HYPERPHRASE TECHNOLOGIES, LLC v. GOOGLE INC.

Filing 136

BRIEF in Reply in Support re: 100 MOTION for Summary Judgment filed by GOOGLE INC. Regarding Noninfringement (Attachments: # 1 Exhibit 1, Summary of Claim Construction and Infringement Theories # 2 Exhibit 2, Terminology Used By the Parties to Refer to AutoLink) (Wolff, Jason)

HYPERPHRASE TECHNOLOGIES, LLC v. GOOGLE INC.

Doc. 136

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 1 of 34

IN THE UNITED STATES DISTRICT COURT FOR THE WESTERN DISTRICT OF WISCONSIN

HYPERPHRASE TECHNOLOGIES, LLC and HYPERPHRASE INC., Plaintiffs,

v.

Civil Action No. 06-cv-199-bbc

GOOGLE INC., Defendant. REPLY IN SUPPORT OF GOOGLE'S MOTION FOR SUMMARY JUDGMENT OF NONINFRINGEMENT OF U.S. PATENT NOS. 5,903,889 AND 6,516,321

Dockets.Justia.com

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 2 of 34

TABLE OF CONTENTS Page I. II. Introduction and Summary ................................................................................................1 HyperPhrase's New Claim Construction of "Data Reference" Provides Another Independent Basis for Finding Noninfringement of All of the Asserted Claims .................................................................................................................4 AutoLink Does Not Infringe the '889 Patent.....................................................................6 A. AutoLink does not infringe under HyperPhrase's old infringement theories. ............................................................................................7 1. 2. 3. B. AutoLink does not employ a "standardized format for addressing" data records in a plurality of databases. .................................7 AutoLink does not modify the first data record to create an address of the referenced record. ........................................................10 AutoLink does not modify the first data record at all..............................12

III.

AutoLink does not infringe under HyperPhrase's new, untimely infringement theories either. ................................................................................14 1. 2. AutoLink does not infringe under HyperPhrase's new literal infringement theory. ......................................................................15 AutoLink does not infringe under HyperPhrase's new doctrine of equivalents theory..................................................................16

IV.

AutoLink Does Not Infringe the '321 Patent...................................................................19 A. AutoLink does not satisfy the "when" or "real-time" limitation of claims 1 and 24. ...............................................................................................19 1. AutoLink does not satisfy the "when" or "real-time" limitation of claim 1 under HyperPhrase's old infringement theory..................................................................................19 AutoLink does not satisfy the "when" or "real-time" limitation of claim 1 under HyperPhrase's new infringement theory..................................................................................20

2.

B.

AutoLink does not satisfy the "modifier reference" limitation of the claims. ............................................................................................................22 1. AutoLink does not satisfy the "modifier reference" limitation of claim 1 under the correct construction of this term. ..................................................................................................22 i

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 3 of 34

TABLE OF CONTENTS (cont'd) Page 2. HyperPhrase cannot meet its burden to show that AutoLink satisfies the "modifier reference" limitation under its new, incorrect construction. ......................................................25

C. D. V.

AutoLink does not satisfy the "seemingly general" and "relatively specific" limitations of claim 86. .......................................................25 AutoLink does not satisfy the "specifying reference" limitation of claim 86. ..........................................................................................................26

Conclusion. ......................................................................................................................28

ii

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 4 of 34

TABLE OF AUTHORITIES Page(s) Cases 02 Micro Intern. Ltd. v. Monolithic Power Systems, Inc., 2006 WL 3300458, at *10 (Fed. Cir. 2006)........................................................................ 3 Engel Indust. Inc. v. Lockformer Co., 166 F.3d 1379, 1382-83 (Fed. Cir. 1999) ........................................................................... 2 Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., Ltd. 535 U.S. 722 at 739-740 (2002)............................................................................ 12, 14, 17 Gagnon v. Teledyne Princeton, Inc., 437 F.3d 188, 191 (1st Cir. 2006)....................................................................................... 3 In re Independent Service Organizations Antitrust Litigation, 114 F. Supp. 2d 1070, 1100 (D. Kan. 2000)....................................................................... 3 Inamed Corp. v. Kuzmak, 275 F. Supp. 2d 1100, 1118 (C.D. Cal. 2002) .................................................................... 3 Network Commerce, Inc. v. Microsoft Corp., 422 F.3d 1353, 1363 (Fed. Cir. 2005)............................................................................... 17 PC Connector Solutions LLC v. SmartDisk Corp., 406 F.3d 1359, 1364 (Fed. Cir. 2005)......................................................................... 16, 18 Rambus, Inc. v. Infineon Techs. AG, 145 F. Supp. 2d 721 (E. D. Va. 2001) ................................................................................ 3 Riefert v. South Central Wisconsin MLS Corp., 2005 WL 20055958, at *5 (W.D. Wis. Aug. 25, 2005)...................................................... 3 W.G. Pettigrew Distrib. Co. v. Borden, Inc., 976 F. Supp. 1043, 1052, 1050 (S.D. Tex. 1996) aff'd, 127 F.3d 34 (5th Cir. 1997) ............................................................................................................................ 3

iii

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 5 of 34

I.

INTRODUCTION AND SUMMARY What the parties admittedly agree on is dispositive of infringement. First, HyperPhrase agrees that Google does not infringe claim 27 of the '321 patent and

has withdrawn that claim from the case. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 41.) Second, HyperPhrase implicitly concedes that its old infringement theories are without merit, because it has abandoned these theories and put forward a new (fourth) set of infringement theories in its opposition, which Google addresses below. If the Court precludes HyperPhrase from advancing new theories at this stage of the litigation, which it should, summary judgment should be granted for this reason alone. Third, HyperPhrase does not dispute that it is bound by this Court's claim construction from the Microsoft litigation, according to which claims 1 and 24 of the '321 patent have a realtime requirement stemming from the "when" language.1 Since the undisputed operation of AutoLink requires manual user intervention, AutoLink does not meet the real-time limitations of at least claims 1 and 24 of the '321 patent and therefore does not infringe those claims. Fourth, the named inventor and HyperPhrase's new expert on validity, Carlos de la Huerga, agrees with Google that the terms "modifier reference" and "specifying reference" were expressly defined in the '321 patent. Since HyperPhrase offers no real rebuttal under Google's constructions of these terms, which are based on those express definitions, claims 1, 24, and 86 of the '321 patent are not infringed. Fifth and finally, HyperPhrase does not genuinely dispute how AutoLink works. It merely recasts AutoLink's operation in different terms, both selectively omitting key steps and conveniently relabeling Google's technology with names from HyperPhrase's patents.2 For instance, HyperPhrase agrees that:

1

HyperPhrase has submitted a table of claim constructions at Niro Decl. Ex. R (Dkt. No. 125-19). Google generally objects to the constructions not addressed in this brief as irrelevant to the issues before the Court, and to the extent that they are at odds with this Court's 2003 constructions and other constructions advanced by the parties. Google will respond to these constructions if further response is necessary in due course. 2 Examples include HyperPhrase's use of the term "link," which this court construed differently in its binding 2003 claim construction order than HyperPhrase uses it now, and the made up term "AutoLink data translation and

1

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 6 of 34

·

AutoLink operates on a separate, in-memory DOM file created from a web page, as opposed to the web page itself. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 38 and 40; Google Reply to Its PFOF (filed concurrently herewith) No. 55.) The URL (the "first URL" or "redirect request URL") that is added to the DOM file by the AutoLink client points to a process on Google's AutoLink server, not to some record. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 6-7; Google Reply to Its PFOF No. 54; Djabarov Decl. (Dkt. No. 33) at ¶ 4(b).) Before the AutoLink server processes the first URL, the user must manually intervene. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 9 ("After the user clicks on the light blue link, the Google `tbproxy' server receives that URL... Once received... the redirect causes a server to build[] what is known as a `redirect URL,' (which is different from a redirect request URL)."; Google Reply to Its PFOF No. 54.) The second URL (also called the "redirect URL") that is built by the AutoLink server is not added to the separate, in-memory DOM file, or to any web page by AutoLink, and HyperPhrase offers no evidence to the contrary. (Djabarov Decl. (Dkt. No. 33) at ¶ 14(g).)

·

·

·

As a result of the key issues which are undisputed, this case now largely reduces to either legal issues or to an essentially an unopposed motion for summary judgment under HyperPhrase's now abandoned infringement theories.3 Realizing this, as noted above, HyperPhrase re-tooled its case a third time and now offers with its opposition brief untimely new and conflicting theories by its experts Paul Thompson (his fourth infringement theory) and Carlos de la Huerga, which include new technical theories, new claim constructions, and the fallback theory of infringement under the doctrine of equivalents. For example:
collection server" ­ there is no such thing. Similarly, HyperPhrase has ignored this court's earlier construction of the terms "record" and "data record." 3 The first four pages of HyperPhrase's opposition are dedicated to objecting that Google has placed more issues before the Court than are allegedly permitted under the scope of the remand from the Federal Circuit. Not only does HyperPhrase misstate the Federal Circuit's opinion and order, which remanded the case to this court to evaluate whether "AutoLink does not meet other limitations of the claims," (Niro Decl., Dkt. No. 125-2 at No. 12 (emphasis added)) but it rehashes the arguments HyperPhrase made in opposition to Google's motion to amend its answer. This Court already found that the parties' stipulated dismissal expressly allowed Google to put new issues before this Court. (See Order Granting Motion to Amend Counterclaims, Dkt. No. 118; see also Motion to Amend/Correct Answer and Counterclaims, Dkt. No. 98 and Brief and Reply ISO Motion to Amend/Correct Answer and Counterclaims, Dkt. No. 114.) Moreover, the authority HyperPhrase cites refers to what the appellant (HyperPhrase) may do on remand, not the appellee (Google). Engel Indust. Inc. v. Lockformer Co., 166 F.3d 1379, 1382-83 (Fed. Cir. 1999) ("The scope of issues presented to this court on appeal must be measured by the scope of the judgment appealed from."). HyperPhrase cites to no authority that precludes Google, the appellee, from presenting issues that were not decided or are not within the scope of the 2006 decision. (Dkt. No. 127 at 4.) Lost in HyperPhrase's misapplication of this rule to Google is that HyperPhrase itself violates the actual rule, because its entire opposition is premised on a new claim construction

2

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 7 of 34

·

Where Dr. Thompson's third infringement report offered no theory of infringement under the doctrine of equivalents, having concluded the claims were literally infringed (indeed "identical" to the patent), he now provides a new conclusory and unsupported doctrine of equivalents analysis. Where Dr. Thompson's first three expert reports make no mention of the DOM file, he does so for the first time now and, in the process, contradicts HyperPhrase's inventor and new validity expert Mr. de la Huerga ­ meaning the patents are either invalid, or not infringed, or both. Where Dr. Thompson's first three expert reports say that the "record" (a.k.a. "data record") required by the '889 patent is the web page with an AutoLink token, his untimely new report concludes that the first URL itself is the data record, and an AutoLink token in that URL is the "data reference," despite claim language that makes this new theory completely untenable.

·

·

Dr. Thompson's fourth expert report is untimely and should be disregarded by the court for that reason alone, but it is also inconsistent and conclusory.4 Riefert v. South Central Wisconsin MLS Corp., 2005 WL 20055958, at *5 (W.D. Wis. Aug. 25, 2005) (precluding "supplemental" expert report submitted after the date permitted under Rule 26(a)(2)(C) as untimely); 02 Micro Intern. Ltd. v. Monolithic Power Systems, Inc., 2006 WL 3300458, at *10 (Fed. Cir. 2006) (upholding exclusion of expert affidavit containing new expert opinions submitted in response to a motion for summary judgment).5

for the term "data reference" that it (the appellant) did not present on appeal, for a term that has already been construed by the Federal Circuit. 4 This is the second time in this case HyperPhrase has offered new infringement theories in response to Google's motions for summary judgment. See Dkt. Nos. 58 and 59. 5 As explained in Google's 2006 motion to strike, there is no shortage of authority for this court to disregard HyperPhrase's untimely infringement theories. See, e.g., Rambus, Inc. v. Infineon Techs. AG, 145 F. Supp. 2d 721 (E. D. Va. 2001) (excluding a "supplemental" expert report served one month after the close of discovery that "made substantial changes to the substance of, and the bases for, his opinions on the issues of literal infringement and the validity of the patents."); Gagnon v. Teledyne Princeton, Inc., 437 F.3d 188, 191 (1st Cir. 2006) (reviewing exclusion of late expert disclosure in summary judgment proceeding and remanding to determine if "harmless error" escape valve of section 37(c)(1) applied); W.G. Pettigrew Distrib. Co. v. Borden, Inc., 976 F. Supp. 1043, 1052, 1050 (S.D. Tex. 1996) aff'd, 127 F.3d 34 (5th Cir. 1997) (disregarding affidavits in summary judgment proceeding to the extent they introduced new matters that, per Rule 26(e), should have been provided in supplemental responses to interrogatory asking for the basis of Plaintiffs' contentions); Inamed Corp. v. Kuzmak, 275 F. Supp. 2d 1100, 1118 (C.D. Cal. 2002) (precluding a defense the defendant had raised for the first time during its opposition to summary judgment, because the defendant failed to supplement its prior interrogatory answers with the defense, and the plaintiff could only respond to the new defense in its reply); In re Independent Service Organizations Antitrust Litigation, 114 F. Supp. 2d 1070, 1100 (D. Kan. 2000) (excluding expert declaration submitted in response to a motion for summary judgment).

3

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 8 of 34

Nevertheless, HyperPhrase has the chutzpah to make a preemptive strike against Google's ability to respond to HyperPhrase's new infringement theories, suggesting any such argument cannot be considered in connection with this motion. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 52-53.) While Google agrees that the Court should not consider new liability theories, application of this rule begins with exclusion of HyperPhrase's new infringement theories, including, but not limited to, its doctrine of equivalents theories. However, if the Court will not disregard Dr. Thompson's fourth declaration and HyperPhrase's new claim constructions as untimely, then the Court ought to consider why HyperPhrase's untimely arguments are wrong too. A table summarizing the parties' claim constructions, HyperPhrase's changing infringement theories, and why in a nutshell Google does not infringe is attached as Exhibit 1. A second table, attached as Exhibit 2, shows some terminology pertaining to the alleged infringing product. While the parties agree how AutoLink operates, they have not always used precisely the same terms to describe it.. II. HYPERPHRASE'S NEW CLAIM CONSTRUCTION OF "DATA REFERENCE" PROVIDES ANOTHER INDEPENDENT BASIS FOR FINDING NONINFRINGEMENT OF ALL OF THE ASSERTED CLAIMS As a general matter, Google addresses HyperPhrase's new infringement theories below after first considering its old infringement theories, which HyperPhrase has largely abandoned. However, Google first wishes to single out HyperPhrase's new claim construction for the term "data reference" because this new construction ­ notwithstanding that the Federal Circuit already has construed the term -- alone compels summary judgment of noninfringement across the board in Google's favor. The Federal Circuit held that all of the claims require a "data reference" and construed the term. (Niro Decl., Ex. A (Dkt. No. 125-2) Federal Circuit Opinion at 10.) While HyperPhrase pays the Federal Circuit construction lip service, it in fact offers a new construction in its opposition to Google's motion for summary judgment of invalidity. This new construction is: 4

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 9 of 34

"Data reference (DR):" This term has been defined by the Federal Circuit in this case as follows: A unique phrase or word which may be used in a record to refer to another record or record segment, and a data reference may refer to one or more than one record. See Niro Decl., Ex. [A], U.S. Court of Appeals for the Federal Circuit Opinion, Dec. 26, 2007, p. 10. Consistent with the Federal Circuit's opinion, HyperPhrase believes that a "data reference" is the text in a record normally displayed for a typical user to read and not hidden computer codes. (Niro Decl. Ex. R (Dkt. No. 125-19) at 18 (emphasis added).) The new language HyperPhrase has added to the definition is underscored.6 This construction, which is not mentioned in HyperPhrase's opposition to Google's motion for summary judgment of noninfringement, is the key basis upon which HyperPhrase and its inventor now attempt to distinguish several prior art references, including, for instance, the Graham prior art. (See e.g. de la Huerga Rebuttal Invalidity Rpt. (Dkt. No. 130-2) at 144-145 of 307 and 173 of 307.) Mr. de la Huerga uses the term "hidden" at least 70 times in his rebuttal report on validity and states that the hidden, separate file approach ­ which is exactly how AutoLink operates ­ is antithetical to the fundamental point of his invention:

(Id. at 25-27 and 144-145 of 307; see also de la Huerga Decl. ISO Invalidity Opp. (Dkt. No. 130) at 5, ¶ 8 ("In the claims and patents, this display text is parsed, analyzed, examined, or identified ... the hidden codes are not parsed, analyzed, examined, or identified.").) The parties do not dispute that AutoLink operates on the hidden DOM file, not on the web page which the user sees, and they do not dispute that the DOM file is a separate, inmemory file (that is, it is a "hidden" file) created from a web page. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 38; Google Reply to Its PFOF No. 55.) In fact, HyperPhrase's opposition
6

Google disagrees that HyperPhrase's construction is correct or supported in any way by the Federal Circuit's decision, the written description, or any other intrinsic or extrinsic evidence.

5

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 10 of 34

papers characterize the DOM file as "a document presented as a logical structure rather than as a collection of tagged words. In essence, DOM is a means of defining a document as a treelike hierarchy of nodes in which the document is an object containing other objects." (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 39 (emphasis added).) HyperPhrase offers absolutely no evidence, because it cannot, that the DOM file is the file that is displayed in the browser or that it is ever "normally displayed for a typical user to read." Thus, if the HyperPhrase inventions preclude operation on and/or modification of a separate, hidden file such as the DOM file or a file that is not normally displayed for a typically user to read, then there can be no infringement by AutoLink of any of the asserted claims because the "data reference" limitation as now construed by HyperPhrase is not satisfied. It is no accident that HyperPhrase has not even tried to satisfy its burden to prove infringement under its new claim construction, which it uses to try to save the claims from invalidity. III. AUTOLINK DOES NOT INFRINGE THE '889 PATENT. Google moved for summary judgment of noninfringement on three independent bases with regard to the '889 patent. First, Google asserted that based on the undisputed manner in which AutoLink worked, and as noted by the Federal Circuit, AutoLink could not satisfy the "standardized format for addressing" limitation.7 (Google's Noninf. Br. (Dkt. No. 103) at 8-10.) Second, Google asserted that the first URL (the "redirect request URL"), which is added to the DOM file by AutoLink, did not satisfy the "address of the referenced record" limitation. (Id. at 10-12.) Third, Google asserted that it did not satisfy the "modifying" limitation because it did not modify the first data record (the web page) at all. (Id. at 12-15.)

7

HyperPhrase asserts that Google's argument is that it is practicing the prior art and therefore it cannot infringe. That was not Google's argument. Rather, Google's argument was that, in the respects relevant to the address formatting requirement, AutoLink works like the system HyperPhrase said its invention was not. HyperPhrase also argues that the '889 patent is the prior art system described in the '567 patent. However, the '567 patent claims priority to the `889 patent (Niro Decl. Ex. J (Dkt. No. 125-11) '567 patent at 1:17-21) and despite that HyperPhrase now claims the `889 patent is prior art to its own '567 patent, HyperPhrase never cited it as such to the Patent Office (id. at cover page).

6

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 11 of 34

In response, HyperPhrase essentially concedes there is no infringement under its earlier infringement theories and offers a mix of new technical theories, new claim constructions, and the new, fallback position of infringement under the doctrine of equivalents.: · First, according to HyperPhrase, the alleged infringing "references" (also called "data references") are no longer the AutoLink tokens (e.g. ISBNs, VINs, postal addresses) in a web page, but the AutoLink tokens in the first URL alone. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 34; Fourth Thompson Decl. (Dkt. No. 129) at ¶ 22.) In other words, HyperPhrase is now saying the first URL itself is the claimed data record. Even apart from the absurdity of this argument on its face, it fails using any construction of the claim language because neither the first URL, nor any of the other URLs, is a record in a database. Second, HyperPhrase re-wrote the limitation "standardized format for addressing" into the different and broader "standard convention for retrieving the data records," or "common addressing conventions," which HyperPhrase alleges is satisfied by any URL. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 15-22.) This claim construction is unsupported by the intrinsic record and contradicted by HyperPhrase's arguments regarding validity. Third, HyperPhrase now suggests that the second URL (also called the "redirect URL"), rather than the first, is the address of the referenced record. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 34 ("AutoLink sends the [first] URL ... to the ... server where it is converted to a new [second] URL ... this new URL is then sent to recover the second data record.").) However, it offers no evidence that the second URL, which is processed by a third server (whether it is Google's or a third party's), is in fact the address of the alleged referenced record that is returned.

·

·

As is shown in further detail below, none of these theories has merit. A. AutoLink does not infringe under HyperPhrase's old infringement theories. 1. AutoLink does not employ a "standardized format for addressing" data records in a plurality of databases.

HyperPhrase's original infringement position was that the second URL (the redirect URL) embodied the "standardized format for addressing" the data records in the plurality of databases.

(Third Thompson Decl. (Dkt. No. 129-4) (2/11/08 Thompson infringement chart) at 35 of 37.) Google moved for summary judgment that this address was not a "standardized format for 7

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 12 of 34

addressing data records" in a plurality of databases, because the undisputed record shows that there is not a standardized format for addressing data records in the alleged plurality of third party databases. (Niro Decl., Ex. A (Dkt. No. 125-2) Federal Circuit Opinion at 2-3 (generally describing how AutoLink works) and 14 (noting that the second URLs have multiple different formats in the context of the '567 patent).) HyperPhrase's only response is to rewrite the claim limitations and argue new infringement theories. It does not dispute that, under its prior infringement theory and Google's construction, there is no infringement. This is as far as the Court need go. Nevertheless, even considering the new arguments HyperPhrase loses. To address HyperPhrase's new construction first, Claim 1 of the '889 patent is a system claim with multiple means-plus-function elements. It reads: 1. A computer system with a plurality of data records on a plurality of databases, and a standardized format for addressing said data records, said computer system comprising: (a) a user interface having an interactive display program for requesting one of said data records and displaying a plurality of interface supported data formats; (b) means for receiving a reference to a first data record from said interactive display program; (c) means for retrieving said first data record; (d) means for parsing said first data record to identify a reference to a second data record; (e) means for modifying said reference to said second data record to create an address, said address being operable to retrieve said second data record; and (f) means for sending said modified first data record to said interactive display program. The parties do not dispute that the preamble is a limitation, but only what that preamble means. Google says it means that the plurality of data records on the plurality of databases are addressed using a standardized format. Thus in claim element 1(e) when the "address" is created, it must have a standardized format for the plurality of data records on the plurality of databases. Recognizing that AutoLink does not work this way, HyperPhrase now says that the "standardized format for addressing said data records" simply means any URL at all. It does this by re-writing the claim phrase "standardized format for addressing data records" as "standard convention for addressing." (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 15 and 22.)

8

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 13 of 34

As support for this new claim construction, HyperPhrase explains that the structure for this means-plus-function claim element is now the URL Cipher 140 that converts an HTTP URL into an FTP URL. HyperPhrase's argument is flawed because both the HTTP URL and FTP URL are already URLs (as HyperPhrase itself acknowledges at pages 21 and 22 of its opposition brief), which makes HyperPhrase's construction meaningless. If any URL qualifies, both URLs are a "standard convention for addressing," and therefore the alleged corresponding structure doesn't in fact take something which is not a "standard convention for addressing" and make it so. HyperPhrase's attempt to escape noninfringement by re-writing the claim also fails for a more fundamental reason. The claim requires a "standardized format for addressing" data records in a plurality of databases, not a "standard convention for addressing" (whatever that is), meaning that the format of the address for accessing the data records in the plurality of databases must be the same. The point of this requirement is clearly stated in the Summary of the Invention: It is the object of this invention to provide a means of processing and converting existing data records formatted, structured, and accessed according to a multitude of disparate standards to common standards by which they may be accessed, controlled, and/or displayed through a single interactive display program. ('889 patent at 2:65-3:3.) Even Carlos de la Huerga, the first named inventor, agrees with Google's construction, since he used it in his rebuttal report to argue claim 1 of the '889 patent is not invalid in view of the prior art. (de la Huerga Rebuttal Invalidity Rpt. (Dkt. No. 130-2) at 303-304 of 307 (arguing that the Anthony prior art "does not propose a standard addressing scheme" because "[t]he link is created ... using whatever (random) address the record with the matching topic name is stored at.").) Contrary to the invention, AutoLink does not employ a standardized format for addressing data records on a plurality of databases. It is undisputed that there are over 20 different formats for the alleged infringing second URL (the redirect URL), which HyperPhrase says are used to retrieve the data records of interest. (Niro Decl., Ex. A (Dkt. No. 125-2) Federal 9

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 14 of 34

Circuit Opinion at 2-3 (generally describing how AutoLink works) and 14 (noting that the second URLs have multiple different formats in the context of the '567 patent).) Thus, claims 1 and 7 of the '889 patent are not infringed.8 2. AutoLink does not modify the first data record to create an address of the referenced record.

HyperPhrase alleged that the first URL (the redirect request URL) was the address of the second data record which satisfied element 1(e) of the '889 patent. (Third Thompson Decl. (Dkt. No. 129-4) at 22 and 23 of 37, ¶ 50.) In response to Google's motion for summary judgment, HyperPhrase seeks to reargue a claim construction issue pertaining to whether that claimed address must point to a particular referenced record or records (as Google says and the patent shows for this means-plus-function element), or can instead vaguely point to anything that might eventually lead to an alleged referenced record (as HyperPhrase says). The Federal Circuit already rejected, at least implicitly, HyperPhrase's attempt to broaden its claims so that merely using some information from a first document as part of a process of retrieving other information in a second document is sufficient to satisfy the claims. (Niro Decl. Ex. A (Dkt. No. 125-2) Federal Circuit Opinion at 10, fn 5 ("Simply being generally `used in the process' of creating a link between records is insufficient because the stated definition in the '321 patent requires that the reference itself refer to a record."). Claim 1 requires that the first data record be modified to create the address of the second, referenced record. HyperPhrase has no evidence that AutoLink does this, because it does not. As to the additional language, "said address being operable to retrieve said data record," contrary to HyperPhrase's argument, that language does not broaden the address limitation it modifies. Rather, the additional words narrow the claim (as more words in a claim always do), and further support Google's position that the claim is not infringed, because the parties agree that the first URL is not the actual address of a data record. (Google Reply to Its PFOF No. 54.)
8

If, on the other hand, HyperPhrase means to say the first URL (the redirect URL) is the standardized format for addressing, that does not work either (as is explained in greater detail below), because the first URL is not the address of a plurality of data records on a plurality of databases. It merely points to a process on the AutoLink server. (Google Reply to Its PFOF No. 54.)

10

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 15 of 34

Google's view of "modifying the first data record to create an address" of the referenced record is also consistent with the intrinsic record, including the prosecution history, which HyperPhrase largely ignores. In particular, HyperPhrase disregards the arguments it made to the Patent Office to obtain allowance of the '889 patent. There, HyperPhrase emphatically argued the address limitations to distinguish the Meske prior art, asserting the address created by Meske was different than HyperPhrase's invention because Meske did not "modif[y] the first data record to create an address of the referenced record." (Woodford Decl. (Dkt. No. 26) Exhibit B ('889 patent file history) at GOOG056999-57000 (10/6/98 Response A at pp. 3-4) (emphasis added).) More specifically, HyperPhrase argued:

(Id.) HyperPhrase thus confirmed during prosecution the natural reading of claim 1: the "address" required by the claim is the "address of the referenced record." So important was this feature of the invention to HyperPhrase that it made the argument three times in the same paragraph: "to create an address of the referenced record," "to refer to another file by its address," and "to create an address of the referenced record." Id. (emphases added). There is no dispute that under Google's construction the first URL is not literally "the address of the referenced record" and that it does not "refer to another file by its address." (Google Reply to Its PFOF No. 54.) Thus, it cannot literally satisfy this claim limitation. Furthermore, while HyperPhrase offered no timely doctrine of equivalents theory initially (as it does now and which is addressed below), it is not entitled to any theory of infringement under 11

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 16 of 34

the doctrine of equivalents in any event because of the narrowing arguments it made during prosecution, which HyperPhrase has not rebutted. Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., Ltd. 535 U.S. 722 at 739-740 (2002). 3. AutoLink does not modify the first data record at all.

HyperPhrase alleged that the data record operated on in the means-plus-function claims was the web page from the plurality of databases. But as already noted, the AutoLink client does not operate on the web page and in particular does not modify the web page in any way; it instead operates on a separate, hidden DOM file created by the browser. These facts regarding how AutoLink operates are not genuinely disputed9, and are perhaps the reason HyperPhrase has abandoned its earlier infringement allegations. If the Court agrees with Google that the claim requires modification of the parsed data record, and not some other record, then it should find that AutoLink does not infringe the '889 patent. Based on the claim language and intrinsic record, there should be no doubt that "modifying" must be performed on the same file (e.g. the web page if it is a data record) which is first retrieved and parsed. (See Google's Noninf. Br. (Dkt. No. 103) at 12-14.) HyperPhrase's validity expert and inventor both agree with Google on this question, since this is the basis upon which they attempt to distinguish much of the prior art. (Nirenburg Decl. (Dkt. No. 45) at 8, ¶ 19 ("Again, limitation 1(e) requires `modifying' the reference itself. If the short form address is a `reference' under Google's theory, then HTML interpretation leaves this string of text intact and unchanged in the base HTML document." (emphasis added)); de la Huerga Rebuttal Invalidity Rpt. (Dkt. No. 130-2) at 5 of 307 (adopting Dr. Nirenburg's declarations and reports in their entirety) at 26 of 307 ("In fact, it is not consistent with the '321 patent to have a reader of a record see a data reference that can be read and a modifier reference that appears in a silent, hidden, and unknown to the reader, separate file." (emphasis added).) To reach a different conclusion which serves its infringement theory, HyperPhrase ignores the claim language and

9

See Google Reply to Its PFOF Nos. 55 and 61.

12

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 17 of 34

does not substantively discuss the prosecution history. Instead, HyperPhrase waxes on in a mostly irrelevant and misdirected discussion of the DOM file and Anne Frank's Diary. The plain language of claim 1 of the '889 patent requires that the self-same file is operated on in each means-plus-function element -- the term "modifying" is used, the antecedent basis is consistent throughout the claim, and each means-plus-function element refers back to the operation of the previous function performed on the same record. The claim repeatedly refers to the thing being operated on ("said first data record") in the "retrieving," "parsing" and "sending" functions. The "modifying" function refers to modifying "said reference," which was found in "said first data record." Since the "sending" function operates on "said modified first data record," the modifying step must refer to "said reference" in the "said first data record," otherwise the claim would be indefinite because it would not convey what the modified first data record is. The prosecution history also confirms Google's view. When HyperPhrase was confronted with the prior art, which included (but was not limited to) Meske, it argued that in its invention the file that was operated on was the first file (the file that was retrieved and parsed) and that this distinguished its claims over the prior art, which did not modify the parsed file but rather created additional files that were modified. (Woodford Decl. (Dkt. No. 26) Exhibit B (`889 patent file history) at GOOG056999-57000 ("...[in Meske] there is no modification of the parsed file to refer to another file by its address... There is simply no teaching of linking records by modifying the parsed record to create an address to the referenced record." (emphasis in original).) The point was that the invention modified the same file and not something else, for example the "additional files that were created."10 (Id.)

10

Returning to HyperPhrase's Anne Frank's diary analogy, it makes Google's point. First, the German and English versions of Ann Frank's Diary are different books. The claim requires operating on one of these books and modifying that version, not another version. Second, regarding whether the modified record is the one in the database or the one in the browser, HyperPhrase's validity expert argued that the record had to be in the plurality of databases (see Nirenburg Decl. (Dkt. No. 45) at 12, ¶ 32 (noting the prior art does not satisfy claim element 1(e) [of the '889 patent] because "the [Anthony] database already contains the identifiers")), though for Google to succeed on this issue it does not matter where the record is located, because AutoLink does not modify either alleged record -- it operates instead on the hidden and separate DOM file.

13

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 18 of 34

Since there is no dispute that AutoLink operates on a second file (the DOM) and not what HyperPhrase says is the "first data record" (the web page which is retrieved and parsed), AutoLink cannot literally infringe claim 1 of the '889 patent. AutoLink further cannot infringe this claim under the doctrine of equivalents because of the narrowing arguments HyperPhrase made to get the claim allowed. Festo Corp. 535 U.S. at 739-40. B. AutoLink does not infringe under HyperPhrase's new, untimely infringement theories either.

Given that AutoLink does not infringe under HyperPhrase's old infringement theories, HyperPhrase employs a two prong response to Google's motion. First, it appears to change its theory of literal infringement for the "modifying means" limitation to point to operation on something other than the web page (or even the DOM). Second, HyperPhrase now says Google infringes under the doctrine of equivalents too, even though HyperPhrase expressly alleged the infringement was literal and offered no doctrine of equivalents theory before. (Third Thompson Decl. (Dkt. No. 129-4) at ¶ 36.) Neither new theory has merit. HyperPhrase suggests in its new infringement theory that the "data record" is the first URL (the redirect request URL), the "reference" or "data reference" is information in this URL, and the second URL (the redirect URL) is the "modified reference " or "modified data reference." (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 31-35 (comparing converting one URL to another URL to the AutoLink process as the same "way" the patent discloses in Figs. 14A and 14B); Fourth Thompson Decl. (Dkt. No. 129) at ¶ 22 ("The AutoLink [ ] server extracts the identifying information from the first URL to create a second URL.").) The trouble with HyperPhrase's new theory is that HyperPhrase never follows through and analyzes the remainder of the claim limitations under this theory, nor does it offer evidence to support the remainder of its infringement allegations that is consistent with this new theory. For instance, HyperPhrase still refers to web pages (not URLs) as the "data records" in the "plurality of databases" in its infringement charts (Third Thompson Decl. (Dkt. No. 129-4) at 32 of 37) and offers no evidence that the first URLs are "data records" in a "plurality of databases," which of course they are not. 14

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 19 of 34

1.

AutoLink does not infringe under HyperPhrase's new literal infringement theory.

First, under HyperPhrase's new construction for the term "data reference," the first URL is a "hidden computer code" because it is not visible. (Djabarov Decl. (Dkt. No. 33) at ¶¶ 4 and 8.) Thus, it cannot be a "reference" as the claim requires if HyperPhrase intends to be internally consistent with its infringement and validity theories. HyperPhrase offers no evidence that the first URL is "the text in a record normally displayed for a typical user to read and not hidden computer codes." Second, the first URL is plainly not a data record. This Court construed the term "record" and "data record" to mean "components of a database composed of fields arranged in a particular predefined structure." (Woodford Decl. (Dkt. No. 35) Exhibit S (6/18/03 claim construction order) at 6-7.) HyperPhrase offers no evidence that the first URL is a "data record" in any database. Third, the operations in the other claim elements must also pertain to the first URL if the first URL is the "data record." But HyperPhrase offers no evidence that the other claim elements, besides the "means for modifying," also operate on the first URL and they plainly do not. If on the other hand HyperPhrase means to say that the web page (or even the DOM) is the data record for some limitations while the first URL is the data record for other limitations, its theory fails for that reason alone ­ it cannot identify one thing as the "data record" for some limitations and something else as the data record for other limitations. Fourth, one of the operative claim terms is "modifying said reference." HyperPhrase offers no evidence that the first URL is modified. Creating a new, second URL (the redirect) is not the same as or equivalent to "modifying" the first URL. HyperPhrase offers no evidence that any part of the first URL is "modified" (because it is not). Fifth, HyperPhrase offers no evidence that a second URL (which typically points to a third party server) is the "address of the referenced record" or that it "refers to another file by its address." In fact, HyperPhrase offers no evidence regarding how the third party servers, or even 15

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 20 of 34

the Google Map server, process or use the second URL (the redirect URL), so it has no basis to conclude that the second URL is in fact the address of the referenced record. Any one of the above grounds is a sufficient basis for the Court to find that AutoLink does not infringe claims 1 and 7 of the '889 patent under HyperPhrase's new literal infringement theory. 2. AutoLink does not infringe under HyperPhrase's new doctrine of equivalents theory.

HyperPhrase's principal new doctrine of equivalents theory pertains to the "means for modifying." This is because AutoLink modifies the hidden DOM file and not the web page, so it cannot literally infringe under HyperPhrase's old infringement theory. However, under HyperPhrase's new infringement theory, the thing to be modified is allegedly the first URL itself and not the web page. Therefore, HyperPhrase's new equivalents theory pertains to its old infringement theory, under which the web page was the data record. HyperPhrase has no equivalents theory that pertains to its new infringement theory, under which the first URL is the data record. Nevertheless, the "new" doctrine of equivalents theory relating to the old infringement theory fails in any event. First, because it is new, the Court should refuse to entertain it. In Dr. Thompson's third infringement report served just two weeks before Google filed this motion, Dr. Thompson stated that "The infringement is literal infringement." (Third Thompson Decl. (Dkt. No. 129-4) at 16, ¶ 36; see also id. at 33-37 (stating that the alleged infringing structure in AutoLink is "identical structure" to each of the limitations 1(b), 1(c), 1(d), 1(e), and 1(f)).) Likewise, in his infringement charts he does not even mention the DOM file, and has no analysis under the doctrine of equivalents. This is not the particularized testimony and linking argument required to meet HyperPhrase's burden under the doctrine of equivalents. See PC Connector Solutions LLC v. SmartDisk Corp., 406 F.3d 1359, 1364 (Fed. Cir. 2005) ("[h]aving presented the district court with only conclusory statements regarding equivalence, without any particularized evidence and 16

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 21 of 34

linking argument as to the `insubstantiality of the differences' between the claimed invention and the accused device ... [Plaintiff] is now foreclosed from invoking the substantive application of the doctrine of equivalents."); see also, Network Commerce, Inc. v. Microsoft Corp., 422 F.3d 1353, 1363 (Fed. Cir. 2005) (holding that evidence purportedly showing the doctrine of equivalents "must be presented on a limitation-by-limitation basis. Generalized testimony as to the overall similarity between the claims and the accused infringer's product or process will not suffice."). Second, the prosecution history precludes any scope of equivalents for the "mean for modifying" limitation. As Google pointed out in its opening brief, the "modifying" limitation was the subject of back-and-forth argument between HyperPhrase and the Patent Office. There, HyperPhrase made clear that the same file is retrieved, parsed, modified, and sent to the "interactive display program." (Woodford Decl. (Dkt. No. 26) Exhibit B ('889 patent file history) at GOOG056999-57000 ("...[in Meske] there is no modification of the parsed file to refer to another file by its address... There is simply no teaching of linking records by modifying the parsed record to create an address of the referenced record." (emphases added)).) HyperPhrase's new equivalents argument swallows not only the plain language of the claim, but disregards the repeated arguments it made that the parsed file, and not something else, is the file that is modified. Thus, HyperPhrase is not entitled to the doctrine of equivalents for this limitation. Festo Corp. 535 U.S. at 739-40. Third, HyperPhrase's new doctrine of equivalents theory fails in substance. To make the argument at all, HyperPhrase must re-draft the function, way, and result of claim element by cobbling together several disjointed and freshly minted embodiments. Regarding the "function," HyperPhrase says the AutoLink process function is to "insert a URL to link the first data record with a second data record." However, HyperPhrase offers no evidence that the second URL (which is allegedly the "modified reference") is inserted in the DOM or any other file, because, of course, it is not. Further still, HyperPhrase offers no evidence of how the second URL corresponds to the referenced data record ­ HyperPhrase has 17

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 22 of 34

taken no discovery of any of the third party servers that store or return whatever information is returned in response to their processing of the second URL. Regarding the "way" in which the invention works, HyperPhrase simply states that AutoLink is the same because AutoLink uses HTML. This is misdirection. HTML is a document format standard ­ the Hypertext Markup Language. HTML is not the "way" in which the "reference" is "modified."11 Among other things, what HyperPhrase's analysis lacks is evidence and linking argument regarding the way in which the '889 patent modifies the reference in the first data record to create the address of the second data record, and how what AutoLink does is insubstantially different. PC Connnector, 406 F.3d at 1363. Thus, HyperPhrase's analysis fails on this front too. Lastly, regarding the "result," the patent requires creating the address of the second data record in the first data record. The first URL created by the AutoLink client, as is noted above, is not the address of the referenced record, but a "redirect request URL," which points to the AutoLink proxy server and to no records at all. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 7.)12 And the second URL (the redirect URL), created by the AutoLink server is not placed in any record, nor does HyperPhrase offer any evidence that it is. Any of the above reasons is a sufficient basis for finding that AutoLink does not infringe claims 1 and 7 of the '889 patent under HyperPhrase's new doctrine of equivalents theory.
11

If the modified data reference is now the second URL, which is not added to the DOM or any other web page, it is unclear how the URL satisfies HyperPhrase's "HTML" requirement in any event. A URL is not HTML, nor does HyperPhrase offer evidence that it is. 12 HyperPhrase disputes that a "redirect request URL" points to no records at all. This is a red herring, because HyperPhrase is conflating the "redirect request URL" or "first URL" with other URLs. In any event, HyperPhrase's position finds no support in the Djabarov Decl. (Dkt. No. 33), because he stated the opposite of what HyperPhrase suggests. The AutoLink creator noted that the information that can be returned by the AutoLink feature is generally stored by third parties, except in the case of Google Maps. He does not state that the AutoLink proxy server (which receives and processes the first URL) stores data records, and in fact says it does not store the alleged data records at all. (Djabarov Decl. (Dkt. No. 33) ¶ 12(f) and 12(g) (describing what happens after a redirect URL is received by a browser: "then the browser would send the second URL [read: the redirect URL] to a third server (the Google Maps server).".) HyperPhrase offers no evidence of how any of the so-called data records are stored on third party servers or on the Google Maps server, or that the AutoLink server and the separate Google Maps server (or any of the other servers) are the same thing (they are not). To the extent the court believes this is sufficient to raise a genuine issue of fact, it would do so only with regard to AutoLink redirecting to Google Maps and not to the third party data providers such as Amazon, Carfax, Yahoo! Maps, FedEx, etc.

18

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 23 of 34

IV.

AUTOLINK DOES NOT INFRINGE THE '321 PATENT Given the undisputed operation of AutoLink, Google does not infringe claims 1, 24, and

86 of the '321 patent. Google will first analyze each limitation under HyperPhrase's old infringement theories, then under its new infringement theories. A. AutoLink does not satisfy the "when" or "real-time" limitation of claims 1 and 24.

Google's argument regarding the "real-time" limitation is essentially the same as that already decided by this Court in the earlier Microsoft litigation with HyperPhrase. For the same reasons articulated by this Court in its summary judgment opinion in that litigation, Google does not infringe at least claims 1 and 24 of the '321 patent because the claim precludes manual user intervention while AutoLink requires it. (Woodford Decl. (Dkt. No . 35) Ex. W (Judge Crabb Order re MSJ) at 24-27.) 1. AutoLink does not satisfy the "when" or "real-time" limitation of claim 1 under HyperPhrase's old infringement theory.

HyperPhrase essentially concedes that the term "when" adds a real-time limitation to claim 1. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 44.) Furthermore, HyperPhrase does not dispute that AutoLink does not create the second URL until after a user has manually selected the first URL. (Id. at 9 ("After the user clicks on the light blue link, the Google `tbproxy' server receives that URL [read: the first URL]... Once received... the redirect causes a server to build[] what is known as a `redirect URL,' [read: the second URL] (which is different from a redirect request URL [read: the first URL])." (emphasis added); Google Reply to Its PFOF No. 57 (undisputed).) The issue is application of this claim limitation to AutoLink, and particularly claim element (ii)(c), which reads in relevant part: (b) analyzing the referencing record in accordance with the MRRS to identify the existence of the MR and, when the MR is identified; (c) identifying the referenced record associated with the DR/MR combination. We note that this Court construed the phrase "identifying the referenced record" to mean "locating or finding the record." (Woodford Decl. (Dkt. No. 35) Exhibit S (6/18/03 claim 19

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 24 of 34

construction order) at 15.) In HyperPhrase's original infringement analysis, which has no meaningful discussion of the "when" limitation (Dr. Thompson ignored it), the second URL (generated by the AutoLink server) allegedly satisfied the "identifying the referenced record" limitation:

(Third Thompson Decl. (Dkt. No. 129-4) at 28 of 37.) Since the second URL only gets built if and after a user selects a hyperlink corresponding to the AutoLink token (which in turn occurs only after the user selects the AutoLink button on the toolbar), Google moved for summary judgment that HyperPhrase could not prove that AutoLink satisfied the "when" or real-time limitation. Having been put to the task of explaining its position on facts that are not in dispute, HyperPhrase has abandoned its prior infringement theories and changes horses again. Thus, there is no longer any dispute that claim 1 of the '321 patent is not infringed by AutoLink under HyperPhrase's old infringement theory. 2. AutoLink does not satisfy the "when" or "real-time" limitation of claim 1 under HyperPhrase's new infringement theory.

HyperPhrase's new infringement theory is the first URL, which points to the AutoLink server, identifies the referenced record, even though that alleged referenced record is not stored on the AutoLink server. (HyperPhrase Noninf. Opp. (Dkt. No. 127) at 44.) This theory turns not on evidence but on the inconsistent speculation of Dr. Thompson. Dr. Thompson says:

(Fourth Thompson Decl. (Dkt. No. 129) at 20.) In considering this passage, the Court should note first that the Google argument to which Dr. Thompson refers was addressing Dr. 20

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 25 of 34

Thompson's earlier infringement analysis, which HyperPhrase has abandoned, so his disagreement is moot. (Supra.) Second, to the extent Dr. Thompson proceeds to speculate how AutoLink "must" work, there is no citation to any evidence to support his surmise. In fact, reading on a few lines, it appears that Dr. Thompson is not even sure of his opinion:

(Id.) HyperPhrase must do more to prove infringement ­ and to establish that a genuine issue of material fact exists ­ than to have its expert speculate that the claims might be infringed. The undisputed facts dispose of all HyperPhrase's theories, new and old. There is no dispute that the first URL is different than the second URL. (Google's Reply to Its PFOF Nos. 54 and 58; Djabarov Decl. (Dkt. No. 33) ¶¶ 12 and 14; Fourth Thompson Decl. (Dkt. No. 129) at ¶¶ 12 and 19.) There is no dispute that the first URL points to a process on the AutoLink server. (Google Reply to Its PFOF Nos. 54 and 57; Fourth Thompson Decl. (Dkt. No. 129) at ¶ 13.) HyperPhrase offers no evidence or even argument to explain how, for instance, the first URL identifies the alleged referenced record on a third party server (or even the Google Maps server) without relying on what happens at the AutoLink server after the user has manually intervened by selecting the AutoLink first URL. Moreover, HyperPhrase's new theory completely ignores this Court's construction of the term "identifying the referenced record" in the Microsoft litigation, which was "locating or finding the record." (Woodford Decl. (Dkt. No. 35) Exhibit S (6/18/03 claim construction order at 15.) Since nothing happens with the first URL until a user manually intervenes and causes the first URL to be sent to the AutoLink server (where it is further processed), AutoLink fails to satisfy the real-time requirement of claim 1 and there is no 21

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 26 of 34

infringement. (Google Reply to Its PFOF No. 58.) Because claim 1 is not infringed, claim 24, which depends from claim 1, is not infringed. B. AutoLink does not satisfy the "modifier reference" limitation of the claims. 1. AutoLink does not satisfy the "modifier reference" limitation of claim 1 under the correct construction of this term.

HyperPhrase offers its fourth construction (in this litigation) of the term "modifier reference" in its opposition brief, agreeing in many respects with Google's construction, but grafting additional clauses to the phrase to (1) broaden aspects of what may be "specified" due to noninfringement problems, and (2) narrow other aspects of the phrase to avoid the substantial prior art Google has advanced in its motion for summary judgment of invalidity. The parties' constructions of the term "modifier reference" are shown in the table below. The common thread between these constructions is that, until HyperPhrase had to re-tool its infringement case based on Google's summary judgment motions, HyperPhrase agreed that the modifier reference pointed to a subset of what the data reference pointed to ­ namely, a particular record or record segment.

22

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 27 of 34

Google's Construction
A word or phrase that further specifies a specific record or record segment.

HyperPhrase's Various Constructions
September 24, 2003: A reference which, in combination with a DR, may be used to refer to a specific and distinct record or record segment.13 October 2, 2006: Additional information identifying or relating to more specific or distinct records when received with a DR.14 November 21, 2006: A modifier term which may be used in combination with a more significant term anywhere in a document (the DR) to resolve ambiguities over the context of the more significant term.15 March 11, 2008: Record text (not programming codes), as defined in col. 3, l. 22-24 of the '321 patent, which states that "... modifier references (MRs) may be used to further specify a specific record or record segment when a DR is identified.16 April 11, 2008: A word or phrase that further specifies a specific record, record segment, or records referred to by a data reference; the modifier reference is the text in a record normally displayed for a typical user to read and not hidden computer codes.17

Google's construction follows the disclosure in the written description, which defines a modifier reference as a word or phrase that further specifies a particular record or record segment identified by a data reference. In other words, the modifier reference is used to pinpoint a subset of what was already referred to by the data reference ­ a specific record or record segment. (See '321 patent at 3:21-32.) For instance, if just the term "ECG" were identified as the data reference, it might be unclear which ECG were meant to be referred to by the record author, which is why the modifier reference "admission" narrows the term "ECG" to identify not one of many ECG records, but a specific record or record segment such as "admission ECG". (See '321 patent at 11:65-12:8.) This conclusion is also supported by the repeated reference in the written description to the importance of linking from one record to another specific record. (See '321

13

Woodford Decl. (Dkt. No. 35) Ex. V (HyperPhrase's Supp. Br. on Claim Construction) at 22. Offered in the HyperPhrase v. Microsoft litigation. 14 Woodford Decl. (Dkt. No. 35) Ex. X (HyperPhrase's 10/2/06 Resp. to Google's Interrogatories) at 12.). Offered before Google filed its 2006 motion for summary judgment. 15 HyperPhrase Opp. to Google's Mtn for SJ of AdSense Noninfringement (Dkt. No. 36) at 23. Offered after Google filed its 2006 motions for summary judgment. 16 de la Huerga Rebuttal Invalidity Rpt. (Dkt. No. 130-2) at 193 of 307. Offered after Google filed its 2008 motions for summary judgment but before HyperPhrase filed its opposition brief. 17 Niro Decl. Ex. R (Dkt. No. 125-19) (HyperPhrase Proposed Claim Const.) at 18. Offered after Google filed its 2008 motions for summary judgment.

23

Case: 3:06-cv-00199-bbc

Document #: 136

Filed: 05/19/2008

Page 28 of 34

patent at 2:22-25, 2:50-62, 3:8-10, 3:13-17, 3:30-32, 4:24-27, 5:5-8, and 5:47-50.) The inventor of the '321 patent, Carlos de la Huerga, in his rebuttal expert report on validity, concurs with Google's definition, himself noting that the patent he drafted defined the very term "modifier reference," which is now disputed by his lawyers and infringement expert Dr. Thompson:

(de la Huerga Rebuttal Invalidity Rpt. (Dkt. No. 130-2) at 193 of 307.) Under Google's (and Mr. de la Huerga's) construction, AutoLink does not infringe because the alleged modifier reference, the AutoLink trigger, does not further specify a particular record or record segment identified by a data reference. For instance, when an AutoLink token (the alleged data reference) is found, it is assumed to be a specific token, meaning, for instance, it is an ISBN, VIN, or shipping tracking number. (Google Reply to Its PFOF No. 52.) The AutoLink tokens, as Judge Shabaz found in his earlier summary judgment decision in this case, "are not references to a particular data record but are references to things about which there may be many data records. A street address, for example, is not a reference to a data record at all, but a reference to a place... The same is true of the other tokens which, rather than identifying a particular data record as in the patent claims, identify a car, package, or book about which there may be number data records." (SJ Order (Dkt. No. 64) at 19.) The AutoLink trigger is used to throw out false positives, meaning tokens that might not really be what AutoLink already thinks they are. (Google Reply to Its PFOF No. 53.) The trigger does not make the thing referred to by the token any more specific than it already was. Even the alleged combination of a token and a trigger do not refer to a specific record or record segment, and HyperPhrase offers no evidence to the contrary.18 Thus, under Google's construction, AutoLink does not satisfy the modifier limitation.
18

The Federal Circuit found, sua sponte, that a token was a data reference. While Google objected to this finding in its petition for rehearing, it notes that the reason Judge Shabaz's conclusion that a token was not a data reference was reversed is because the Federal Circuit believed Judge Shabaz found the term data reference required pointing to "one and only