that's exactly what happened with Vorbis (1 FOSS guy) and Opus (2 FOSS guys, 1 Skype guy).

The MPEG process is full off politics and massive overheads.

Those are audio codecs, not video codecs. Xiph worked on Theora for 7 years and never got close to H.264, and Daala is promising but HEVC is amazing right now.

Theora and Daala are both hamstrung by working around minefields of patents on essential processes to video processing.

Interestingly, the Vorbis project only really took off after mp3 started facing licensing issues. From http://en.wikipedia.org/wiki/Vorbis

> Intensive development began following a September 1998 letter from the Fraunhofer Society announcing plans to charge licensing fees for the MP3 audio format.

Who says patents don't promote innovation? :-)

(I'm only half-joking - invention forced by working around patents has long been a post-facto justification of patents.)

Also, nit, it wasn't all done by just "1 FOSS guy". From the same wikipedia link:

> Chris Montgomery began work on the project and was assisted by a growing number of other developers.

In my case, a substantial amount of my work on Opus was done literally from my basement (sorry, I didn't have a garage).

To be fair a lot of the research that lead to these codecs was made by publicly-funded research institutions.

Spending big money and organizing meetings with many people does give you right to prevent others to invent their own codecs? Does Linus sue other OS makers, that they make OS kernels too and therefore they must (by definition) infringe on his rights?

And according to my math, I should be able to use MPEG-2 (publication date 1996), AC3 (1995) or MP3 (1992) already, with MPEG-4 ASP (1999) getting there soon.

> Spending big money and organizing meetings with many people does give you right to prevent others to invent their own codecs?

No, making a novel and non-obvious contribution to the arts gives you the right to apply for a patent on your invention, which then gives you the right to prevent others from using that specific invention. Spending big money and organizing meetings may or may not be involved.

Nothing prevents others from inventing their own codecs.

Now commercializing those codecs may be another matter entirely, probably requiring resolving licensing issues depending on how much they rely on other patented methods. Ostensibly, pools like MPEG-LA exist to make this easier.

> Now commercializing those codecs may be another matter entirely, probably requiring resolving licensing issues depending on how much they rely on other patented methods. Ostensibly, pools like MPEG-LA exist to make this easier.

That would be fine. However, confusing the issue, claiming infringement by these new codecs in general, without providing any proof and otherwise trying to destroy the adoption of these new codecs is not OK. MPEG-LA and other right-holders conspiring to damage the alternative codecs uptake did cross this line.

Meetings for these codecs were more than 300 people and those were just representatives of larger groups....

That's easy to do when your goal is to have as many different people and patents involved as possible, rather than simply develop an efficient codec using an efficient process.

Reply on news.ycombinator.com