Google: No 'shred of evidence' Android team saw Sun's patents

Google: No 'shred of evidence' Android team saw Sun's patents

Summary: Google followed up with its closing arguments in phase two of the trial, reiterating that it had no knowledge of Sun's patent portfolio.

TOPICS: Oracle, Google

SAN FRANCISCO -- Google's counsel followed up with its closing statements in phase two of its legal battle against Oracle at the U.S. District Court of Northern California on Tuesday morning.

See alsoOracle argues 'Google's reckless path to patent infringement'

Attorney Robert Van Nest delivered closing arguments on behalf of the Mountain View, Calif.-based company, starting off by reminding the jury that "this case is not about Java versus Android."

With that in mind, Van Nest outlined Google's three core points in its case:

  • Google made "fundamentally different design choices for Android" without any specific knowledge of Sun Microsystems's patent portfolio
  • Android's Dalvik Virtual Machine does not infringe U.S. Patent No. RE38,104
  • Android's dx tool does not infringe U.S. Patent No. 6,061,520

"There isn't a shred of evidence that anyone at Google on the design team had seen these patents," Van Nest said, asserting that Oracle "admitted" that it didn't mention these specific patents to anyone at Google until the lawsuit was filed by Oracle back in July 2010.

Van Nest asserted that the '104 patent requires that symbolic references must be used in the instructions, backing up that "Android never uses symbolic references in the instructions, but rather numeric references."

Incidentally, one of the big debates that has emerged between experts testifying for both sides in phase two has been terminology, which could confuse and even sway the jury one way or another.

For example, on the '520 patent, Van Nest said that every expert witness acknowledged every step of the method has to be there, including simulation of the bytecode. He explained that Android doesn't implement simulated execution like Oracle argues, but rather "pattern matching."

In his closing arguments earlier in the morning, Oracle attorney Michael Jacobs defined a symbolic reference as “a reference that identifies data by a name other than the numeric memory of the location of data, and that is resolved dynamically rather than statically.”

Citing the testimony of Dr. David August, an associate professor of computer science at Princeton University who testified as an expert witness on behalf of Google, Van Nest recalled that August confirmed he looked at every implementation of the instructions, testifying "with certainty" that there aren't anything but numeric references in Dalvik's instructions.

Van Nest also brought up the testimony of Dr. John Mitchell, a professor of computer science at Stanford University and an expert witness for Oracle. Earlier, Jacobs has tried to defend Mitchell for making a "mistake" in his deposition video by testifying that indexes are numeric references, which would really favor Google's case and not Oracle's.

While Jacobs aimed to assure the jury that Mitchell corrected himself later, Van Nest countered by saying that it wasn't a mistake, and he has simply since "changed his opinion."

"I wouldn't criticize someone for making a single mistake, but these are not mistakes," Van Nest reiterated.


Topics: Oracle, Google

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • It's over folks, Oracle lost!

    Google has launch the torpedo of truth and the Oracle's rotten ship is sinking! Elli$on can't patch it with lies anymore!
    The verdict will find in favor of Google and there will be no phase 3 since there is no liability.
    Once again FOSS wins, adding Elli$on to the same league of sore losers like $CO!
    The Linux Geek
    • Once again, FOSS wins?

      I had no idea that the GPL was under assault.
      Your Non Advocate
      • it was Dalvik this time

        Apache not GPL!
        The Linux Geek
      • RE: it was Dalvik this time

        The Linux Geek wrote:
        [i]Apache not GPL![/i]

        Apache's Harmony was Apache 2.0. Google dipped into Harmony for Dalvik. Sun GPL'd [most of] Java nearly at the same time that Google released Dalvik.

        P.S. Android (minus the Linux kernel) is Apache 2.0 as well. However, while Google's services are built with open-source software, the software comprising the services is mostly proprietary. Just try to download Google Docs to serve on your local network. Or try to download Google+ to set up private social networking amongst you and your family and friends. And how many Android apps in Google Play (formerly Android Market) are open-source?
        Rabid Howler Monkey
    • professional opinion?

      professional opinion or what exactly is it?
    • These are closing aruments

      NOT a jury decision or decision by the judge... seriously go check your shorts.
  • No shred of evidence Android team saw Sun's patents?

    Of course there isn't - well, at Google HQ anyhow.

    Because Google just shreded all the documents that proved they saw Sun's patents. ;)
    William Farrel
    • Any more...

      ...of your proof-less phantasies we should hear about?
      • Are you a Phillies fan?

        as most spell it as [i]fantasies[/i], with an F as opposed to a PH. And no, we don't live in Filladelphia, we live in Philadelphia, so we're allowed to use PH in place of F.

        Oh, and if you couldn't take the post as a play on words...
        William Farrel
    • what engineer would waste time searching poorly worded descriptions

      the more expansive the claim the more ambiguous the wording
    • High priced talent is busily counting

      How many Google engineers are reading patent 6,000,001 6,000,002 etc?
  • Doesnt matter if they saw them or not. You cant use patented ip whether you

    know about it or not. Not that anyone would believe a word out of googles mouth to begin with but still.
    Johnny Vegas
    • Certainly can affect likelihood of infringement

      A prudent juror might well conclude it is less likely that Google infringed if they came up with their own solutions on their own and as opposed, for example, copying or even disassemblying the Sun/Oracle code.

      I have to say that Dr. Mitchell's testimony seemed to tear up everything I ever read about the meaning of symbolic referencing of data. Seems like he purposely conflated "symbolic reference" and "indirect reference", just so Oracle can make some questionable claims.

      According to him, the code fragment "a(ix)" out of a 30 year old Fortran program, having been translated into MOV R1, X(R2) is a "symbolic" reference. Gonna have to shred some text books for that to be true.

      To be fair to the good Dr. Mitchell, patents define their own terms. Sometimes in non-sensical ways.
      • Nonsense dominates software patents. No models to demo function.

        If the could demo the likelihood that it would work as designed, maybe I'd believe the validity of Oracle's symbolic vaporware.
  • pathetic excuse

    "Google made ???fundamentally different design choices for Android??? without any specific knowledge of Sun Microsystems???s patent portfolio"

    - Lack of knowledge is not an excuse. If you are in big business its a must to know competitors portfolio. It's just shows google's utter disrespect for patents.

    And the hypocrites at google didn't forget to patent 'project glass'
    • does not apply to open software

      only closed software companies must proactively scour the patents to check for infringement.
      The Linux Geek
      • Warning

        If anyone here is in need of legal advice, call 1-800-ASK-DOOFUS. Do not under any circumstances rely on The Linux Geek for such advice.
        Robert Hahn
    • Umm

      neither is hate!
  • Very little has been said about Google's clean room used for Dalvik

    Google has many ex-sun Java engineers on their payroll. Many of these ex-Sun Java engineers worked with Sun's legal team to draft patent applications that were sent to the USPTO to build up Sun's Java patent portfolio.

    Has a list of Google employees that comprised the Dalvik clean room team been released? Were any of these employees ex-Sun Java engineers? If so, what patent applications did these ex-Sun Java engineers contribute to while at Sun?
    Rabid Howler Monkey
    • the room was very clean

      there is no evidence that patents or java source entered that clean room.
      Oracle has no proof!
      The Linux Geek