Xiph.org Comments For the FTC's Patents Workshop

Freddybear writes "Xiph.org, makers of ogg audio and theora video codecs, submitted a detailed proposal to the FTC for the patents workshop. Their proposal recommends changes which would help to eliminate the practice of 'submarine' patents regarding standardized technologies. Quoting: 'The Xiph.Org Foundation recommends that the FTC work to require specific, ex ante disclosure of patents or patent applications that would read on standards under development, that failure to disclose exhaust the patent, and assertion of such a patent ex post be deemed anti-competitive. This should apply not only to standards development activities that the patent holder participates in or knows about, but those it should have known about. Furthermore, vague infringement allegations or activities designed to avoid an SSO's disclosure requirements or undermine the standards process should also be deemed anti-competitive.'"

Comments

I wish to support this submission

[Anonymous Coward]

I would like to add my name to a list of people who support this submission.
Does such a list exist?

Re:I wish to support this submission

[davester666]

Yes. Everybody that cares monitors /dev/null. Pipe your name there.

Re:I wish to support this submission

[stms]

Yeah but looks like you've already signed Mr Anonymous Coward.

Re:I wish to support this submission

[Sunshinerat]

I would send money to Xiph.org, and this time not anonymously.
Money to help fight the fight.

Thank you xiph.org

[jhoegl]

I appreciate the time spent by xiph.org to think about and put together the information. Unfortunately I know little about patent processes, but I know that patents have been a major concern and pain, because one can, through no fault of their own, create a "method" like something that had been patented.
I also know that there are many patent troll companies out there and they need to be taken care of, regardless of how much the government thinks it brings in on fees during these processes.
I understand parts of what was said, and found nothing that I can disagree with. Although I would rather they do away with software patents completely, in our reality that will not be the likely case.
Thanks xiph.org.

Re:Thank you xiph.org

[bzipitidoo]

I read it. Yes, it's complicated.

Xiph chose to focus on a small part of the problem: how patents make it hard to create standards. They gave a number of examples and scenarios. Then they suggest a few small reforms that may help a little bit. Rather than advocate against the patenting of software, they aim for more modest changes that they hope make it impossible to hold a standard hostage years later with submarine patents. Surface immediately, or lose all right to challenge the standard.

Is Xiph's recommendation a good idea? I don't know. Sounds good as far as it goes. But I think they do not go far enough. I fear clever legal experts will find ways around Xiph's solutions. And that they will turn those solution against us. Large companies could use Xiph's proposed mechanism to make an end run around any number of patents. Just make up another standard, and never mind whether the standard is ever adopted, as the point is to quash related patents. Would little folk be able to use this technique? And who decides what is a standard, what the requirements are to create a standard? Are we going to appoint ISO as our new overlords to decide that? MS's OOXML campaign springs to mind as an example of what could go wrong. In contrast, if we simply eliminate the patenting of software, we don't have those issues.

My suggestion is....

[Joce640k]

a) The patent office needs to employ some experienced, unbiased software engineers who understand the word "obvious".

b) They need to break the link between "profit" and "number of patents accepted".

c) They need to make it much easier to invalidate a patent due to prior art and obviousness (this isn't necessarily a good thing but it's needed because of all the "a" and "b" they've done in the past).

Re:WTF???

[MysteriousPreacher]

That means that if a company, no matter how big or small, ever files a patent that relates to video compression, they have to suddenly join every single standards body just so they can tell those standards bodies that things they're looking into infringe on patents? At what point is this company supposed to move on and do engineering work?

I believe these companies would only be under an obligation to disclose patents if they are active in the setting of a related standard. The goal here us to prevent people from helping to develop a standard reliant on a patent they just happened to forgot they held.

And My suggestion is....

[SmallFurryCreature]

A) Joce640k Should vote for the increase in salary of goverment employees and not vote for the guy who offers the biggest tax cut.

B) Joce640k Should stop trying reduce spending by trying to make institutions be profitable regardless of the effects.

C) Joce640k Should stop voting for senators that are wholly in the pocket of big business just because they promised him a tax cut.

The patent office used to work, then cut backs came along. Now it doesn't. Who voted for the cutbacks?

Re:And My suggestion is....

[rtfa-troll]

It's important to remember also that quite a few of the things the patent office does nowadays they do because of successful appeals against their patent rejections. E.g. the standard for "obvious" is at least in part determined by patents that the office rejected and then lost lawsuits about.

Clear Path to the Public Domain

[Jahava]

I propose the following change to the current patent system:

• When a patent is initially filed, the patent filer may optionally include an itemized list of costs incurred to directly develop the patent.
• At any time after the patent is granted, the USPTO offers the following option: If members of the public can accumulate and pay 150% of the stated development cost of the patent to the patent holder, then the patent irrevocably enters the public domain.
• Failure to disclose development cost will result in a default value (around $100,000, can be raised by USPTO as needed) being assigned to the value of the patent.
• Misleading or incorrect information on an itemized list disqualifies the list and results in default value being assigned to the patent.
• Challenges and negotiations regarding the value of a patent can be brokered in a public setting through an institution established by the USPTO.

Under this system, inventors have a clear path to profit from the effort they invested to create a patent. No matter how much they invest, they will always make 50% of that investment back in profit. There is also a clear path to the public domain for the patent - anything so fundamentally critical can be purchased and contributed to the public domain with the USPTO as the intermediate broker. It is likely more profitable for any given company to place a patent in the public domain for all to enjoy than it is for them to license it from the individual company.

The buyer of the patent can be a company, community pool of money, or even the US Government itself (think cancer cure) based on the criticality of that patent to any entity's set of interests.

So patents aren't gold mines anymore. You can't build a business model around exclusivity. Who cares? Innovation will continue, as it always has, and now everyone can participate. I dunno; I like the idea.

Re:Clear Path to the Public Domain

[soren.harward]

This system already exists, and already functions better than it would if the USPTO were involved. If you want a patent to go into the public domain, do the following:

1. Start a nonprofit organization, like a 501(c)(3) in the USA. Legal costs for doing so are minimal compared to the purchase price of the patent.
2. Raise the money.
3. Buy the rights to the patent.
4. Either allow the patent to expire by not paying the maintenance fee, or by explicitly disclaiming the remaining term of the patent and granting the rights to the public domain.

Not only does this procedure not need the involvement of the USPTO, but it has the added benefit of you not needing to pay any more for the patent than the selling price the owner is willing to accept.

On the other hand, determining the "selling price" is not quite as simple as "150% of the development costs". The value of a patent is not that it is a way of recouping your development and patent prosecution costs; there are plenty of accidental discoveries that turned out to bring in a lot of revenue from very little development costs. The value of a patent is that it gives you an exclusive right to an invention. So the true value of a patent is the revenue it will allow you to bring in over its lifetime, from sale or licensing, or from litigation awards. Most companies also treat patent acquisition as an arms race: as long as you and your closest competitor both have big patent portfolios and are each infringing on the other's patents, and as long as patent litigation is as costly as it currently is, then you can have occasional skirmishes with out-of-court settlements being traded back and forth while avoiding an all-out infringement litigation war.

So, in neither case is a patent owner going to be enthusiastic about selling if a nonprofit shows up and says "we'd like to destroy your patent at the price of its development costs". To buy the patent, you'd have to cover the cost of anticipated revenue from that patent and the cost for having lost a deterrent against litigation.

Re:Clear Path to the Public Domain

[evanbd]

Most ideas don't pan out. This applies to patent-worthy, reasonably thought out ideas as well. Plenty of things look perfectly reasonable at the patent stage, but just don't quite work for one reason or another. This causes a problem for your idea: if I put in the effort to develop ten patents, all of them reasonable, and pursue them further, perhaps one or two or three will actually turn into a saleable, profitable product. If I can only recoup 1.5x my costs on those 3, and nothing on the others, I lose money.

Patents are, in general, a bet on an unlikely outcome. Much like startup companies. Most of them fail, too. In order to make that work, you need a possibility of a high return. Maybe not an astronomical return, but a high return. Of course, none of these comments are relevant to patent trolls and such, but the problem isn't quite as trivial as you suggest.

Re:So what will happen?

[rtfa-troll]

"Defacto Standards" will be all that remains.

Why? Unless I'm missing something, your comment seems a bit nonsensical. This should actively encourage creating public open standards. If you want your product to be safe from patents, then the best way will be to state clearly how it works in a patent. Basically you would want to do that for every part of the products functionality apart from those bits where you can get a patent yourself or keep the functionality secret.

length

[Anonymous Coward]

With the pace of competition today, why do patents last more than 2-4 years?

Patent duration

[Crouty]

Because there are developments that require years and millions of $, think pharmaceuticals.

Software must not be patentable. Why? Because with software patents it is possible to monopolize mathematical concepts, see http://vimeo.com/11206098

Re:So what will happen?

[sxeraverx]

The "standards" will be open, sure, but no one will use them. ODF is an open standard. MS uses OOXML, and would use it even if they hadn't messed with ISO. And just because they are the largest player in the market, OOXML would become the de facto standard that any other office suite would need to support. It's true this isn't a patent case, but I'd imagine the same applies.