How Google’s Open Sourcing of VP8 Harms the Open Web

Much of the initial commentary on Google’s open sourcing of the VP8 codec it acquired in purchasing On2 has breathlessly, and uncritically, centered on the purported game-changing impact of the move.

But unfortunately, these commentaries miss an essential point that Google has studiously avoided mentioning the need to standardize royalty free codecs (not just release an open source snapshot).

But since forward motion is good simply because it is forward motion, shouldn’t one hesitate to look this gift horse in the mouth?

Unfortunately, in the case of multimedia codecs and technologies, ignoring open standards and instead presenting open sourcing as a fait accompli solution just works to the detriment of the entire open community.

The open Web needs royalty free standards (true, multi-stakeholder run standards, not unilateral actions) — that is its essential genius.  And without them, proprietary, vendor-controlled projects, even those that self-label as “open”, do little good and more likely more harm than good.  We all have the right to expect, and demand, that the Web’s current beneficiaries and leaders stay true to this fundamental open standards proposition, and not just forget it when convenient.  And this includes Google.

It is well known that many experts consider it now feasible to standardize serviceable royalty-free codecs.  MPEG (the standards group, not the unaffiliated license administrator MPEG LA) has even put out a resolution to that effect, and IETF has recently launched a royalty free codec activity in a similar spirit.  Google should get on board on this important trend, not undermine it with studied avoidance.  So far they have not.

It is important to understand that patent claims are typically handled under confidential non-disclosure agreements.  So unless there is a forcing function (litigation or standardization-required disclosure and review), there is no effective way to know who is actually claiming, and who is paying, what.  And there are documented cases of this going on for literally years.  So leaving VP8 code out in the open with nothing but a mutual non-assert license leaves the patent issue not only unaddressed, but up for capture by those with uncharitable agendas, and on their turf and time frame (let’s at least hope that’s sooner rather than later — but remember, forming patent pools rarely disclose all their patents up front).

Not a smart move, and hopefully one Google will realize the error of and correct quickly (here’s a useful cover story: we intended to smoke out patent holders all along, and we were going to get around to working with standards groups when we had the chance).  Contributing VP8 to a standards group with a strong patent disclosure policy would be a good corrective move; it would force lurking patent holders to come fully into the public. Not perfect, but a step forward.

Google’s open sourcing of VP8 is very different from Sun’s Open Media Stack codec work, and for that matter other responsible open video initiatives, which have based their work on identifiable IPR foundations, documented their patent strategy, and have been willing to work with bona-fide standards groups to address and resolve IPR issues.  When companies like Google ignore standards and go on their own in such important areas as video codec standards, they just undermine the very standards groups the open Web needs to thrive and grow.

We’d never accept a brand name company unilaterally declaring control of the next version of TCP/IP, HTML, or any other of a host of foundational Internet and Web standards simply by open sourcing something they’d bought.  Codecs will also be such a foundational component, a critically important one.  Just because the technology of codecs might be less familiar than some other technologies is no reason to abandon the royalty-free standardization philosophy that has built the Web.

Certainly not based on the complete feel-good-marketing non-explanation for this radical abandonment that Google has offered so far.   Because patent pool licensing is out of control?  No argument about that from me (or antitrust complainants Nero, VIZIO, and others).  Because Google “must have done its patent homework”?  OK, if so why not hand that homework in as a contribution to a standards group where it could get some expert scrutiny?

So I would encourage Google to do the right open standards thing — join and contribute to responsible standards groups that are working to solve the royalty-free open standards need.  Be a part of the royalty-free, open-standards solution, not part of the problem.

UPDATE

Tip of the hat to Xiph’s leader, Chris Montgomery, for good tongue-in-cheek humor:

I, FOR ONE, WELCOME OUR NEW WEBM OVERLORDS

Not to confuse: Xiph is wholeheartedly supporting WebM, but another interesting remark by Montgomery:

“But Monty isn’t worried about the MPEG-LA suing him or anyone at the WebM Project.

“The recent saber-rattling by Jobs felt more like a message to his own troops than
a warning shot to ours,” he says. “MPEG itself has always has an internal contingent
that has pushed hard for royalty-free baselines from MPEG, and the missives about
video codecs and patents were probably meant for them, not us.”


Comments

6 responses to “How Google’s Open Sourcing of VP8 Harms the Open Web”

  1. It’s kind of reminiscent of the W3C vs. WHAT-WG web standards spat a few years back.

    For those unaware: the W3C (in charge of web standards) kind of lost its way for a while, working on an improved HTML called XHTML that was of little or no use to the real world. Much of their standardisation energy was diverted by telcos working on lucrative content delivery schemes within the walled gardens for the mobile web (i.e. not the actual real web you get on an iPhone, or even with Opera Mini or Mobile, but WAP stuff and things that the carrier charged you for viewing).

    Rather that just sit by and let this disaster unfold, Apple, Google, Mozilla and Opera set up their own group outside of the standardisation process and started work on what would become HTML5. After a few years of rapid progress the W3C realised they’d picked the wrong horse and HTML5 was brought back into the fold, though not without a lot of wailing and moaning.

    I’m guessing that’s at least partly what Google is hoping for with this move.

    1. Rob Glidden Avatar
      Rob Glidden

      Dave:

      The Whatwg group analogy may or may not be apt — it was a different era in 2004 for founders and original copyright holders Apple, Mozilla, and Opera, when the group was announced “with the intention of submitting the results to a standards organisation. This submission would then form the basis of work on formally extending HTML in the standards track” (http://www.whatwg.org/news/start, http://www.whatwg.org/specs/web-apps/2005-09-01/). Google and WebM may have such an intent, but it has not been yet stated. Inclusion in Android, though, has been announced.

  2. Also, what’s the chance of Brazil (or other similar market) adopting this for Digital TV or a shiny-disc based format similar to SVCD? Any thoughts?

    1. Rob Glidden Avatar
      Rob Glidden

      Dave:

      If by “Digital TV” you mean over the air free broadcasting, no country I am aware of has ever adopted a proprietary solution instead of a standard for codec (perhaps someone can site an exception, I’d be interested in hearing of one). Of course this is not true for subscription TV services, but even then I don’t know of any that did not end up adopting a standardized codec (albeit royalty bearing). There is a wider variety in optical media formats.

  3. Tab Atkins Avatar
    Tab Atkins

    Rob,

    You seem to be either mistaken or naive on the ability of a standards body to produce a good result when patents are involved.

    In particular, let us look at this statement from your article: “Contributing VP8 to a standards group with a strong patent disclosure policy would be a good corrective move; it would force lurking patent holders to come fully into the public.”

    Standards bodies have no power over companies that are not members of that body. For example, if VP8 were given over to the W3C, only W3C member companies, and particularly only those who had joined the relevant Working Group working on the problem, have to disclose any patents. Any other company can merrily ignore the call for disclosures.

    The only reason to join a standards body in the first place, and the working group in question, is if you want a hand in the development of the standard around the technology. If you’re a patent troll, though, you have no incentive to do so. You want to ensure that the standard incorporates some technology that you own a patent on, but beyond that you have no interest in contributing to the technology, and would in fact prefer that nobody knows of your interest so that you can stealth-sue sometime in the future.

    So, having the technology owned by a standards body would not guarantee, or even strongly suggest, that all relevant patents are known about.

    Further, “disclosing” is not the same things as “licensing royalty-free”. A company can disclose a patent they hold during the call for disclosures, and maintain their license. This merely means that they reserve the right to police their patent, and in practice means that the standard must avoid infringing on that patent.

    So, even among the companies who are legally bound to disclose their relevant patents, disclosure does not in any way guarantee a royalty-free standard.

    So, your argument that Google submitting VP8 to a standard body would create a better patent landscape is false. Such an action certainly has advantages, but this is not one of them.

    1. Rob Glidden Avatar
      Rob Glidden

      Tab:

      You are right that standards disclosure processes are not perfect, but they are much better than you imply, and in recent years they have been both upheld in litigation and tightened. The Broadcom v. Qualcomm case is well worth reviewing — Qualcomm got in huge trouble, including penalties, for trying to get around the JVT disclosure process.

      Generally disclosure processes allow both disclosure of the patents of others and of non-participants. In the case of non-participating patent holders, the standards group typically has an obligation to contact the identified patent holder and request they sign the disclosure form. You are right that the third party could refuse to sign (or simply declare the patents are only available on a royalty basis), in which case the standards group if so inclined would simply remove the patent. Even ANSI, which some see as a “pro-patent” process, has this kind of procedure.

      There is also the real possibility of a patent holder transferring patents to another company to evade disclosure requirements — whether someone could really get away with such deception is the subject of the VIZIO antitrust suit mentioned in the article. But that would obviously be a pretty high risk strategy on the part of a patent holder.

      Consider the distinctly unappealing alternative of doing nothing and letting existing patent pool administrators take their time and put together pools on their own time frame and under their own controlled, confidential essentiality determinations and negotiations (check the recent Nero lawsuit — patent pools are taking huge leeway in determining what patents to put in their pools and when to disclose what patents are in them). Assuming Google has done even some of the patent homework they claim, they should disclose and not leave the industry on its own on this.