On Patent Infringement Trials And Their Jurors

Newegg lost their lawsuit with TQP Development Erich Spangenberg before the Thanksgiving break.  How does such a bad decision come out of such a pretty courthouse, is what I want to know!


from Joe Mullins’ Ars Technica post. Did you take that photo? Lovely!

And you know what?  I do know.  I know exactly how this stuff happens, and because I’m cool like that I’ll go ahead and share the love.

Follow me along the trail here, if you will:

  • The people who file patents, by and large, have law degrees.
  • The people who issue patents, examiners at the USPTO, have engineering degrees.
  • The people who send out demand letters threatening an infringement suit are lawyers or self-described thugs.
  • The people who argue patent infringement cases have law degrees.
  • Yet, inexplicably, the people who decide patent infringement cases are…butchers?  Bakers? Candlestick makers?

Does anyone else see a problem here?

This is why, as I wrote about well over a year ago, so many companies settle with patent trolls.  Not only do the not have the money to fight a lawsuit, they don’t want to take their chances with a jury if they do.  You could end up with Velvin Hogan as your foreman, for heaven’s sake!  This is what I said then, and it’s apropos now:

Then there is the problem of putting very technical arguments in front of the general public.  That’s not a slam on the general public, for I are one of them.   Patent infringement trials are fraught with all manner of industry-speak and jargon and terms that people have to look up in order to understand.  Or worse, they need the lawyers to explain it them and we all know how that is likely to end up.  (Hint:  lawyers are terribly partisan explainers, in that they explain only the part of the definition they want you to know, the part that will tip the verdict in their favor.)  Unless you just enjoy spending your time reading about the ins and outs of your newest gadget, all that stuff is going to fly over your head.  And if you buy into the rhetoric that corporations are E.V.I.L. and don’t deserve to make money, then you’re almost always going in with the attitude that Deep Pockets is wrong and the Patent Troll is right.  It’s an easy assumption that is difficult to overcome no matter how good your lawyer is.

It’s probably not fair for me to blame this verdict on the jury, when the blame squarely belongs on poor patents and companies that abuse them like IP Nav.  But good grief, Charlie Brown.  This is not a situation where a “jury of your peers” applies.  If I’ve been mugged or my neighbor’s septic tank has overflowed into my backyard and they refused to pay for the resulting damage (not that that happened to my family as a child and has forever scarred me and now I can’t live in homes where there’s a septic tank) or if I were to spill hot coffee on myself and try to get money out of McDonalds then things would be different.  In those cases?  I need my peers.  People like me who live similar lives, and who do and experience similar things that I do and experience.


Patent infringement is so not a mainstream “thing” that any of my peers get.  You want to know how I know this?  Because when I tell people that I write about patents and patent trolls and lawsuits and such their eyes glaze over they respond with a head-tilt and a very polite “Well.  That’s…interesting.” And then they nod off to sleep and their heads bob forward and slam onto the table at the little cafe where we’re having lunch, the cafe that I will never get invited back to because I talk about boring things like patent trials.  That?  That’s how I know.

I don’t have a solution for who should determine verdicts in patent infringement cases.  I know that outside of the patent troll issue there is certainly plenty of legitimate disagreement over patents and those disagreements need to be heard and vetted by a group of someones.  I’m just not sure it should be a group of someones who live in Marshall, Texas.

Which, by the way, does anyone track juror service up there?  I mean, there’s not but 67,000 people in the whole darn county.  My freshman English class at UT had that many,  (OK, not really.  It had 350 which is pretty much exactly the same.)  and probably only about 1/4 of those are even eligible for service.  How do we know that the same people aren’t being dragged into the courtroom every two weeks for another stint?

We don’t, but it doesn’t matter.  They would still have made the wrong decision in this case.

I’m glad Newegg’s going to appeal and I hope that they win and it’s not just because that’ll mean a loss for IP Nav/Erich Spangenberg/TQP Development.

Actually?  It’s totally because of that.



{Charlie Brown image found here. I’ll cop to the (super simple) photoshopping.}

3 thoughts on “On Patent Infringement Trials And Their Jurors

  1. When the trial transcripts are published and accessible via pacer.org for this case, you should take a look at them. I purchased them for the most relevant days and have researched this case extensively and I disagree with your assertion that the jury is at fault here. The fault in this case lies first with the Plaintiff whose infringement theory was baseless and inappropriate (though the Court obviously didn’t see it that way), second with the District Courts (twice) for an improper claim construction and avoidance of Summary Judgment, and third with the Defendant whose attorneys used risky and confusing arguments at trial.

    Let’s start with the perspective of the jury and how they could find the patents to be valid and the Defendant to be infringing.

    • Jury instructions usually make clear that patents are presumed to be valid and that it takes clear and convincing evidence to prove otherwise.

    • The 5,412,730 patent went through a USPTO re-examination requested by TD Ameritrade after reviewing “150 additional prior art documents not considered during the original patent examination” (http://gametimeip.com/2011/10/06/the-pa … onary-tale) and confirmed patentability of all original claims and added several new claims.

    • The Defendant’s arguments were not about non-infringement, but were about patent invalidity. One of the invalidity arguments was that the RC4 encryption algorithm was publicly disclosed and offered for sale prior to the “one year before patent priority date”. With this argument, the Defendant in effect admitted infringement since they admitted use of RC4 so if the Jury didn’t believe the prior disclosure argument, then they would likely return the verdict that they did (i.e. infringement). The public disclosure was via Lotus Notes distributed to application partners via a development and license agreement and though not under NDA it was not published or sold (and “commercialization” is a subtle topic for a jury to grasp). The offer for sale was said to be at a trade show and was also not clear cut since any actual sale with its collection of payment and delivery occurred after the prior art cutoff date.

    • The improper claim construction from the Court broadened the claims to become obvious under prior art, but the Defendant did not argue obviousness but rather anticipation which is a more difficult standard where a piece of prior art must meet every element of the claim. Also, the improper claim construction put the Court at odds with the USPTO and their re-examination so would confuse a jury to be cautious with invalidity.

    • The improper claim construction from the Court has the patent claims seemingly apply to RC4 with SSL (or TLS).

    • The Defendant put forth arguments about public key cryptography (with Withfield Diffie on the stand) that were irrelevant to this case except that, were it not for the Court’s improper claim construction, they rendered the techniques described in the patent moot. To the jury, this sounded like an irrelevant distraction.

    To understand how the Court erred, you need to understand a little bit about cryptography. There are two broad types of cryptographic techniques. The first technique is used with block ciphers which take a block of plaintext and encrypt it as a whole into cipher text. This is done through a combination of diffusion which shuffles the bits in the block and confusion which randomizes (substitutes) those bits. A key is used to define the unique way this diffusion and confusion occurs and the process is reversible so that decryption converts the ciphertext back into the original plaintext. The distinguishing characteristic for the block cipher is that the message itself is redistributed within the block (i.e. the diffusion aspect), not just modified in place. The key is what must be transmitted securely or otherwise exchanged in advance to make this work.

    The second technique is used with stream ciphers where pseudo-random numbers are generated and a simple combining algorithm converts plaintext to ciphertext and vice versa. The most common combining algorithm uses XOR which is a bitwise operation that is reversible such that applying it twice gets you back where you started. The sequence of pseudo-random numbers is sometimes called a key stream though when using XOR it is really a bitstream and in any event the key stream is NEVER transmitted. The starting point for the pseudo-random number generator (aka the key generator) is usually called a seed. For stream ciphers, it is this seed that must be transmitted securely or otherwise exchanged in advance to make this work.

    The patent recognized that using the same key for every block in a block cipher was less secure and that it would be better if the key changed after a pre-determined number of blocks were transmitted. The novelty of the patent is that it proposes using synchronized pseudo-random number generators at both ends to generate the sequence of keys. You may immediately recognize that this sounds very much like how stream ciphers work and in fact the USPTO objected to the patent application because stream ciphers had long predated the patent. To overcome this objection, the applicant added language to the patent claims regarding the “data comprising a sequence of blocks” and that the key values changed “each time a predetermined number of said blocks are transmitted over said link.”

    While the patent claim language might make one limit the patent to block ciphers, one could still reasonably apply it to stream ciphers by interpreting “key” is being the seed to the stream cipher since that is analogous — it is what needs to be transmitted or exchanged initially to start the process and it is also what the patent would want to regenerate (i.e. a new key for a block cipher or a new seed for a stream cipher) for improved security. The thing is, no modern cryptographic protocol uses this technique for changing the key or seed. Instead, the protocol is restarted using public-key cryptography to securely transmit the new key or seed.

    The Plaintiff’s infringement theory was that RC4 (a stream cipher) in conjunction with SSL (or TLS) infringed the patent claims. They tied the patent claim element language to the RC4 pseudo-random number generator and key stream that was generated and applied one byte (8 bits) at a time. Because the Court had defined a block as being more than 1 bit, the byte-wise encryption used in RC4 was claimed to be infringing (along with the transmission protocol provided by SSL, but there is nothing unique there). You can see how this is nuts. Their argument is the same one the USPTO was hoping to prevent with the “block” language.

    The Court should have looked at the patent prosecution history and defined “block” as more than one NON-SEPARABLE bit, meaning that the generation of a one byte (8-bit) key in a key stream and the use of it in XOR encryption and decryption is a processing efficiency and doesn’t change the fact that the keys themselves are still 1-bit and are applied bitwise. The Plaintiff’s own expert witness on infringement admitted as much in describing XOR that combines the key and plain or ciphertext together “effectively bit by bit”.

    Furthermore, the Court should have looked at the “principle feature” in the patent description where it describes the pseudo-random number generators creating new keys “without these keys being transmitted in any form over the transmission facility”. The prior-art stream ciphers (as well as RC4) already met this requirement so the Court should have defined “key” as the information that needs to be transmitted or exchanged initially (and for the patent is regenerated every so often) and specifically stated that it is NOT any part of the key stream of a stream cipher.

    Had the Court done a proper claim construction, then it would have been obvious to declare Summary Judgment in favor of the Defendant (though the East District of Texas has one of the lowest rates of Summary Judgment). Or course, even if that was done, the $45+ million taken from over 140 other companies would not have to be returned. The Plaintiff should never have asserted this patent in the first place as it was inappropriate and baseless (IMHO).

  2. I want to add some additional context to the flaws in this case because there were three related cases with claim constructions for this patent. The first case was TQP v. Merrill Lynch:

    Click to access 0.pdf

    What is most notable in this first claim construction is the lack of a precise definition of a “key”, especially in the context of the two primary cryptographic techniques of block ciphers and stream ciphers. Though it is noted that the data is encrypted and decrypted using encryption keys, there is no discussion about the distinction between the bit keys in the key stream of a stream cipher vs. the seed used for its pseudo-random number generator as the former is never transmitted or exchanged while the latter has to be (at least once). In fact, there is no discussion of such ciphers at all. The second case was TQP v. Barclays:

    Click to access 0.pdf

    In this second claim construction, the topic of block vs. stream ciphers is discussed. Unfortunately, the Defendant used an extreme argument to try to completely exclude stream ciphers from the patent claims, but as I noted in the previous comment this is not necessary if one allows for the patent claim word “key” to apply to the seed of the pseudo-random number generator of a stream cipher. The Court did note how stream ciphers are encoded bit-by-bit while block ciphers are not, but did not then notice that their definition of a “block” requiring more than 1 bit would mean that the patent claims could not apply to a bitstream since a block is more than 1 bit and with 1-bit keys this means that more than one key must be generated per block. The fact that these 1-bit keys are generated in parallel or as a group is irrelevant and there is nothing in the patent about efficiency (except in compressing the data to have less to transmit). The third case was TQP v. 1-800-Flowers.com which was the case against Newegg:

    Click to access 0.pdf

    This claim construction basically rehashed the prior ones. Again, there was no discussion of excluding “key” from the key stream of a stream cipher (especially when it’s a bitstream) since, by definition, it is never transmitted and instead applying “key” to the “seed” or other initial information needed to be exchanged (at least once). I suppose the idea of having one pseudo-random number generator that generates (after a certain amount of data has been transmitted) a new key (seed) to essentially reset another pseudo-random number generator of the core stream cipher was beyond the comprehension of everyone in these cases. This is the only logical rational way consistent with the invention that one can apply the patent claims to stream ciphers (including RC4).

    I should note that I simplified the stream vs. block discussion since in practice many ciphers are hybrids. Nevertheless, any algorithm using XOR between the generated pseudo-random numbers (bitstream no matter how efficiently generated) and the plaintext/ciphertext is a pure synchronous stream cipher and that includes RC4. The following book, “Cryptography and Data Security” by Denning was a reference to prior art in the Newegg trial and is an excellent reference (albeit older) on cryptography:

    Click to access Denning-CryptographyDataSecurity.pdf

  3. Pingback: The Best Post on Patent Reform (Not Written By Me, I Mean) | IP Troll Tracker

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s