Valueclick Inc v. Revenue Science Inc

Filing 37

DECLARATION of KENDYL ROMAN IN SUPPORT OF DEFENDANT REVENUE SCIENCE, INC.S RESPONSIVE MARKMAN BRIEF filed by Counter Claimant Revenue Science Inc, Defendant Revenue Science Inc. (Attachments: # 1 Exhibit 1# 2 Exhibit 2# 3 Exhibit 3)(Subhedar, Nitin)

Valueclick Inc v. Revenue Science Inc

Doc. 37

1 ROBERT T. HASLAM (State Bar No. 71134) 2 NITIN SUBHEDAR (State Bar No. 171802) 3 BHANU K. SADASIVAN (STATE BAR NO. 233429) 4 HELLER EHRMAN LLP

[e-mail]

[e-mail]

[e-mail] 275 Middlefield Road

5 Menlo Park, CA 94025-3506 6 Facsimile: [phone]

Telephone: [phone]

7 JOHN C. ULIN (State Bar No. 165524) 8 HELLER EHRMAN LLP th 10 Facsimile: [phone]

[e-mail]

9 Los Angeles, CA 90071-3043

333 South Hope Street, 39 Floor Telephone: [phone]

11 MATTHEW C. LAPPLE (State Bar No. 193546)
th 12 4350 La Jolla Village Drive, 7 Floor

[e-mail]

13 Telephone: [phone]

San Diego, CA 92122

Facsimile: [phone] Attorneys for Defendant

14

15 REVENUE SCIENCE, INC. 16 17 18 19 20 VALUECLICK, INC., a Delaware Corporation, 21 22

UNITED STATES DISTRICT COURT CENTRAL DISTRICT OF CALIFORNIA WESTERN DIVISION Case No.: CV 07 2052 MMM (JCX) Plaintiff, v. DECLARATION OF KENDYL ROMAN IN SUPPORT OF DEFENDANT REVENUE SCIENCE, INC.'S RESPONSIVE MARKMAN BRIEF

23 REVENUE SCIENCE, INC., 24 a Washington Corporation, 25 26 27 28

Defendant.

Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

Dockets.Justia.com

1 I, Kendyl Roman, declare as follows: 2 3

I.

INTRODUCTION 1. I am the same Kendyl A. Román who submitted a declaration entitled

4 "Declaration of Kendyl Román in Support of Defendant's Opening Markmark Brief," 5 dated October 14, 2007, ("Opening Declaration"). My qualifications, materials 6 reviewed, technical background, methodology, understanding of one of ordinary skill 7 in the art, and claim term interpretation, as found in my Opening Declaration are 8 included herein by reference. Attached as Exhibit 1 is a list of additional material I 9 considered. 10

2.

I have been asked to provide expert testimony with respect to U.S.

11 Patent Number 5,848,396 ("the `396 Patent") 1 and U.S. Patent Number 5,991,735 12 ("the `735 Patent"), 2 (collectively the "ValueClick Patents"). 13 II. 14

ONE OF ORDINARY SKILL IN THE ART 3. As I describe in my Opening Declaration, a person of ordinary skill in

15 the art is "an individual with a variety programming techniques and languages 16 including object oriented programming and Internet client-server programming 17 experience obtained through a bachelor degree in computer science or equivalent 18 work experience and education. Additionally, the person of ordinary skill would 19 have experience applying the above-listed skills to advertising and market research 20 including user profiling." 21

4.

I disagree with ValueClick's expert, Peter Kent's, description of a

22 person of ordinary skill in the art as "someone with a bachelor's degree in a computer 23 science field, or equivalent, plus 2-3 years of experience in computer networking and 24 related system design and programming, with a basic awareness of HTML 25

U.S. Patent No. 5,848,396, attached as Exhibit 1 to the Declaration of Bhanu K. Sadasivan In Support Of RSI's Opening Markman Brief. 27 2 U.S. Patent No. 5,991,735, attached as Exhibit 3 to the Declaration of Bhanu 28 K. Sadasivan In Support Of RSI's Opening Markman Brief.
26
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

1

2

1 programming." (emphasis added) 2

5.

In my opinion, understanding the only embodiment disclosed by the

3 ValueClick patents requires a fairly high level of expertise with object oriented 4 programming (and at least Java). Mr. Kent, however, omits object oriented 5 programming from the definition of one of ordinary skill in the art. 6

6.

I disagree with Mr. Kent's conclusion that the field of invention would

7 be understood to be "Online Marketing System Design." 8

7.

For the reasons set forth below, I believe the field of invention would be

9 "computer programming of Internet server software using a networked client-server 10 architecture," and in particular "programming a method of use of a computerized 11 device, for performing a business process where information is polled (i.e. obtained 12 by query and response) over a computer network and systematically analyzed to 13 forecast demand (or in this case, interest)." 14

8.

The ValueClick Patents omit a "Field of Invention" section commonly

15 found in patents. 16

9.

Due to the omission of the field of invention, I look to the U.S. Class,

17 which is 705/10, and to the claims to determine the art field of the invention. 18

10.

The class 705 is "Data Processing." Subclass 10 is indented under

19 subclass 7, which is indented under subclass 1. Subclass 1 refers to "a device or to a 20 method of use of a device for performing" a business practice process, including in a 21 "network." Subclass 7 brings in the concepts of "computerized" and 22 "systematic...analysis" to which subclass 10 adds the concept where the 23 computerized device is "polling" in order to obtain market data in order to forecast 24 demand. 25

11.

Neither the phrase "online marketing" nor "system design" appears in

26 the specification. The term "marketing" only appears once in the claims and once in 27 the specification. In contrast the word "object" appears over 165 times. Java (an 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

3

1 object oriented programming language) is mentioned more in the written description 2 than "marketing". 3

12.

Value Click has put forth Peter Kent as a proposed expert in the field of

4 "Online Marketing System Design." In ¶4. 5

13.

In my opinion, Mr. Kent has made several erroneous statements in his

6 declaration. 7

14.

First, Mr. Kent refers to "HTML programming" in his definition of one

8 of ordinary skill in the art. HTML is a markup language which is used to mark up 9 and format text documents. HTML is not a programming language. HTML is not an 10 object oriented programming language. It would be more appropriate to describe 11 HTML as document "formatting" not programming. 12

15.

Second, Mr. Kent's statement that "Programming techniques to

13 accomplish these algorithms would have been trivial" is inaccurate in light of the fact 14 that the functions are quite complex and the patent does not provide any flow chart or 15 source code describing the algorithms. 16

16.

Third, in the context of the patent, Mr. Kent's equating the data

17 assembly of the objects disclosed in Fig. 3A and Figs. 5A through 5D to a relational 18 database or a "flat-file" text database (Kent ¶11, p. 141 line 11-13) is inaccurate. 19

17.

In object oriented programming, data elements (known as variables or

3 20 fields) are encapsulated into modules called objects along with the routines or

21 algorithms (known as methods) that act on the encapsulated data. As summarized by 22 the ValueClick Patents, "Each object is formed of data [variables or fields] and 23 subroutines (methods) for acting on the data." (`396, 5:42-43). An object oriented 24 program is comprised of multiple objects. 25 26

18.
3

Similar objects that share the same features (i.e. data fields and methods)

Different texts and different languages use different terms to refer to the data member of an object. Apple Media Language used the term "field," other references 28 used the terms "instance variable" and "data member."
27
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

4

1 are defined by an "object definition" (also known as a class definition in some object 2 oriented languages). Thus, an object definition (also known as object "type") is a 3 template used to "instantiate" multiple copies of similar objects that share the same 4 features. An object definition is like a cookie cutter. In making holiday cookies, a 5 family may have a variety of cookie cutters: one shaped like a pumpkin, one like a 6 star, one like a tree. Although the same cookie cutter (template) is used, each 7 instance of tree cookie can be filled in "populated" with different specifics. This is 8 analogous to different colors of frosting and different types of sprinkles or candy 9 being used to decorate the cookies. A plate full of tree cookies all share the same 10 basic attributes (e.g. shape, size, number of branches, etc.) Likewise, a plate full of 11 star cookies shares the same basic attributes (e.g. shape, size, number of points). 12

19.

When the ValueClick Patents talk about "respective tables of objects"

13 (`396 5:46) it is not suggesting conventional relational database management system
4 14 with relational tables full of tuples, or a "flat-file" database, rather it is describing

15 what is analogous to respective plates of cookies, i.e. a collection of object instances 16 that are all based on the same object definition (template). 17 III. 18

THE VALUECLICK PATENTS 20. The ValueClick Patents describe using agate information to determine

19 the psychographic profile of a computer user and providing customized pages to a 20 targeted user. 21

21.

The only disclosed embodiment is an object oriented program 31

22 running on a single Internet server 27 (Fig. 1 and Fig 2.) which program is 23 implemented as various objects (Fig. 3A). 24

22.

The program 31 generally comprises a program controller 79 that runs

25 on the server 27 having three components: agate data assembly 71, user profile 26 member 73 (also known as "user tracking and profiling member 73"), and ad module 27

In math and relational database theory, a set of ordered elements is referred to 28 as a tuple.
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

4

5

1 75. 2

23.

In Fig. 3A, more detail is shown regarding program 31. Program has

3 two routines, represented by rectangles, i.e. main routine 39 and a reporting 4 subroutine 41. All of the objects, represented by ovals, are shown associated with the 5 main routine 39. The user tracking and profiling member (73) is shows as a set of 6 user objects 37 (37a-37f). The agate data assembly (71) is shown as a set of page 7 objects 35 (35a-35c). The ad module (75) is shown as a set of Sponsor (or Ad) 8 Objects 33 (33a-33d). 9 IV. 10

ADDITIONAL DISCUSSION OF CERTAIN TERMS 24. This section will provide further discussion of certain terms in addition

11 to what I provided in my Opening Declaration. 12 13

A. 25.

Computer Apparatus I discussed the term "computer apparatus" in my Opening Declaration

14 (see ¶31-36) 15

26.

In 1996, because the computers were still relatively slow compared to

16 the demand placed on a particular Web site, the load of processing the HTTP requests 17 were distributed to different CPUs by clustering in one of two ways: a) a cluster of 18 multiple CPU in one computer server, and/or b) a tightly coupled group of "similar" 19 (i.e. cloned) servers all running identical software and processing requests delegated 20 by a load balancer which had a single network address. However, in both these 21 cluster structures for load balancing, all the requests were directed to a single 22 network address in a single physical location. This is what meant by "server 27" 23 being comprised of a) a DEC AlphaServer cluster or b) a "multiplicity of similar 24 such servers". ('396, 3:63-64) Exhibit 2 is a PR Newswire document5 that discusses 25 DEC's cluster solutions, referring to both OpenVMS clusters and the newly 26

Digital Unix TruCluster Solutions Shatter the Barriers of Performance, Availability and Affordability for Enterprise Computing, PR Newswire, April 17, 28 1996.
27
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

5

6

1 announced UNIX cluster solutions. One of ordinary skill in the art would have 2 understood that DEC's "OpenVMS" Clustering and "Unix TruCluster Solutions"
6 3 provided hardware and software that allowed "multiple AphaServer" nodes to be

4 rack mounted in a cluster as a "single computing resource." In a DEC cluster, the 5 CPUs are tightly coupled as a "single computing resource" (e.g. server 27) as 6 opposed to the "loosely coupled" digital processors 11, 13, 15, 17 and servers 21a, 7 21b, 21c. As the patent states: 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Each network 19 includes a multiplicity of digital processors 11, 13, 15, 17 (e.g., PC's, mini computers and the like) loosely coupled to a host processor or server 21a, 21b, 21c for communication among the processors within that network 19. ('396, 3:41-45) That is, each of the networks 19 are themselves loosely coupled along a communication line 23 to enable access from a digital processor 11, 13, 15, 17 of one network 19 to a digital processor 11, 13, 15, 17 of another network 19. ('396, 3:49-53) For example, the "Distributed Available Disk... allows directly connected, standard SCSI storage devices to be accessed by all nodes in the cluster" (i.e. the cluster uses the same hard disk drives) and "High-speed Memory ChannelTM interconnect ... provides a direct memory-to-memory connection between AlphaServer" cluster nodes (i.e. the CPUs in the cluster use the same memory).7 This allows the tightly coupled CPUs in the cluster to run the exact same software program. 27. The patent specification further supports such a reading. For example,

the written description says the program has a single network address, which is both

the Website address of the software program 31 and the server 27 hardware, stating: 22 "The user logs onto the Internet 29 and enters the URL or Website address of program 31 which initializes main routine 39. The URL 23 request is received by Web server 27 ..." ('396, 13:38-42)
24 25

DEC's clustering hardware and software included for example a "Distributed 26 Available Disk" and a "High-speed Memory ChannelTM interconnect" which allowed two to eight AlphaServer nodes to be interconnected, as well as "Netscape-based" 27 Web server software. See Exhibit 2. 7 28 Ibid.
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

6

7

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

B. 28. (see ¶37-39) 29.

A Computer Program Embodied On A Computer Readable Medium I discussed the term "computer program..." in my Opening Declaration

I disagree with ValueClick's statement that "object-oriented programs

(like the ones described in the Gerace Patents) were known to be an aggregation of multiple self-contained subprograms." (ValueClick Brief, pl. 21, lines 14-17). The objects in an object oriented program are not "self-contained subprograms," rather they are all compiled together into a single, interdependent computer program with a single "main" routine. C. 30. 31. User I discussed the term "user" in my Opening Declaration (see ¶40-44). The ValueClick patents teach that the purpose of the User Objects 37

(including User 37a and User Interface Profile 37c) is to identify users (37a) and to

maintain a user profile (37c) for each individual user: "Turning to FIG. 3a, the purpose of the set of User Objects 37 is to 16 identify users and maintain a user profile for each user" ('396, 5:62-64) 17
18 19 20 21 22 23 24 25 26

"User Object 37a identifies a respective user..." ('396, 6:2-5) "The User Interface [Profile] Object 37c provides a unique ... identifier of the user. The User Interface Object 37c also provides indications of categories of interest to the user and a primary screen display for each category customized to that user" (`396 6:22-26) (emphasis added) 32. After the user logs in, the program also "assigns a user name and

password at the user's convenience." (`396, 5:5-7). Thus, contrary to ValueClick's statement that "logging is optional. See 5:5-7 ("at the user's convenience")," the

phrase "at the user's convenience" does not refer to optional logging: "Program controller 79 also obtains user identification information 27 from the user to assign a user name and password at the user's convenience." (emphasis added). 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

8

1 2 3 4 5 6 7 8

"For a new user, the Home Page 43 effectively requests a user name and password. In response to the user-provided data, main routine 39 immediately builds a cookie if possible. Included in the newly built cookie is a unique user identification code (preferably numeric), time and date of login, and computer identification number to distinguish between home and work logins." (`396, 13:6214:1)(emphasis added). "The User Interface Object 37c provides a unique (preferably numeric) identifier of the user." (`396, 6:22-23). 33. The user provided data is stored in addition to the unique identification

9 code (generated internally by the program controller). Thus, the login is required to 10 enable the "unique user identification code" which is a numeric "unique ... identifier 11 of the user". In other words the claimed invention will not work without being able 12 to uniquely identify the set of User Objects 37 which correspond to the respective 13 user. 14

34.

Further, the unique identifier associated with the User Interface Profile

15 37c object is used to customize the content and format for uniquely identified 16 individual user. 17 18 19 20 21 22 23

"...the user's User Interface [Profile] Object 37c holds indications of his categories of interest, including specific items of interest in each category of information, and his display/format preferences (colors, design, layout, etc.). Based on these recorded details, program 31 constantly and automatically tailors screen views (content and presentation) and advertisement selection (subject matter and presentation) for the user." (`396, 17:1-9) (emphasis added). 35. 36. Kent's opinions regarding "user" are provided at Kent ¶10. I disagree. That the user is an individual user is supported by the written description

24 which shows that the invention provides advertiser (sponsor) reports which "Allows 25 drill-down through to individual user level" (`396, 33:43) 26

37.

Further, the intrinsic record shows that Wilkin also defined user "...psychographic information about each audience member...The
9

27 psychographic information to an individual member level: 28

Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

channel selector/decoder (100) unit associated with each member's receiver compares the selection profile with the demographic/psychographic information about the audience member and selects the appropriate media message for that audience member." (Wilkin `918 Abstract; see also 8:10-11 and 8:33-34) (emphasis added) 38. ValueClick is technically wrong when they state that "...the preferred embodiment itself uses the computer's ID as one way to identify a user." (ValueClick Brief p. 22, lines 10-15) A proper reading of the preferred embodiment shows that each user has both "a User Computer Object 37b and a User Interface Object 37c." (`396 6:13-14). Further, the cookie contains multiple pieces of data including both a unique user identification code, and a computer identification number. "For a new user, the Home Page 43 effectively requests a user name and password. In response to the user-provided data, main routine 39 immediately builds a cookie if possible. Included in the newly built cookie is a unique user identification code (preferably numeric), time and date of login, and computer identification number to distinguish between home and work logins." (`396, 13:6214:1)(emphasis added). 39. Thus the unique user identification code identifies the user, not the

computer ID number that instead identifies the computer (e.g. whether it be the home or work computer for the same unique individual user). D. 40. Psychographic Profile Of A (Each) User I discussed the term "psychographic profile of a (each) user" in my

21 Opening Declaration. (see ¶45-50) 22 41. I disagree with Mr. Kent's statement "There is no disclosure that such 23 profiles are built by prompting the user..." Kent ¶6. For example, ValueClick 24 admits in their own brief that "portfolio" and "last specified city" are associated with 25 the "psychographic profile" (ValueClick Brief p. 5, lines 1-3, citing `396 16:19-29). 26 The user "specified city" is indicative of a user prompt. 27 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

42.

Further, the written description teaches that for example menu selections
10

1 are used to build the psychographic profile. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

"..., program controller 79 responds to user selections and viewing actions (screen formatting commands/requests, menu selections, etc.) ... using the user profiling member 73 to record the user's activities and thus build a psychographic/behavioral profile of the user." (`396, 5:8-14; `735, 5:36-42) (emphasis added). 43. The specification of the `396 and `735 Patents supports that the

psychographic profile is "based on that person's responses to prompts regarding that person's preferences and lifestyle." For example, "Main routine 39 prompts the user for his zip code or the name of the city for which he wants weather information." (`396, 16:2-4) (emphasis added). "The screen view also prompts the user to a directory of symbols for use as needed. Near the lower portion of the screen view, there is displayed an area for the user to enter a new stock symbol and an option "button" to effect addition of the corresponding company to the user's portfolio." (`735, 14:62-66) (emphasis added). "An alternative is displaying in Financial Pages several blanks in which users can place company symbols, with check boxes for the options of `Add these to my Portfolio' or `Add these to my `Follow these` list'" ('396, 20:66-5) (emphasis added) 44. The term "user profile" would be understood by those skilled in the art

18 at the time of filing of the patents to include both the psychographic profile and the 19 demographic profile. 20

45.

In particular, `396 claim 1 "generates a psychographic profile of a user"

21 which contains "categories of interest" and "display format preferences for each 22 category". Claim 5 then adds a distinct "target profile of desired users" which is part 23 of the advertising component (detailed as Sponsor/Advertiser Objects 33, distinct 24 from the User Objects 37). This is not inconsistent with claim 10 where "each user 25 profile provid[es] an indication of categories of interest to the user and display 26 preferences for each category." 27

46.

Further, the intrinsic record shows that Wilkin also defined user
11

28 psychographic information to an individual member (see Wilkin quotes regarding
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

1 "each audience member" above). 2

47.

In the extrinsic dictionary documents offered by ValueClick, there were

3 multiple definitions, including multiple definitions directly related computers. 4 However, ValueClick choose a definition that was not computer specific, and thus 5 was the least relevant. For example, ValueClick Exhibit G page 5 of 6 (page 103) 6 contains two computer related definitions from the Free Online Dictionary of 7 Computing and the Jargon File where "profile" as it relates to computers is a 8 "control file for a program" for "each user." Both sources also point to related "dot 9 file" and "rc file" definitions. One of ordinary skill in the art would have understood 10 a profile of computer user to be, for example, a "dot file" such as "~/.profile" which 11 is "personal initialization file" (see, for example, the UNIX online documentation 12 regarding the Bourne shell ("sh") including ".profile"). "rc files" were also known to 13 contain a user's personal display/format preferences on a per user basis. 14

48.

ValueClick's selected extrinsic evidence, and expert testimony, is

15 inconsistent with the "otherwise unambiguous meaning of a claim term." For 16 example, ValueClick Exhibit E, Wikipedia "psychographic" break profiles down into 17 three (3) not two (2) groups (namely, behavior, psychographic, and demographic). 18 However, in 1996, and in the context of the ValueClick Patents, as admitted by 19 ValueClick, there are only two groups (see ValueClick Brief p. 14 lines 21-22). 20

49.

Another misleading faux quote is "target psychographic profile" which

21 does not exist in the Value Click Patents (ValueClick Brief p. 15 line 4) 22

50.

ValueClick skips over contradictory information regarding preferences

23 in `396, 6:23-34 and 12:22-41 (ValueClick Brief, p. 2 lines 12-17). 24 25 26 27 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

12

1

51.

In another case, ValueClick skips over contradictory information

2 without even providing an elipse "..." after the word "views" in the excerpt from 3 `396 4:18-21 (line 22 ignored) (ValueClick Brief, p. 3 line 18). 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

52.

Further as stated above, the purpose of the User Objects 37 (including

User Interface Profile 37c) is to maintain a user profile (37c) for each individual user indicating that users display preferences for each category of interest, and then to customize (tailor) the display screen view (page) based on those psychographic

profile format preferences. "Turning to FIG. 3a, the purpose of the set of User Objects 37 is to identify users and maintain a user profile for each user" ('396, 21 5:62-64) (emphasis added)
22 23 24 25 26 27 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

"The User Interface [Profile] Object 37c provides a unique ... identifier of the user. The User Interface Object 37c also provides indications of categories of interest to the user and a primary screen display for each category customized to that user" (`396 6:22-26) (emphasis added) "... the user's User Interface [Profile] Object 37c holds indications of his categories of interest, ..., and his display/format preferences (colors, design, layout, etc.). Based on these recorded details,
13

1 2 3

program 31 constantly and automatically tailors screen views (content and presentation)..." (`396, 17:1-9) (emphasis added). 53. My opinion differs from ValueClick's interpretation in another aspect

4 related to preferences. In contrast to ValueClick's interpretation, a person of ordinary 5 skill in the art reading the patent would understand that the terms "display 6 characteristics" and "graphic references" listed in Fig. 5D relate to the "daily start 7 time" and "daily end time" for displaying the ad, and to the graphical or multimedia 8 portions of the advertisement: 9 10 11 12 13 14 15 16 17 18 19 20 21

"Ad Object 33d also includes the starting and ending time for display of the ad each day. Ad Object 33d also provides references to graphic, sound, and multimedia portions of an advertisement." (`396 12:48-52) (emphasis added) E. 54. 56). 55. browser. ". . . the term `page' is used synonymously with screen view." (`396, 20:39-40) (emphasis added). "Preferably the Home Page 43 (FIG. 4a) is an HTML (HyperText) document generated through the set of Page Objects 35a,b,c. ... The Home Page 43 is formed of several graphical and text documents in the HTML and Java formats." (`735, 14:3-9) (emphasis added). 56. Kent's opinions regarding "screen view" are provided at Kent ¶7. I The term "screen view" means the "entire page" for display in the Screen View I discussed the term "screen view" in my Opening Declaration (see ¶54-

22 disagree. 23

57.

ValueClick's proposed construction which appears to be loosely based

24 on Kent's statement, "a page of content or some other form or amount of content" is 25 clearly wrong in that it contradicts the only disclosed embodiment, which always 26 references the screen view as the whole page, or the outermost HTML document, as 27 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

14

8 9 1 opposed to Java window or banner advertisement which could also be within the

2 page. (A Java window is contained inside a screen view (web page). See also Fig. 3 4A and Appendices I and II). 4

58.

Further, PointCast (a cited prior art example) would have been known to

5 one of skill in the art. It provided a view which filled exactly the entire screen of a 6 PC running Windows in 1996. Thus, Kent's comments regarding the "Windows 7 taskbar" are not accurate. 8 9

F. 59.

Physical Activity I discussed the term "physical activity" in my Opening Declaration (see

10 ¶57-59). 11

60.

I disagree with Mr. Kent's articulation regarding "physical activity."

12 Kent ¶12. The term "physical activity" must be defined as understood by a person 13 of ordinary skill in the art at the time of filing of the patents, not a layperson. 14 Ordinary activities that a layperson today would consider a "physical activity" like 15 dancing, boxing, or shooting a gun, would not have been understood by one of 16 ordinary skill in the art in 1996 to be the type of computer inputs to which the
10 17 ValueClick Patents limit the term.

18

61.

Further ValueClick's proposed phrase "which can (after processing)

19 indicate what has been viewed and/or what has been clicked" is not accurate for this 20 term. First, the ValueClick Patents explicitly discuss line mode web browsers (e.g. 21 "lynx" and "www") which did not even return mouse click information, thus the 22 "what has been clicked" clause is not consistent with, and would exclude, the 23 24 25 26 27 28

See "Format 2. News Page (windows will be Java scrolling including new news where possible)" (`396, 25:57-58) See also "Other news windows" (`396, 25:63) 9 See `396, 23:18-65, in particular "Top banner" line 21 and "Bottom banner" line 65. 10 In contrast, today each of these is a computer input, for example, to home computer games.
15

8

Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

1 preferred embodiment. Second, the keyboard typing (or alternatively mouse cursor 2 locations) that would have been received by the server 27 alone would not be able to 3 determine "what has been viewed". Only a program controller 79, running on a 4 single server 27, which had concurrent information about what screen view (Page 5 Display 35c object) was displayed and the layout of the "Page Data" 35b objects 6 within the entire page (as defined by the "Page Display" 35c object) would be able to 7 determine what item was selected based on keyboard or mouse data received by the 8 program controller 79 at the time the HTTP request (for the next page) is received. 9 (See Figs. 2, 3A, 3F, and 3G). 10 11

G. 62.

User Response I discussed the term "user response" in my Opening Declaration (see

12 ¶60-61). 13

63.

Kent's opinions regarding "user response" are provided at Kent ¶13. I

14 disagree. The term "user response" has a special meaning as defined by the 15 ValueClick Patents. See my discussion regarding "physical activity" above. Further, 16 I conclude that the two distinct claim terms have distinct meaning (see my Opening 17 Declaration). 18 19 20 21 22 23 24 25 26 27 28

H. 64.

Target Profile Of Desired Users To Whom To Display The Advertisement/Target Audience Profile Of Each Advertisement I discussed the term terms "target profile of desired users to whom to

display the advertisement" and "target audience profile of each advertisement" in my Opening Declaration (see ¶62-64). 65. Kent's opinions regarding "target profile" are provided at Kent ¶15. I

disagree. The term "target profile" has a special meaning as defined by the ValueClick patents. 66. ValueClick is logically wrong in their implication about "psychographic

profile and demographic profile" matching the "target profile" (ValueClick Brief p. 25, lines 22-27). Logically, for example, any round peg can match a round hole, but
16

Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

1 that does not mean that a round peg is the same or equal to a round hole. It only 2 means that a round peg (analogous to a "psychographic profile") can be matched 3 based on some subset of similar criteria to a distinct item, namely, the round hole 4 (analogous to a "target profile"). Any number of individual round pegs (individual 5 "psychographic profiles") would match a given round hole (an intended "target 6 profile"). In this analogy the required characteristic defined by the round hole is the 7 shape and size of the peg but not the color. Both a red round peg and a blue round 8 peg would match the target round hole. 9

67.

Further ValueClick's pseudo quote "target psychographic profile" does

10 not exist in the ValueClick Patents. (ValueClick Brief, p. 15 line 4) 11

68.

Finally, ValueClick admits that ads are selected to show to "a user" (i.e.

12 on an individual user basis, not a "group" of users). (ValueClick Brief p. 5, lines 1713 22) 14 15

I. 69.

Advertisement I discussed the term "advertisement" in my Opening Declaration (see

16 ¶65-67). 17

70.

In summary, sponsored advertisements are in the domain of the

18 advertisement module 75 as detailed in the Sponsor/Advertising Objects 33 with each 19 ad being represented by an instance of the Ad Object 33d (See `396, 4:43-46 and 20 `396, 12:42-43, respectively). On the other hand, agate data (discussed above) 21 contains un-sponsored advertisements (e.g. classifieds) that are in a distinct domain, 22 namely, the agate data assembly 71 as imprecisely detailed in the Pages Object 35. 23

71.

Kent's opinions regarding "advertisement" are provided at Kent ¶16. I

24 disagree. As discussed above, the term "advertisement" has a special meaning as 25 used in the claims and as defined by the ValueClick patents. 26

72.

The specification uses the term "sponsorship ad" when referring to

27 banner advertisements that "are positioned along the periphery (i.e., above, below, 28 left or right) of the agate data" ('396, 7:31-34) (emphasis added):
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

17

1 2 3 4 5 6 7 8 9 10 11 12

In each of the foregoing formats, the preferred embodiment includes incorporation of ads or sponsorship indications as top and/or closing banners. ('396, 8:13-15) (emphasis added) Incorporation of a sponsorship ad is provided at the top and/or bottom of the screen view (termed "banners" in Appendix I). ('396, 8:63-65) (emphasis added) 73. In each claim clause (see the "target" section above), the word "advertisement" is used in context with a target that is set by a sponsor, and thus is limited to "a sponsorship ad," distinct from agate. Claim 6 also shows the distinction between "the advertisement" and "agate information." 74. Here again, ValueClick omitted language in their citations regarding the

distinction between sponsorship ads and agate data. One example of skipped

sentences is: "Preferably advertisements are positioned along the periphery (i.e., above, below, left or right) of the agate data, as defined by a 13 respective Page Display Object 35c. Accordingly, Page Data Objects 14 35b support Page Display Objects 35c which outline the possible screen content and presentation formats in which agate data 15 advertisements are to be displayed." (emphasis added)
16 17 18 19 20 21 22 23 24 25 26 27 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

It is clear from the entire quote that "agate data advertisements" refers to "advertisements" which are displayed separately, in a position relative to "agate data." V. "MEANS PLUS FUNCTION" CLAIM ELEMENTS (35 U.S.C. § 112, ¶6) 75. The following section provides a general discussion that relates to all of

the means plus function terms. A. 76. What is Missing As I stated in my Opening Declaration regarding each of these terms,

there is no algorithm disclosed in the specification for performing the function of each claim limitation. In the only disclosed embodiment, the invention "program 31 is implemented as an object oriented program . . . Each object is formed of data and subroutines (methods) for acting on the data." (`396 5:40-43; `735, 6:1-4)
18

1 (emphasis added). The patents fail to disclose the routines (methods) for acting on 2 the data of the various objects. Similarly, while Program controller 79 is described as 3 a series of routines (methods) on Web server 27, the routines (methods) themselves 4 are not disclosed. The algorithms for these routines could have been disclosed by 5 flow charts for each routine or by exemplary source code, but were not. Therefore, 6 there is insufficient corresponding structure disclosed for this limitation. 7 8

Incomplete Object Definition 77. As explained above, in object oriented programming each object

9 instance has a "type" defined by an object (or "class") definition. One of ordinary 10 skill would expect the object definition to be a precise specification of every field 11 (data variable) and every method. In object oriented programming languages such as
11 12 Java, C++, or Apple Media Language, there is "strong typing" i.e. every field (data

13 variable) and every method has a well-defined type. Specification of the type is 14 required. 15

78.

Each type of object has a name. For example, "User Interface Profile" is

16 the name for the type of object represented in Fig. 3D, i.e. 37c. Each of the "data" 17 variables (fields) of that object, i.e. "User computer ID," "categories," and "category 18 display" each would also have a specific type. Each of the methods (routines) of that 19 object would also have a return value of a specific type. A proper object definition 20 would specify, for example, that "User computer ID" is of type "User Computer" as
12 21 opposed to type "User" and thus precisely define what the variable is and how it

22 may be used. 23 24

79.
11

Likewise, the "category display" would have a type. However the type

Some other object-oriented-like languages have weaker typing. However, the `396 F.H. shows that the applicant amended the specification to correct and 26 capitalize the names of the objects, i.e. the types, showing the importance of strong typing in the disclosure. See, e.g. `396 F.H. in particular, RSI00000941, 942, 947, 27 948, 949, 950, 953, 954, 959, and 1033 through 1048. 12 28 See discussion, below, of the ambiguity of "Computer user ID".
25
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

19

1 is not specified. Had a precise type definition been provided, some of the questions
13 2 before the Court now in the Markman brief would have precisely clear answers.

3

80.

One of the impacts of the missing "type" information is that the sources

4 of the various files are not precisely disclosed. For example, it is not clear what the 5 following fields refer to: page ID (37e), object clicked ID (37e), ID (37f), 6 precipitating action ID (37f), related object ID (37f), or item ID (37f). For example, 7 "item ID" (37f) might refer to Page Data (35b). But this requires guesswork and 8 experimentation, especially because Page Data 35b not outlined. 9

81.

Each of the methods of that object would also have a name and a type.

10 The type of a method would be the type of its return value. Each parameter to the 11 method would also have a type. For example, in the following Java source code for
14 12 the "ImageViewer" object, the type of the "handleEvent" method is "boolean" and

13 the method takes one parameter "evt" which is of type "Event".: 14 15 16 17 18 19 20 21 22 23 24 25
public class ImageViewer extends Frame { ... public boolean handleEvent(Event evt) { if (evt.id == Event.WINDOW_DESTROY) System.exit(0); return super.handleEvent(evt); } ... private Image image = null; }

"ImageViewer," "Frame," "Boolean," "Event," and "Image" are object names, i.e. types. "id" and "image" are field (data variable) names which are part of the "Event" and "ImageViewer" objects, respectively. 82. In contrast to what one of ordinary skill in the art would have needed to

understand and implement the disclosed invention, the ValueClick Patents, in Figs.
13

See discussion, below, of the inconsistent disclosure, and dispute, regarding 26 "category display" 14 See "ImageViewer.java" file dated March 4, 1996, from Core Java part of 27 the Sunsoft Press Java Series CD-ROM, Sun Microsystems, Inc. Attached as Exhibit 28 3.
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

20

1 3B through 3G and Figs. 5A through 5C, only provide an imprecise, incomplete, and 2 inconsistent "outline" of some of the data field names, but not types. 3

83.

The outlines do not provide either names or types for the routines

4 (methods). And, as will be discussed in more detail below, no algorithms for the 5 methods have been provided as source code, pseudo code, or flow charts. 6 7

B. 84.

Lists of Method (Routines) As disclosed in the patent, each object is made up of both data (which is

8 partially disclosed in the outlines) and methods (routines). "Each object is formed of
15 9 data [fields] and subroutines (methods) for acting on the data." (`396, 5:42-43). A

10 listing of routines could have been disclosed but was not. 11

85.

Fig. 3A identifies two routines: a) main routine 39, and b) reporting

12 routine 41, which are essential to the invention but no source code, pseudo code, or 13 flow chart has been provided as a supporting disclosure for those two routines. 14 Further other routines are mentioned but not shown in any drawing figure. These 15 include: 16 17 18 19 20 21 22 23

· "User Object ...updating routine" ('396, 2:50-60) · "regression analysis" subroutine ('396, 2:42-44) · "Agate Object routine" (`396, 2:66-3:1) · "advertiser profile building routine" (`396, 3:7-10) · "performance routine" run remotely by advertisers (`396, 3:15-19) · travel plans inference routine ('396, 16:52-54) · weather related flag setting routine (`396, 16:54-57) 86. One example of the missing algorithms associated with the user profile

24 object 37c is: 25

The written description uses the word "subroutines" once (`735, 6:1-4) and 26 "routines" once (`735, 5:9-10) and both times equates those words with "methods". One of ordinary skill in the art would understand "routines" and "subroutines" both 27 to be substantially equivalent to "methods" which are the algorithms that act upon the 28 data in the objects that make up an object oriented program.
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

15

21

1 2 3 4 5

"Finally, the user's User Interface Object 37c records his travel plans (as inferred from the user's activity with the displayed Travel Page/Screen View). User Interface Object 37c also sets a flag in program 31 to send the user an appropriate weather forecast the day before he travels." (`396, 16:54-57) (emphasis added). 87. One of ordinary skill in the art would have understood that hundreds of

6 routines would have been associated with the disclosed objects and would have 7 understood that those hundreds of routines have not been disclosed or even hinted at. 8 The outlines do not identify the names or types of the routines (methods) that would 9 be needed to implement the invention. 10

88.

Other than indirect reference to Java associated with the content of the

11 pages, the applicant did not disclose a programming language that could be used to 12 use to implement these objects, let alone indicate a preexisting, well-known set of 13 routines that could be used to implement these methods. 14

89.

ValueClick raised the AllVoice case in its Opening Brief (p. 27, l. 28

15 through p. 28, l. 8). The AllVoice Patent does include detailed algorithms that are 16 represented as flow charts. The step identified by ValueClick, "DETERMINE 17 WORD LOCATION IN TEXT" (step S52 of Fig 8A), combined with the disclosure 18 in `273 7:7 et seq. further disclosed "well-known features of windows operating 19 system (messages, operating system function calls, and hooking" (AllVoice) which to 20 one of ordinary skill in the art would identify specific routines that were known for 21 implementing the specific function. Step S52 is a very low level operation that 22 corresponds to one or more Windows operating systems function calls or routines 23 that were available to programmers in developer libraries. The developer libraries 24 were well documented and used by most Windows programmers. In contrast, the 25 ValueClick Patents provide no flowcharts, and no identification of specific routines 26 available from a well-known, third party source. The ValueClick Patents do not even 27 clearly identify a programming language which may be used to implement the 28 invention.
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

22

1

90.

Thus, in my opinion, one of ordinary skill in the art would be not be able

2 to identify structure including the hundreds of undisclosed routines associated with 3 the complex data, which is only partially disclosed for the objects in the ValueClick 4 Patents. 5 VI. 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

ADDITIONAL COMMENTS ON SPECIFIC 112 ¶6 TERMS A. 91. Data Assembly For Providing And Supporting Display Of Agate Information To Users Of The Computer Network I discussed the term "data assembly for providing and supporting

display of agate information to users of the computer network" in my Opening Declaration (see ¶89-93). i. 92. 93. No Outline for Page Objects 35a through 35c There is no outline for the data fields of the page objects 35a-35c. There are no drawing figures that provides an "outline" of the data

stored in each Page Object. This is a significant inconsistency which highlights the partial nature of the disclosure of the ValueClick Patents. There is no disclosure of the outline of the following objects: Figure NONE NONE NONE 94. Item 35a 35b 35c 3A Name Page Page Data Page Display Missing Figure Name MISSING MISSING MISSING

The only drawing that discloses an outline of a page is the User Home

Page 43. (Fig. 4A) ii. 95. Missing Details for 37b (Fig 3C) Also as discussed above, the title and apparent content expected to be

24 found in Fig 3C are missing. The written description states that "limitations" and 25 "capabilities" are contained in the "User Computer Object", namely "limitations and 26 capabilities of the user's computer system" (`396 6:16). 27 96. However, in my opinion the data provided only lists the capabilities and 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

23

1 it appears that other data elements (in addition to the title) have been erroneously 2 omitted from the drawing (i.e. Fig 3C). 3

97.

Thus in my opinion, one of ordinary skill in the art would not have

4 understood what the "corresponding structure" for this term, "data assembly for 5 providing and supporting display of agate information to users of the computer 6 network," was, and would not be able to determine what was or was not covered by 7 claims using this term. 8 9 10 11 12

B.

User Profiling Member For Recording Information Regarding Each User Including Indications Of User Responses And Physical Activity With Respect To Screen Views During Display Of Said Screen Views, Such That The User Profiling Member Enables Creation Of A Psychographic Profile Of Each User From Said Recorded Information And Indications Of User Responses I discussed the term "a user profiling member for recording information

98.

13 regarding each user including indications of user responses and physical activity with 14 respect to screen views during display of said screen views, such that the user 15 profiling member enables creation of a psychographic profile of each user from said 16 recorded information and indications of user responses" in my Opening Declaration 17 (see ¶84-88). 18 99. One of ordinary skill in the art would expect that every object, especially 19 those disclosed in Figs. 3B through 3G and 5A through 5D would have an ID field 20 that acts as a key for the object, however one of the outlines has an unqualified "ID" 21 (37f, Fig 3G). As discussed above, a proper object definition provides the template 22 (like a cookie cutter) upon which object instances are formed. The ID field is then 23 used to uniquely identify which instance of a particular type of object is being 24 referred to. Thus it appears that even the disclosed outlines for data fields that make 25 up the objects are incomplete. 26 100. The incomplete specification of the object ID leads to awareness of 27 further ambiguity and omissions. Referring the User Viewing History 37f object in 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

24

1 Fig. 3G, there are several entries in the outline with "ID" or "identifier": 2 3 4 5 6

· ID · ordinal sequence identifier · precipitating action ID · related object ID · item ID

7 What do these IDs reference? If they are inter-object references, the ID fields are 8 missing from the disclosure, and these references are not shown as arrows on Fig. 3A 9 because there are four or five IDs and only one arrow shown. Further, if "item ID" 10 refers to a Page data 35b object that was viewed, we again run into the lack of 11 disclosure regarding the data outline and methods associated with the Page Data 35b 12 object. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

i.

"User computer ID"--Incomplete, Ambiguous and Inconsistent

101. This confusion then leads us back to the specific dispute where the parties are trying to say that "User computer ID" in User Interface Profile 37c (Fig. 3D) means a particular thing as it relates to claim terms. The imprecision, ambiguity, and lack of disclosed algorithmic structure may make it impossible for anyone to understand the bounds of the claimed invention. 102. Had Fig. 3C contained a title, Fig. 3C may have provided one possible answer. One could argue that the title for object 37b should have disclosed as "User Computer" based on the inconsistent disclosure of Fig. 3A (discussed above) and `396, 6:13. Under this understanding, then "User computer ID" would be the interobject reference to an instance of the "User Computer" 37b object. 103. The other possibility in this ambiguity is that "User computer ID" is a "computer ID" for identifying a user (see ValueClick Brief p. 22 lines 10-15; Kent ¶10), as opposed to the various "identifiers" in the "User" object 37a (Fig. 3B). 104. Further neither Fig. 3B nor Fig. 3D consistently account for all the data
25

Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

1 described in the written description. For example, the "unique user identification 2 code" and the "computer identification number": 3 4 5 6 7 8 9 10 11 12 13 14 15 16

"For a new user, the Home Page 43 effectively requests a user name and password. In response to the user-provided data, main routine 39 immediately builds a cookie if possible. Included in the newly built cookie is a unique user identification code (preferably numeric), time and date of login, and computer identification number to distinguish between home and work logins." (`396, 13:6214:1)(emphasis added). "The User Interface Object 37c provides a unique (preferably numeric) identifier of the user." (`396, 6:22-23). Note that User Interface Profile 37c object is only disclosed as having three elements, one of which is "Computer user ID". The above quotation could be satisfied due to the one-to-many relationship that could exist between "User" 37a object and "User Computer" 37b object. However, this level of detail for the data schema has not been provided by the incomplete and imprecise disclosure of the ValueClick Patents. ii. "category" and "category display"--Incomplete, Ambiguous and Inconsistent

105. The lack of precise type specification regarding the objects leads to

17 other disputes. The "category" and "category display" in User Interface Profile 37c 18 (Fig. 3D) are also important pieces of data as it relates to the claims and claim 19 construction. Regarding the User Interface Profile 37c object, the specification 20 provides the following: 21 22 23 24 25 26 27 28

"... the user's User Interface Object 37c holds indications of his categories of interest, including specific items of interest in each category of information, and his display/format preferences (colors, design, layout, etc.). Based on these recorded details, program 31 constantly and automatically tailors screen views (content and presentation)...." (`396, 17:1-9) (emphasis added) "In the preferred embodiment, the various categories of interest include stock trading portfolio, sports, news, weather, theater and television schedules, telephone directory, travel data, classified ads and personals information, and the like. Display preferences include orientation, color scheme, screen quadrant/location and the like,
26

Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

1 2 3 4 5 6 7 8 9 10 11

indicated with respect to the category of information. For example, one user may tend to like stock information displayed in tabular form on a blue background and weather displayed on a map scene. Another user may prefer stock information displayed in a running 1line quote at the bottom of the screen and weather displayed in a tabular format by city on a green background, and so forth. (`396, 6:27-39) (emphasis added) "Further, the tracking and profiling member records presentation (format) preferences of the users based on user viewing activity. Preferences with respect to color schemes, text size, shapes, and the like are recorded as part of the psychographic profile of a user. In turn, the psychographic profile enables the data assembly to customize presentation (format) of agate information, per user, for display to the user." (`396, 2:16-23) (emphasis added) 106. Based on these three citations, is clear that the "per user"

12 "psychographic profile" includes "categories of interest" (more precisely the 13 "categories" entry in User Interface Profile 37c object outline of Fig. 3D) and 14 "presentation (format) preferences" (more precisely the "category display" entry in 15 User Interface Profile 37c object outline of Fig. 3D). Further the "category display" 16 is referred to as "Display preferences" and includes details of the "screen" view such 17 as "orientation, color scheme, screen quadrant/location for each of the "categories of 18 interest".16 19 107. However, the "type" of "category display" is not specified in the outline. 20 The specification further states, 21 22 23 24 25 26 27

"The User Interface Object 37c provides a unique (preferably numeric) identifier of the user. The User Interface Object 37c also provides indications of categories of interest to the user and a primary screen display for each category customized to that user. The foregoing information is held in records illustrated in FIG. 3d." (`396, 6:22-27) (emphasis added). This suggests that the type of "category display" is an array (a plurality of same type
16

PointCast which was a screen saver took up the entire screen and allowed the user 28 to specify the display and content preferences.
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

27

1 objects) of "Page Display" objects 35c. However, as discussed above, there is no 2 drawing figure providing an outline for Page Display object 35c. Fig. 4A gives an 3 example of the User Home Page 43. However Home Page 43 is not an object but is 4 "an HTML (HyperText) document generated through a set of Page Objects 35a, b, c" 5 (`396 13:47-50). 6

108. For other object variables, there is no disclosure that would provide even

7 a hint at what the object types are. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

iii.

Another Example--"Computer ID" versus "Computer user ID"

109. Another example of the imprecision and inconsistency of the disclosure are the "Computer ID" field of User Session 37d object (Fig. 3E) versus the "Computer user ID" field of User Interface Profile 37c object (Fig 3D). A person of ordinary skill in the art would wonder ­ and be unable to determine ­ whether these object references to the same type of object (e.g. "User Computer" 37b object), or (as discussed above) are they different? iv. Cursor Position v. Object Clicked 110. The specification states "The User Action History Object 37e stores each click of a mouse and corresponding cursor position to effectively record the user's motions/movements in a session." ('396, 6:57-59) However, Fig. 3F does not have fields for storing the (x,y) coordinates of the mouse cursor location, or the user's movements. The ValueClick Patents don't disclose how, for example, the "object clicked ID" is determined. Either there is missing data fields or the information is insufficient to determine how the "indications of user responses and physical activity with respect to screen views during display" of this claim term are stored and used to determine what object was clicked (especially as correlated to what was viewed "during display"). 111. Thus in my opinion, one of ordinary skill in the art would not have understood what the "corresponding structure" for this term, "a user profiling
28

Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

1 member for recording information regarding each user including indications of user 2 responses and physical activity with respect to screen views during display of said 3 screen views, such that the user profiling member enables creation of a 4 psychographic profile of each user from said recorded information and indications of 5 user responses," was, and would not be able to determine what was or was not 6 covered by claims using this term. 7 8 9 10 11 12

C.

A Program Controller Responsive To User Commands Of A User For Generating Screen Views To The User, The Program Controller (I) Obtaining Information From The Agate Data Portion And User Profiling Member, Including Creating And Obtaining The Psychographic Profile Of The User And (Ii) Generating And Displaying Appropriate Screen Views To The User Based On The Created Psychographic Profile Of The User

112. I discussed the term "a program controller responsive to user commands

13 of a user for generating screen views to the user, the program controller (i) obtaining 14 information from the agate data portion and user profiling member, including 15 creating and obtaining the psychographic profile of the user and (ii) generating and 16 displaying appropriate screen views to the user based on the created psychographic 17 profile of the user" in my Opening Declaration (see ¶99-103). 18

113. Kent provides his conclusions regarding "program controller" at Kent

19 ¶14. As discussed above, I disagree with Kent's conclusions. As detailed above, the 20 term "program controller" provides inadequate structural support for a means plus 21 function claim term. The program controller is clearly defined as a series of routines 22 (methods) i.e. software running on a CPU. 23 24 25

"Program controller 79 is a series of routines (methods) on Web server 27." (`396, 4:48-49; `735, 5:9-10) (emphasis added). 114. ValueClick is technically wrong when it states that "controller of a

26 program [is] (e.g. a computer CPU)" (ValueClick Brief p. 25, lines 4-6). This 27 statement is inconsistent with the otherwise unambiguous statement in the 28
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)

29

1 specification. 2

115. I also was the testifying expert in the Markman hearing for ACTV v.

3 Disney. An ACTV patent is mentioned in Wilkin, i.e. "U.S. Pat No. 4,602,279 4 Freeman, M. J. `Method For Providing Targeted Profile Interactive CATV Displays' 5 assignor to ACTV, Inc. ..." (Wilkins 2:29-36) 6

116. In the ACTV v. Disney matters, I provided an opinion regarding various

7 means plus function claim elements. For example, I provided an opinion regarding 8 the meaning and supporting algorithm for a "controller means... for interpreting". 9 The District Court adopted my opinion in its legal opinion. The Federal Circuit later
17 10 affirmed that interpretation over Disney's appeal.

11

117. Thus in my opinion, one of ordinary skill in the art would not have

12 understood what the "corresponding structure" for this term, "a program controller 13 responsive to user commands of a user for generating screen views to the user, the 14 program controller (i) obtaining information from the agate data portion and user 15 profiling member, including creating and obtaining the psychographic profile of the 16 user and (ii) generating and displaying appropriate screen views to the user based on 17 the created psychographic profile of the user," was, and would not be able to 18 determine what was or was not covered by claims using this term. 19 VII. CONCLUSION 20

118. I understand that certain aspects of discovery have not been completed

21 in this matter. The findings and opinions set forth in this declaration are based on my 22 work and examinations to date. I may continue my examinations. I may also receive 23 additional documentation and other factual evidence over the course of this litigation 24 that will allow me to supplement and/or refine my opinions. I reserve the right to add 25 26 Federal Circuit 2002-1491 "In particular, Disney argues that the district court 27 ignored the ordinary meanings of these terms, accepted definitions contrary to the patents' written descriptions, .... we disagree that the district court erroneously construed the 28 terms `interpreting' and `decoding'..."
Heller ROMAN DECLARATION IN SUPPORT OF DEFENDANT'S RESPONSIVE MARKMAN BRIEF Ehrman LLP CASE NO. CV-07-2052 MMM (JCX)
17

30