Wednesday, August 17, 2011

Oracle v. Google update: summary judgment pressure and Motorola Java license fallacy

Over the last couple of days, a number of new documents have been filed in Oracle v. Google. I'll take a look at the most important ones of those. The two key issues here are a piece of evidence that could almost singlehandedly win the willful infringement part of the case for Oracle, and a set of four summary judgment motions filed by Google in hopes of having certain aspects of the case (including some absolutely critical ones) resolved in its favor by the judge even ahead of trial. The most important one of those proposed motions would have wide-ranging ramifications if granted in full. It would basically allow open source and other free-of-charge software to infringe a large percentage of all software-related patent claims.

Subsequently to discussing those new activities in the lawsuit, I'll comment on a fallacy spreading virally on the Internet concerning how Motorola's Java license might resolve Google's legal problems with Oracle after the closing of Google's proposed acquisition of Motorola Mobility. While this comes from a highly-respected veteran of information technology journalism, what he says flies in the face of everything I know about intellectual property licensing, and further below you'll see why I'm right and he's wrong on this particular one.

If you're only interested in certain parts of this post, which is basically three posts in one, here's a list of links that get you to different sections:

  1. Evidence controversy over Lindholm email

  2. Summary judgment pressure

  3. Motorola Mobility's Java license

Lindholm email controversy continues

A week and a half ago I reported on the continuing wrangling over the so-called Lindholm email, a document that the judge presiding over the case described (in different words) as extremely problematic for Google with a view to the trial.

That document shows that, shortly before Oracle's lawsuit, Google was well aware of the need to negotiate a Java-related IP license with Oracle. But Google's lawyers leave no stone unturned in order to prevent that email from being shown to the jury. Their legal argument is basically that even though it was, on the surface, a typical business email that begins with "Hi Andy" (addressing Android boss Andy Rubin), it was only an auto-saved draft of an email that was actually meant to be privileged attorney-client communication involving lawyers.

Google has already tried in a couple of other ways to have that document -- and quotes from that document at a recent hearing -- removed from the record. So far the judge doesn't agree with Google, but Google's lawyers keep pursuing their objective in different ways and from different angles.

As a result, a draft email of only a few lines has given rise to pleadings by both parties that already fill dozens of pages.

[Update] I have meanwhile done a subsequent post on a new Oracle filing related to the Lindholm email controversy. It's a very forceful document in which Oracle tells Google the following: "Damning facts are still facts." [/Update]

On Monday, Google's lawyers filed two new declarations on this matter that are interesting. While this seems to be a long shot, those declarations provide a very detailed account of the genesis of that controversial document and the circumstances surrounding it.

One of the declarations was made by Tim Lindholm (the author of the email at issue), and the other one by Benjamin Lee, a former Senior Counsel at Google and now (since November 2010) Legal Counsel at Twitter. As a Twitter employee, Lee shouldn't have a motivation to distort facts in Google's favor.

As we look at this, keep in mind that the question at issue here is only one: was it attorney-client correspondence or just regular business communication? If it's attorney-client stuff, Google has a case for this not being admitted as evidence. If it's just regular business correspondence, it must be shown to the jury. Now let's start with Lindholm's declaration.

Lindholm writes that "they [different versions of that document] constitute a communication to a Google attorney and were the result of work that [Lindholm] performed at the direction of Google attorneys". He says that in "late July 2010", Lee informed Lindholm that Lee "had attended a meeting in which Oracle made patent infringement claims against Google. Lindholm's understanding of Lee's information was that "Oracle was threatening to sue Google over those claims".

According to the declaration, Lee asked Lindholm "to gather certain information related to Orale's infringement claims", which Lindholm thought "would assist Google's legal analysis of Oracle's claims". On July 30, 2010, Lindholm "was asked by Google General Counsel Kent Walker to attend a meeting [at the Googleplex], convened by [Walker]" to discuss Oracle's infringement claims. The meeting was attended by Walker and Lee on behalf of Google's legal department as well as "Google top management and several Google engineers [the latter group including Lindholm]". Walker asked Lindholm "to continue to work under Mr. Lee's direction, and to work with Google engineer Dan Grove, also under Mr. Lee's direction, to gather information for Google's lawyers and management to consider in evaluating technology issues related to Oracle's infringement claims".

A week later, on August 6, 2011, at 11:05 AM, Lindholm "sent an email reporting to Mr. Lee and to Andy Rubin, who was then a Google Vice President in charge of Android, concerning certain topics that Mr. Walker and Mr. Lee had asked me and Mr. Grove to investigate". Lindholm says that at the end of preparing that email, but before sending it out, he added "Attorny Work Product" and "Google Confidential" at the top. In the To field of the email, he placed the addresses of Lee and Rubin. In the Cc field, he placed Grove's email and his own.

But before he did all of those things (which are key parts of Google's argument that this is attorney-client communication), the email was automatically saved several times. Lindholm refers to "eight different stages of preparation during the roughly four-minute period between 11:01 a.m. and 11:04 a.m.". Those auto-saves, Lindholm goes on to say, have empty To fields because he added the recipient's addresses only as "one of the last things that [he] did [before sending the email]". Another one of those last things was the addition of the potentially magic words "Attorney Work Product" and "Google Confidential".

The other declaration -- the one made by then-Google-now-Twitter in-house counsel Benjamin Lee -- basically tries to corroborate from that witness's perspective the substance of the Lindholm declaration. In particular, Lee, the lawyer, stresses that Lindholm, the engineer, was "supervised" by Lee during the relevant process, and acted at Lee's "direction".

The way I see it, the involvement of lawyers in this is credible, and the auto-save story, too -- though I'd have to see all of those documents to be sure.

However, there are still some strong indications that this was business correspondence, even if in a legally relevant context. It seems that the whole idea was to write an email to Rubin, and the addition of Lee as a To recipient was just an afterthought in order to declare something that is more of a business communication as attorney-client correspondence. The judge and Oracle's lawyers previously pointed out that the mere addition of the words "attorney work product" can't turn a business email into attorney-client correspondence.

The two declarations submitted by Google seek to emphasize the involvement of the company's legal department in this. If we were talking about an email in which, for example, Google engineers tell Google in-house lawyers about their specific research into legal issues, such as their analysis of specific patent infringement issues (claim charts etc.), Google would probably be right that such correspondence should enjoy certain protection. But an email that sums up such a general conclusion and a resulting business recommendation made by an engineer to a vice president is, in my view, a business email. That position is furthermore supported by the fact that the email itself talks about a request from Google founders Larry Page and Sergey Brin.

[Update] I have meanwhile done a subsequent post on a new Oracle filing related to the Lindholm email controversy. It's a very forceful document in which Oracle tells Google the following: "Damning facts are still facts." [/Update]

Summary judgment pressure: Google wants certain issues resolved prior to trial

Late on Tuesday, Google's lawyers filed motions for leave (permission) to file summary judgment motions on four issues. They can't file the actual motions right away without the judge's prior consent. Back in mid May, Oracle and Google agreed that "[o]ther summary judgment motions [than the now-ongoing process on copyright infringement] will be entertained only upon obtaining leave pursuant to the Court's previous order to that effect."

Judge Alsup, who presides over this case, has repeatedly stressed the court's limited judicial resources. On the one hand, this means the judge would rather avoid having to deal with too many summary judgment motions. Every "SJ" means he has to make a decision on a complicated issue, with a whole briefing process (a motion, an opposition brief, and a reply in support of the original motion). On the other hand, the judge also wants to streamline the case for trial so as to reduce the number of questions put before the jury, and SJ's can have that effect. Therefore, Google's motions for leave tout this potential benefit.

Three of the four SJ proposals are easily explained. I'll address them quickly before we get to the most important one, which relates to contributory (indirect) patent infringement.

  • Proposed SJ against assignor estoppel:

    Apparently Oracle argued that Google does not have the right to challenge the invalidity of three patents (the '520, '205, and '104 patents) that were assigned to Sun by four engineers who now work for Google but made the related inventions while they worked at Sun. Oracle argues that this is a case of asssignor estoppel. The definition of assignor estoppel cited by Google (from a different case) is that of "an equitable doctrine that prevents one who has assigned the rights to a patent (or patent application) [usually the inventor] from later contending that what was assigned is a nullity".

    If this particular SJ motion is allowed and granted, Google can challenge the validity of those patents. If not, Google can only argue against those patents on other grounds than invalidity, such as saying that they aren't actually infringed or that there are reasons for which Oracle is not allowed to enforce them against Google.

    On this one I tend to side with Google, not based on a detailed analysis of applicable case law but simply because I want the truth to come out. If there are any reasons for which those patents are invalid, they should be considered by the court regardless of the company those engineers work for now. I'm against bad patents regardless of who owns them. Patents are a public thing anyway -- substantive invalidity arguments shouldn't be considered confidential.

  • Patent marking:

    Google argues that "Oracle may not recover damages for the period pre-dating notice to Google of alleged infringement of the patents-in-suit" because it failed to expressly mark products "with the patent number of any of the Patents-in-Suit or Related Patents".

    The whole question of patent marking obligations has also created an opportunity for trolls who shake down companies who mark products with patents that have expired. U.S. patent reform will address that nuisance.

    If this SJ motion is accepted and granted, Oracle's entitlement to past damages would be reduced substantially. As I pointed out on previous occasions, the big commercial issue here isn't even going to be damages. Sure, Oracle tries to get billions of dollars on that basis, but the judge didn't buy too many of Oracle's aggressive damages theories (about which I had also expressed doubt months before). But the really relevant commercial issue here is the question of an injunction. If Oracle had an injunction in its hand, or if Google tried to negotiate settlement terms while Oracle is likely to obtain an injunction, then the prospect of a shutdown of the platform for about 200,000 Android applications represents such enormous leverage that the question of past damages becomes only a secondary theater of war.

  • Invalidation of '104 patent:

    Google wants the judge to throw out Oracle's '104 patent, alleging that Oracle "impermissibly broadened the patent's scope beyond the original patent in an application filed after the two year statutory limit on broadening reissues".

    The '104 patent is a reissue patent of U.S. Patent No. 5,367,685. The original one was issued in November 1994. Sun succeeded in obtaining a broadened reissue once (based on an application Sun filed for that purpose in 1996). But Sun wanted even more, and the '104 patent apparently results from a "continuation" reissue application filed in March 1999 -- four and a half years after the original patent was issued, and Google says the limit is two years.

    If the judge accepts this motion, agrees with Google's interpretation that the '104 reissue patent indeed constitutes a broadening of the original one, and furthermore agrees with Google that Sun applied for that broadening after a statute of limitation, then Oracle would lose, according to Google's count, 11 patent claims of the 41 that were at issue in the case as per August 8, 2011 (down from a shortlist of 50).

    I don't like overly broad patents and, like on the previous invalidation issue, I'd like to see Google succeed with this one, too -- again, not based on a legal opinion on the proposed motion but on my opinion that patents shouldn't be broadened again and again, and that bad patents generally shouldn't be asserted at trial if they can be thrown out before.

And now we finally get to the fourth and most important SJ motion proposed by Google: contributory patent infringement.

If Google succeeded with all of the three previously mentioned SJ motions, Oracle would still have pretty good chances at trial. But the infringement SJ could have devastating impact on Oracle's case. That's why it deserves a closer look.

Google would like the judge to determine that there's no infringement of any of the patents-in-suit under the particular theories set forth in two paragraphs of the patent infringement article in U.S. law. The infringement article is 35 U.S.C. 271, and the two paragraphs to which this SJ proposal relates are (c) and (f). (c) is about contributory infringement in general, and (f) about contributory infringement with a view to foreign markets. Easy to memorize, actually: c = contributor, f = foreign.

Contributory infringement is an alternative infringement theory to direct infringement. Oracle also accuses Google of infringing its patents directly, and the proposed SJ motion doesn't relate to that (otherwise this would be an SJ that could do away with the whole patent part of the case).

Direct infringement is the case if Google's Android software itself practices the inventions claimed by Oracle's patents (to the extent those patent claims are valid, of course). Contributory (indirect) infringement is the case if what the patent describes is infringed by those who take Google's Android software and distribute or use it.

It depends on the language of a given patent whether software can infringe it directly or only contributorily.If a patent claim covers a "method", it depends on what the technical prerequisites of that method are. If it's all just software, there can be a case for direct infringement. But if the language of that patent claim also relates to hardware components, then one may have to sell the hardware and the software to infringe directly. The need for the hardware is even clearer if a claim is not on a "method" but on an "apparatus".

Practically speaking, a patent holder is entitled to the same material benefits whether an infringement is direct or contributory. But if a patent cannot be infringed by software all by itself, direct infringement is a no-go, making contributory infringement the only basis on which the patent holder can possibly prevail.

Google's proposal first discusses (f), then (c).

On (f), Google argues that the Supreme Court of the United States already decided in Microsoft Corp. v. AT&T Corp. (see Wikipedia for a description of that case) that software that is "uncoupled from a medium" cannot give rise to contributory infringement claims under paragraph (f). Otherwise, U.S. patent law could affect foreign hardware companies who ship U.S. software on their devices even though those patents aren't enforceable outside the U.S.

I support Google against Oracle's claims with respect to foreign markets under paragraph (f). In terms of business impact, this won't affect the most important question, which is whether Oracle can obtain an injunction against Google in the U.S. market. Whatever happens or doesn't happen outside the U.S. would be a secondary consideration. If Oracle can shut down Android in the U.S. market, I can't imagine that Google will decide to take an unbelievably hurtful hit in that market just because certain ways to use those patents outside the U.S. are excluded from the U.S. decision. This again also relates to damages, but like I said, whatever the court may award in damages would be dwarfed by whatever Oracle can dictate if it has obtained, or is reasonably likely to obtain, a U.S. injunction against Google.

After arguing against infringement under (f), Google then also tries to get rid of all allegations related to (c). That part would have wide-ranging consequences. If Google succeeded on this one, not only would Oracle's chances of prevailing in this lawsuit be substantially reduced but this would also have wide-ranging effects on open source and other free-of-charge software, which could all of a sudden infringe with impunity a fairly high percentage of all patent claims that relate to software.

While this is intriguing in a way, I absolutely can't imagine that the court will agree with Google.

Google presents two arguments for this radical proposal:

  1. Google says "the reasoning of Microsoft [v. AT&T] applies equally to section 271(c), which applies only to 'a component of a patented machine, manufacture, combination, or composition, or a material or apparatus for use in practicing a patented process....'" In Google's view, "Android source code unconnected from a medium cannot be a component because it is intangible", and even more so, it can't be a material or apparatus, which are "tangible things".

    Google claims that those terms (component, material, apparatus) must be interpreted the same way under (c) and (f). However, Google doesn't seem to cite any legal authority (other cases etc.) that is applicable to the situation in Oracle v. Google. For now, I'm unconvinced.

  2. Then Google argues that Android can infringe patents contributorily with impunity because it's available free of charge:

    "[S]ection 271(c) applies only to offers for sale, sales, and importation, but Google does not sell, offer to sell, or import Android or Android SDK."

    Google says that "offer to sell" and "sell" should have their ordinary meaning:

    "The ordinary meaning of 'sale' requires transfer of ownership for a price. [...] Android and Android SDK are offered free of charge to anyone who wishes to download them off the Internet, however, and there is no evidence that Google ever receives any consideration in return. That undisputed fact precludes a finding of contributory infringement."

    So this would put all free-as-in-beer software above patent law with respect to contributory infringement. There are many patent claims that software can only infringe contributorily, so this finding would do away with a pretty substantial part of the patent problem open source and other free-of-charge software faces.

    This argument is similar to the one that Google made in connection with damages. Basically Google's argument was that since it doesn't charge a price for Android, it doesn't owe damages. When Oracle reminded the judge of that take, Judge Alsup said that "zero [damages] is ridiculous" and not going to happen. I believe he'll view this zero contributory infringement idea in a similar light and will simply have to come up with a legal reasoning to reject something so utterly absurd.

    The judge (and Oracle's lawyers, if this motion is allowed in the first place) will be in a much better position than me to dismiss that "free" argument. Everyone knows that there is an Android business model in place. It's a new kind of business model, and by the time most of the body of law that exists today was drafted by lawmakers, such business models weren't taken into consideration because they weren't nearly as relevant then. In fact, some of those business model work only in the times in which we live, i.e., in the digital era. But the judicial system is pretty good at figuring out ways to interpret such statutes appropriately in light of the evolution of technology and business models.

    Google itself actually indicates a way in which its theory can be rejected: it refers to the "consideration" it receives in exchange for the license. Google does make money with Android, mostly through online advertising, and its OEM license agreements are all about locking users into Google's services in order to generate such revenues. There are probably ways in which judges and lawyers can identify a "consideration" and/or argue that infringements simply cannot be legalized through free-of-charge distribution. In a way, infringements by software available free of charge are even a worse problem for the patent holder than infringements by, say, packaged software: something that's free may be installed on an infinite number of devices.

Motorola Mobility's Java license cannot make "Oracle lose grounds for its lawsuit"

Robert X. Cringely is a well-respected information technology journalist. I read his column in Infoworld in the 1990's, and I sometimes find interesting thoughts and rumors (not necessarily reliable, but interesting) on his blog. Yesterday he blogged about Google's proposed acquisition of Motorola Mobility and said something that doesn't make sense from an IP licensing perspective:

"If this Motorola Mobility deal goes through (and I think it will) then Oracle loses grounds for its lawsuit, which is part of why Google is even doing the deal."

Maybe Cringely believed a Google insider too easily. I suspect this because he affirms that this is "why Google is even doing the deal", which suggests he believes to have knowledge of Google's strategy. Google would have everyone believe that the MMI deal is about patent protection when it's actually about a business strategy that will give regulators pause.

If you're looking for a deal that Google recently did that could somewhat realistically affect Google's settlement negotiations with Oracle, or failing those negotiations result in a countersuit, Google's recent purchase of over 1,000 IBM patents is a far more interesting one. But not MMI.

I've been negotiating and analyzing license agreements involving different types of intellectual property since the late 1980's. There's no realistic way -- not even a remotely realistic one -- that Sun (which apparently did a license deal with Motorola long before being acquired by Oracle) would have granted Motorola under a Java license

  • the right to do anything it wants with Sun's Java-related intellectual property, including what Android's Dalvik does, a virtual machine that doesn't comply with the Java specifications, and

  • the right to grant the same right to third parties and/or future parent companies.

Totally theoretically speaking, a right holder could grant such open-ended, limitless rights. But if you as a right holder do that, you have to assume that one such deal may be the only business opportunity you ever have because that one licensee could then provide everyone else in the world, for all purposes imaginable, with a license.

In commercial practice, which I know from the dealmaking side and as a ligitation watcher, license deals with companies always come with field-of-use restrictions. Such limitations are usually expressed very clearly in those license agreements.

A license to implement Java for certain purposes is not a license to do what Android's Dalvik virtual machine does. It could be, but that's highly unlikely, and even if that were (contrary to conventional wisdom and Oracle's consistent representations to the court) the case, a license that a device maker like Motorola received won't just cover third-party device makers shipping Android.

The limitations of any realistic benefit to Google from MMI's Java license are such that a Java implementation on MMI's own Android devices would be safe if (unlike Dalvik) it fully adhered to the Java standard. But it's not like MMI could take a license to ship Java on certain devices and suddenly enable an acquirer like Google to ship Android to third parties or publish any such code on open source terms.

The limited benefit I just described is actually nothing new. A long time ago, Oracle made a promise to license (on FRAND terms) its Java-related intellectual property to those who implement Java according to its specifications. So if Google wanted to comply, it could get a FRAND license anytime (and in limited fields of use even a free-of-charge one).

In light of the above, there's no plausible theory under which Google's acquisition of MMI would do away with even the smallest part of Oracle's lawsuit against Google. Even though Sun wasn't always managed too well at the highest level, Sun wasn't that stupid.

If you'd like to be updated on the smartphone patent disputes and other intellectual property matters I cover, please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents and Google+.

Share with other professionals via LinkedIn: