You'll be happy to know that mirror_slap is in the courtroom for Groklaw
today, and so we'll be getting all the details. Meanwhile, according to Rachel King of CNET the jury has asked
for a readback of the answer from the judge to an earlier question. I hope it isn't the one I thought he was unclear on myself, and I know the answer, but he now
has told the jury that if they are still not clear, they can ask another question. I think they drew the impression that they can't consider Dr. Terence Parr's testimony as evidence that Google doesn't infringe Oracle's patent, and of course that is precisely what his evidence was for. If the jury misunderstands and thinks it can only consider Dr. Mitchell's testimony for Oracle as being somehow true, which some of the tweeting journalists thought had been told to them, they are going to have been seriously misled.
And we have our first report from the courtroom. And now I see that the question was actually two questions, one possibly the one that worries me or the one from Friday about a unanimous verdict, and the other is about definitions regarding one of the patents.
Jump To Comments
Update 2, Update 3
Here's report #1:
Update 4, Update 5, Update 6]
Oracle v. Google, Monday May 21, 2012
One thing is very clear. This jury, or at least someone on the jury, is sincerely trying to understand.
Arrive 7:50am Mr. Van Nest and Baber and 2 associates are here for Google. Mr. Jacobs is here for Oracle, along with a retinue of 4 other lawyers.
Dawn arrives (the clerk), and greetings are exchanged. Mr. Van Nest asks if everyone is accounted for, and Dawn says that they are.
8:05... everyone for Oracle splits, Dawn says that if anything happens, she knows how to reach them. Mr. Baber and one of his associates remain at the Google table.
9:20... activity. Oracle's lawyers come back in. 7+ 2 associates. Google's lawyers enter... same numbers. Discussion with Cathy, the court reporter, who handed back some documents. Some confusion ensues from Oracle about why they are being given the documents, when there was a question about them. About a question that was answered around noon (Friday?) about (transcript page) 4262. Mr. Van Nest and Mr. Jacobs are talking with her. Van Nest: starting at page 4263, line 3.
9:35 Judge Alsup enters... Asks whether everyone enjoyed watching the eclipse, said that he did. There is a note from the jury regarding question number 2 on 104 patent. Have you agreed?
Google: page 4263 line 3, through [the point where they were told to go deliberate].
Jury has a new note, and it refers to old note... what the instructions refer to in the 104 patent.
Court reporter reads back: What is meant by instructions containing one or more symbolic references? [Missed most of the readback of Judge Alsup's previous instructions ]
[Sounds to me like arguing what the definition of the meaning of the word "is" is, but that's just me. The jury is apparently really trying hard to understand what symbolic references are, or hopefully in this case, what they are not.]
Jury, when asked by Judge Alsup, indicates that their question has been answered, and they go back to deliberating.
Judge Alsup: We will now wait for the next question. [Leaves the courtroom, as does all of Team Oracle and all but three of Google's lawyers.]
Me too. It turns out we have a second volunteer there today, who just at the last minute decided to join in the fun. And his information is that the question was about the second question the jury sent to the judge. And King now tweets that there seems to be a third question, "Can we determine a numeric reference instruction to become symbolic because of what happens downstream?" Oracle says yes, she says. Of course it would.
The jury instructions for the patent phase of the trial and the verdict form are here. This is what the jury is struggling with. The definitions in the instructions include this section on patent '104:
I am now going to instruct you on the meaning of some of the words and phrases in the claims of the ’104 patent. You must accept and use these meanings in your deliberations.
Intermediate form code and intermediate form object code: Both “intermediate form code” and “intermediate form object code” mean “executable code that is generated by compiling source code and is independent of any computer instruction set.”
Symbolic reference: The term “symbolic reference” means “a reference that identifies data by a name other than the numeric memory location of the data, and that is resolved dynamically rather than statically.”
Resolve and Resolving: The term “resolving” a symbolic reference means “at least determining the numerical memory-location reference that corresponds to the symbolic reference.”
You must accept these definitions as established for purposes of your deliberations and verdict. You may, however, consider all of the evidence in the case as to whether or not the accused product or method meets these definitions. If a witness has based his view on meanings of the terms contrary to my stated definitions, you should discount that part of his testimony accordingly.
Struggling is the right word, as King says they're now asking, "Regarding symbolic reference in '104 claims, does the resolution of symbolic references need to happen immediately?" Google says the answer to that should be yes.
The jury questions indicate that once again the jury may be divided, some pushing for Oracle, some for Google. The part they don't know is that the USPTO already has preliminarily decided that the '104 patent is invalid. If that in the end is the final finding, then nothing the jury does will matter. That's why Google asked that the trial not happen until the reexaminations of the patents were final. It looks like Google was right. It would have been a better way.
Another question from the jury, via King, this time about '520 patent: "Jury also asks on determining claim language on '520, is stack synonymous with memory?"
The judge has answered the questions, and after the jury left, both parties objected to his answers. That's for the appeal.
Update: Here's our second report from mirror_slap:
Oracle v. Google, Monday May 21, part 2
If the judge means that this was a stupid patent case from the get go, I'd have to agree.
10:20... activity... lawyers filing into the courtroom.
10:35 Dawn shows both sides a document. Mr. Jacobs shakes his head and smiles. Apparently the document is 3 questions from the jury that the judge wants addressed by counsel. Mr. Van Nest says to Dawn that Google is ready on the first two, and asks for some time to work on the third.
Dawn comes back and provides copies for everyone on both sides.
10:45 Judge Alsup enters... "have you seen these three notes?
Regarding symbolic reference issue, does the resolution of the symbolic references need to happen immediately, or can they contain linking of numeric references and be resolved later?
Oracle: We believe that the jury is asking questions that should be addressed in claim scope.
10:10 question should be yes [ referring to the questions by their timestamp. 10:10 is question 1, 10:15 is two-part, questions 2 and 3 below ]
10:15 first should be no.
Second should be yes.
Judge Alsup requests that Dawn find question #2... he is having trouble locating it. Mr. Van Nest offers his copy to the judge, but Judge Alsup finds his copy.
Mr. Van Nest and Mr. Jacobs are before the judge.
Judge Alsup: What does the defendant say?
Google: They are both different versions of question 2,
regarding 10:10 question, in order to be consistent, we have to say no.
Symbolic reference resolution... from Dr. Mitchell, the symbolic reference must be contained in the instruction. Reading back from page 3483 of the court transcript, "if the instructions only use numeric references, this does not infringe? Dr. Mitchell- would not infringe."
Google: Regarding 10:15 questions [ 2 and 3 as Judge Alsup refers to them],
1) should be yes, or, the symbolic reference must be in the instruction
2) can the instruction be considered symbolic if it has a numeric reference? the answer should be No.
Oracle: The contained question was a very narrow question. How do we determine what is contained is a numeric or symbolic reference? It depends on what happens downstream from it. Resolved dynamically rather than statically. No limitation in the court's instruction regarding directness or immediacy. A pointer to a field in a table [ is a symbolic reference ].
Absent a restriction, this is not the answer that the jury's question should be based on.
Google: Observation… both the jury questions' gist is, the jury is looking at the instructions as being numeric, and they are asking if they can subsequently become symbolic. This is not what the case is being tried on.
[Oracle wants to respond, but Judge asks them to wait. In his usual thinking mode, Judge Alsup has his left hand on his face, index finger beside his ear, the remaining three fingers curled in front of his mouth as he looks back and forth at documents on his bench.]
Judge Alsup: All right, you wanted to say something?
Oracle: Google wants to constrain using "exclusively" or "only", adding additional limitations. We looked at this in our briefing to the court. A pointer to a field in a table.
Judge: I don't agree with either of you. I will give my own answer, and you can address it on appeal.
Judge: Second question- is a stack the same as memory?
Oracle: A stack is not the same as memory. The claim is not limited to stack. Want clear no's to both questions.
Google: That wouldn't be right. Oracle said it correctly, the stack is a portion of memory for use in storing operands. All stack is memory. Not sure what synonymous means. All memory is not stack.
Judge: You both agree that stack is memory, but that all memory is not stack?
Oracle: Would confuse the jury.
Judge: Is it a correct statement that stack is memory, and that…
Oracle: Stack doesn't show up in the asserted claims. 520 patent.
Google: Either version would be correct.
Judge: Bring jury in... you lawyers aren't going to agree on anything, so bring the jury in.
[ jury enters ]
1) Is stack synonymous with memory? Is all memory stack?
Answer to both is No.
Judge: The '104 claim calls out symbolic references. The term symbolic reference means a reference that identifies data by a name other than a numeric reference, and is defined dynamically and not statically.
2) "Can a determined numeric reference in the computer instruction become a symbolic reference because of what happens downstream?"
3) regarding the symbolic reference in the '104, does the resolution of the symbolic reference need to happen immediately, i.e at the field ID position in the demonstration, or can the instructions be considered to be symbolic if they have numeric references such as the setup in the demo of string id table, class columns?
Judge: [ to the jury] If the computer instructions contain a numeric reference, then it does not turn into a symbolic reference downstream. But how do you determine if it is a numeric or symbolic reference in the first place? Refer to my previous instructions. A numeric reference doesn't get transmogrified into a symbolic reference downstream.
I think that's the best that I can do for you here. We welcome your notes. Please send on as many as you need. Please go back and continue your deliberations.
[jury leaves ]
Judge: Any objections? [ for appeal purposes ] Speak now or forever hold your peace.
Oracle: We still think the instructions are confusing.
Google: We object to the instructions on the '520.
Judge: I hope that both sides learn something from this on patent cases. Things aren't as clear cut here as they are when you are in your rooms with your whiteboards. I am making this observation to both sides, the one that decided to sue and the one that defends... That's our system.
[Both sides seem to be either amused or in at least one instance at the Google table, taken aback by the judge's imparting of his wisdom. There are strained smiles and some limited shaking of heads after the judge leaves.]
Update 2: Report #3:
12:50 PM. Google's team is all here. Dawn delivered another note from the jury. Some shaking of heads, grim faces. I hear footfalls in the hallway... yep, here come Oracle. Mr. Jacobs reads it and says under his breath, "Jesus". They want to know if something matches 3 out of 4 claims, is that infringement? The judge tells them no. It has to match everything. Then he tells the parties' lawyers, after the jury leaves, that if there is another hung jury, they'd best begin thinking about what comes next.
Update 3: Mirror_slap has the details:
Oracle v. Google, Monday May 21, part 3
12:40 ... most of Google's team files in.
Judge comes in. Reads a note from a juror, "Is it true that the DVM would be infringing the '104 patent via any resolution of a symbolic reference anywhere in the data fields as shown in the demonstrative" [ DVM is Dalvik VM ]
Judge: What do we say to this?
Oracle: The answer to the question is yes. Could recite the definition of "resolving", and note that there is no additional limitation.
Google: This is the same question in different form. The answer is no. Symbolic reference has to be in the instruction. Asking whether anything in the data is symbolic ignores the way the case has been tried. What is resolved is the resolution in the instruction itself. The clear answer to this is no.
Oracle: Mixing issues again... "if all other claim limitations have been met".
Google: I don't read the question the same way. Urge the court to be consistent with prior instructions.
Judge: Bring the jury in. They said that they weren't going to go home until they got an answer to this.
[ jury comes in]
"Is it true that the DVM would be infringing the '104 patent via any resolution of a symbolic reference anywhere in the data fields as shown in the demonstrative"
Judge: This is an evidentiary point. I can tell you the law. You aren't asking me if there is a claim instruction. I cannot give you any help on the answer to this question. This is why we try these cases to a jury. It is so that you figure it out. If you want to try your hand at another question, I will try to help you with that.
So I am sorry that I cannot answer every question. I know you are working very hard on this.
Is today one of the days when you are leaving at 1 o'clock?
Judge: Asks question about the '104. The other one has gotten all the way through?
Judge: So, not to be too negative about this, but you should be considering what the next steps will be if we get a hung jury. We will wait to get a note from the jury regarding when their deliberations will end today.
[ Dawn comes in with the note from the jury saying that they are done for the day. 1:15 PM. Will start again tomorrow morning, 8 AM.]
Update 4: Rachel King has now filed her report for the day's events, and she lists the questions and then the judge's answers. Here a taste, the part about the '104 patent. The questions were:
Regarding the symbolic reference issue in ‘104 claim, does the resolution of symbolic references need to happen immediately?
And King records the judges answers:
Or can the instructions can be considered to contain a symbolic reference by virtue of the linking of numeric references that happen first followed by resolution?
Can we determine a numeric reference in the instructions to become a symbolic because of what happens downstream?
Warning that this was not going to be a yes or no answer, Alsup told the jury that sometimes there are numeric references called out in claim language.
“In earlier instructions to the jury, I defined symbolic reference to mean a reference that identifies data by a name other than a numeric memory location of the data, and that is resolved dynamically rather than statically,” Alsup continued.
He explained that if the computer instructions contain a numeric reference, then “it does not some how turn into a symbolic reference because of what happens downstream.”
However, Alsup told the jury the question then becomes how do you determine if it is a numeric reference in the first place. He answered, “In determining whether the item is a numeric versus symbolic reference in the first place, you must consult the definition that I gave you of what a symbolic reference is.”
What does it all mean? It means no matter what decision the jury reaches, it is unlikely to have much link to technical reality. First, anyone with any technical ability in the jury pool was screened out. Second, nonsense was spewed by one of the technical experts. Third, the jurors can't get along, and some at least seem to be the type that want a parent figure to decide their spats, which would, I would imagine, incline them to find for a patent holder, if there is any way possible. As one comment here on Groklaw put it:
They are passing judgement on a patent on simulated execution for
a stack-based virtual machine, and they don't know what a stack is.
Extra humor points for the fact that they are arguing about a patent that the USPTO itself says preliminarily is invalid, although the jury has no idea about that, but Oracle just couldn't wait for that to be decided (maybe because they figured the jury wouldn't know their eyeball from their ear lobe), and the judge let them win that argument. So here we are. Hurried up and now we wait for a jury that doesn't know what a stack is. Charles Dickens could write yet another novel from this situation, don't you think?
This would be quite funny in a less serious context. But the prejudice to
Google is palpable. There is absolutely no way they can get a fair result
from such a broken and unfair legal system.
If they win it will be a complete fluke. Against such an unjust system, the
best defense money could buy might turn out to not be enough.
What will happen will be an immediate appeal, no matter who "loses" or who "wins". That, of course, favors Oracle, because the Federal Circuit hardly ever sees a patent it doesn't love. It was told by the US Supreme Court again today that it got a case wrong, in light of Mayo v. Prometheus, and were told to try again.
Honestly, software patents are a mess. That is the real problem. And the USPTO has issued invalid patents, and that leaves victims like Google to try to straighten out the mess they made, in a context where everyone is forced to take the ridiculous patents seriously, presumptively valid, because the USPTO issued them. When they shouldn't have.
Do you see why I tried to avoid writing about patent litigation? It's infuriating. And hilarious. And so, so sad. Shame on Oracle for pursuing patent aggression with the likes of the '104 patent. Really. Shame. This case will be taught in law schools, I've no doubt, and it will go down in history as the stupidest patent litigation in the history of the world.
I'm sure we'll see worse before the Big Boys decide to do something about the broken patent system. If this case helps people to see how broken it is, so much the better. I can't see any other good coming out of this.
Update 5: Our "volunteer" volunteer got sick and left at noon, but he still found a way to send us his report, to fill in a few more details, and I certainly wish to say thank you to him and all our volunteers, as doing this is a real sacrifice:
So - I will just try to add things that the reports of mirror_slap did not mention.
I do not really have anything meaningful to add to the first report. For the second report:
The first question from the jury was:
Can a determined numeric reference become a symbolic reference of something that happens downstream later?
The second question from the jury was along the same line, but from another member of the jury:
Does the resolution of symbolic references have to happen immediately, or [something about numeric references by the virtue of linking symbolic references. It was quite long, I never managed to get it completely].
These two questions were answered first together.
Oracle answered first and wanted to answer:
* yes to the first question
Google answered second and wanted to answer:
* no to the first part of the second question (does the resolution have to happen immediately)
* yes to the second part of the second question
* no to the first question
After that Oracle stepped up again and argued that:
* yes or "the symbolic reference must be contained in the instruction stream" to the first part of the second question
* no to the second part of the second question
*the contained questions are narrow and should be answered in a narrow way
After that Google stepped up again and argued that the questions should be answered as it first proposed because that is exactly what the expert witnesses told earlier. Google especially tried to nail the importance of the word "determined" I think.
*how does one determine if something is numeric or symbolic? And then something about that Google uses numeric addresses that point to a table that contains symbolic references and that things can be indirectly symbolic.
After that Oracle tried to step up again and was made to wait a few minutes. The lawyer then tried to argue that the "determined" argument of Google opens up something new and argued more about numbers first pointing to a table, but I think the judge's mind was already made up at that point.
After that the judge said:
"I do not agree with either of you" and continued with question 3:
"Is a stack synonymous with memory" and is all memory stack?
Oracle: clear no's.
Google: stack is a portion of memory.
Judge: something I did not catch.
Google: all stack is memory, memory is not confined to the stack.
The judge tried to get Oracle and Google to agree on this, but Oracle argued that it is factually correct but not in the claim language. And something about that it answers more than the jury asked. And that stack clearly is not synonymous with memory.
The judge seemed pretty much fed up at that time and said that they are not going to agree, to bring the jury back in, and that he will do the best he can.
He then read back the questions and answered the jury:
No and no to the stack question (the answer Oracle wished to give)
After that the jury got sent away again.
For the other two questions, he answered in a very lengthy way. He said that a determined numeric reference can not become a symbolic reference because of something that happens downstream. A computer can not make a numeric reference symbolic later. But he told the jury that to figure out if a reference is numeric or symbolic, they have to consider what happens downstream with it. They should look at the value contained in the instruction stream. If this value refers to an actual memory location, it is numeric. And that it is their job to determine this.
The judge still seemed a little upset to me, asked the lawyers if they have any objections. After they were noted he told them that things do not come down in the courtroom as in your rooms with the whiteboards. And I think that it is not as easy to understand for other people what all this is about. And that this went differently than either Oracle or Google imagined.
Update 6: We now have the transcript [PDF] of the day.