Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`
`FACEBOOK, INC., INSTAGRAM, LLC, and WHATSAPP INC.,
`Petitioners
`
`v.
`
`BLACKBERRY LIMITED
`Patent Owner
`
`
`
`
`IPR2019-00516
`U.S. Patent No. 8,279,173 B2
`
`REPLY DECLARATION OF SANDEEP CHATTERJEE, PH.D.
`
`
`
`
`
`
`
`
`
`
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`001
`
`

`

`Table of Contents
`
`
`Page
`
`
`
`
`
`I. 
`
`RESPONSE TO DR. SURATI’S OPINIONS RE WHETHER “TAG
`SOURCES” HAVE TO BE “SEPARATELY SEARCHABLE” .................. 1 
`RESPONSE TO DR. SURATI’S OPINIONS RE ZUCKERBERG
`REFERENCE (GROUNDS 1-5) .................................................................. 13 
`A. 
`Zuckerberg Discloses Distinct Tag Sources ...................................... 13 
`B. 
`Zuckerberg “Dividing Line” Argument ............................................. 17 
`III.  RESPONSE TO DR. SURATI’S OPINIONS ABOUT
`COMBINATION OF ZUCKERBERG WITH ROTHMULLER AND
`PLOTKIN (GROUNDS 2-5) ........................................................................ 18 
`A.  Dr. Surati’s Argument That Rothmuller and Plotkin Do Not
`Disclose Distinct “Tag Sources” Is Irrelevant to Grounds 2-5 .......... 18 
`Dr. Surati’s Challenge to the Motivation to Combine
`Zuckerberg with Rothmuller and Plotkin Fails .................................. 20 
`C.  MacLaurin Provides an Express Motivation to Combine .................. 22 
`IV.  RESPONSE TO DR. SURATI’S OPINIONS RE ROTHMILLER-
`BASED OBVIOUSNESS (GROUNDS 6-7) ............................................... 30 
`CONCLUSION ............................................................................................. 33 
`
`II. 
`
`V. 
`
` 
`
`B. 
`
`
`
`
`
`-i-
`
`
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`002
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`I, Sandeep Chatterjee, Ph.D., declare as follows:
`
`1.
`
`I have been asked to review and respond to certain points raised in the
`
`“Patent Owner’s Response” filed with respect to the IPR petition for U.S. Patent No.
`
`8,279,173 in IPR2019-00516. I understand that Patent Owner submitted a
`
`declaration from Dr. Rajeev Surati (Ex. 2001) (“Surati Declaration”) in support of
`
`its Patent Owner’s Response. I have therefore been asked to review and respond to
`
`statements in the Surati Declaration as well.
`
`I.
`
`RESPONSE TO DR. SURATI’S OPINIONS RE WHETHER “TAG
`SOURCES” HAVE TO BE “SEPARATELY SEARCHABLE”
`Dr. Surati devotes a significant discussion to the term “tag source” and
`2.
`
`argues that the term should be construed as “separately searchable collections of
`
`tags.” (Surati Decl., ¶¶82-107.) I have carefully reviewed Dr. Surati’s arguments,
`
`and for the reasons below, I respectfully disagree.
`
`3.
`
`Dr. Surati appears to rely exclusively on the figures and textual
`
`description in the ’173 patent specification to support his “separately searchable”
`
`limitation on “tag sources,” but in my opinion, those statements do little more than
`
`restate the claim language and cannot be fairly read, by a person of ordinary skill in
`
`the art, as imposing a “separately searchable” requirement. Dr. Surati does not cite
`
`anything from the ’173 patent prosecution history for his construction, or identify
`
`anything in the claim language itself that would impose such a requirement.
`
`4.
`
`Turning first to the claims, a person of ordinary skill in the art would
`
`
`
`1
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`003
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`have found nothing suggesting that a tag source must be “separately searchable.”
`
`Independent claim 1, for example, merely recites “displaying a tag list including
`
`tags from one or more tag sources matching a search string,” and a substantially
`
`similar limitation appears in the other challenged independent claims. This language
`
`merely recites an end result – the display of a tag list with particular content, i.e. tags
`
`from one or more tag sources matching a search string. The claim does not address
`
`how the tag list was constructed or the mechanics of how information was located
`
`or retrieved from the one or more “tag sources.” The recitation of “tags from one or
`
`more tag sources matching a search string” may suggest that search occurred at
`
`some time before the display, but this claim language does not specify how any such
`
`search was conducted or suggest that tag sources must be separately searchable.
`
`5.
`
`Turning next to specification, I am informed by counsel for Petitioner
`
`that, under the patent laws governing construction of claim terms, it is generally
`
`improper to import limitations or details from the specification into the claims. I am
`
`further informed that this rule applies even if a patent specification describes only a
`
`single embodiment. I am further informed that this rule stems from the differences
`
`in the purposes of the claims and the patent specification; the former defines the
`
`scope of the invention and the latter teaches and enables persons of ordinary skill in
`
`the art to make and use the invention. I am further informed that an embodiment
`
`from the specification can impose a claim limitation where statements in the
`
`
`
`2
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`004
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`specification provide a clear and unmistakable disavowal or disclaimer. I am further
`
`informed that a disclaimer or disavowal will not be found when the statements in the
`
`specification are ambiguous or susceptible to multiple reasonable interpretations.
`
`6.
`
`Although I have articulated the principle that it is generally improper to
`
`“import” a limitation from the specification into the claims, Dr. Surati’s position
`
`cannot even be fairly characterized as attempting to do that. A person of ordinary
`
`skill in the art reading the ’173 specification would find nothing in the specification,
`
`for any embodiment, requiring that “tag sources” be “separately searchable.” Dr.
`
`Surati does not identify any clear disclosure of separately searchable tag sources.
`
`Dr. Surati has at best inferred a “separately searchable” characteristic based on a
`
`high-level description of the exemplary “tag sources” in the specification, and then
`
`imported that inferred characteristic into the claims. As I will explain below, the
`
`specification does not support such an approach.
`
`7.
`
`The specification describes the searching of tag sources in the following
`
`passage, which describes the search in a high-level fashion without details about the
`
`actual search or how matching information in the underlying tag sources is accessed,
`
`identified, or retrieved:
`
`Now referring to FIG. 4A, shown in screen 400A is an illustrative tag
`selection user interface 404 for displaying a tag search facility as may
`be presented by photo tag selection module 148B. As shown in FIG.
`4A, the user is initially presented with a tag entry field 406 indicating
`
`
`
`3
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`005
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`
`that he should start typing a tag. Upon completion of typing, the user
`may click “OK” 408 to select the tag.
`In an embodiment, as the user begins to type, photo tag selection
`module 148B may be configured to search one or more selected “tag
`sources” for tags that match the currently entered text. As shown by
`way of illustration in screen 400B of FIG. 4B, these tag sources could
`include, for example, a list of friends from an online service like
`Facebook™, a list of contacts from the user's address book 142, a list
`of the user's browser bookmarks (in Internet browser 138), a cache of
`recent free-form text entries, etc.
`
`(’173, 5:32-47 (emphasis added).) The passage above states that the photo tag
`
`selection module 148B “may be configured to search one or more selected ‘tag
`
`sources’ for tags that match the currently entered text” (’173, 5:39-42), but the
`
`specification does not explain how a search of selected tag sources should be carried
`
`out. The specification does not disclose any algorithm or technique for searching
`
`tag sources, nor does it require that the “tag sources” be stored in any particular way
`
`or in any physical location. Dr. Surati agreed, in fact, that the claims do not impose
`
`such a requirement. (Ex. 1021 (Surati Depo.), 165:23-166:2 (“Q. For a system that
`
`uses multiple tag sources, does Claim 1 impose any limitations on where each of
`
`those tag sources are physically stored? A. I don’t believe so.”).)
`
`8.
`
`A person of ordinary skill in the art would have recognized that there
`
`are numerous ways to implement “search[ing] one or more selected ‘tag sources’ for
`
`
`
`4
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`006
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`tags that match the currently entered text” (’173, 5:39-42), that would not require
`
`that each tag source be separately searchable.1 The search mentioned in the block-
`
`quoted passage above could involve a single search operation that extends across all
`
`tag sources, a separate search for each tag source, and any combination in between.
`
`A person of ordinary skill in the art would have interpreted the dearth of information
`
`about how a “search” of “tag sources” must be carried out as evidence that the claims
`
`do not impose any restrictions about how the “tag sources” are arranged, let alone a
`
`requirement that each tag source be “separately searchable.”
`
`9.
`
`Dr. Surati’s opinion relies primarily on the exemplary “tag sources”
`
`described in the ’173 patent specification. Dr. Surati argues that “every example of
`
`a ‘tag source’ in the ’173 patent specification refers to obtaining tags from separate
`
`software or hardware components. The specification, consistent with the claim
`
`language, thus confirms that different ‘tag sources’ can be separately searched to
`
`obtain tags.” (Ex. 2001 (Surati Decl.), ¶101.) I respectfully disagree.
`
`10. First, even if it were the case that the specification clearly disclosed that
`
`
`1 It is not entirely clear from the Surati Declaration what “separately searchable”
`means or how one would go about ascertaining if the limitation is satisfied in a
`particular system of a prior art reference. At his deposition, Dr. Surati provided
`several different formulations of the “separately searchable” concept, generally
`focusing on an ability to retrieve information from a particular “tag source” using a
`query that does not involve or make any reference to other tag sources. (Surati
`Depo., 119:16-123:13.) For purposes of my analysis, I have interpreted Dr. Surati’s
`explanation as meaning that a particular tag source is “separately searchable” it can
`be searched without having to search any other tag source. (Surati Decl., ¶105.)
`5
`
`
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`007
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`every exemplary tag source could be separately searched (which it does not), this is
`
`not enough to transform the “separately searchable” characteristic into a claim
`
`limitation. The ’173 patent specification consistently refers to the tag sources
`
`identified in the patent as mere examples of possible tag sources. (’173, 5:42-47
`
`(“As shown by way of illustration in screen 400B of FIG. 4B, these tag sources could
`
`include, for example, a list of friends from an online service like Facebook™, a list
`
`of contacts from the user’s address book 142, a list of the user’s browser bookmarks
`
`(in Internet browser 138), a cache of recent free-form text entries, etc.”), 6:6-13
`
`(“Significantly, as the matching tag list 412 includes possible tags that may be used
`
`from various selected tag sources (such as the user’s Facebook friends, the user’s
`
`address book 142, a list of the user’s browser bookmarks from Internet browser 138,
`
`a cache of the recent free-form text entries, etc.), the user is provided with a simple
`
`way to associate subjects or objects in a photo with a predefined ‘tag’ from one of a
`
`number of selected tag sources, as may be defined by the user.”) (emphasis added).)
`
`The Surati Declaration itself consistently refers to the tag sources identified in the
`
`’173 specification as merely “exemplary.” (Surati Decl., ¶97 (“[E]very exemplary
`
`‘tag source’ in the ’173 patent specification is consistent with separately searchable
`
`collections of tags.”), ¶98 (“These exemplary ‘tag sources’ correspond to tags
`
`obtained ‘from’ separate software applications.”).) Dr. Surati also acknowledged
`
`this at his deposition. (Surati Depo., 148:8-150:15.)
`
`
`
`6
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`008
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`
`11. Accordingly, even if one could conclude that the specification clearly
`
`describes exemplary “tag sources” in which each tag source is stored in a distinct
`
`hardware or software component (which as I will explain below it does not), Dr.
`
`Surati has not explained how this represents anything more than a happenstance
`
`characteristic based on the particular examples in the specification. The patent is
`
`clear that the tag sources identified in the ’173 patent are exemplary, and the claims
`
`do not preclude further tag sources beyond the examples in the specification. And
`
`because the storage of “tag sources” in a distinct hardware or software module is
`
`merely a characteristic that Dr. Surati has at best inferred from the specification, it
`
`is not clear why that type of storage relationship would necessarily apply to
`
`additional tag sources that a developer may create. A person of ordinary skill in the
`
`art would have understood that a developer would be free to add any number of
`
`additional “tag sources,” some of which may be stored in a memory area shared with
`
`other tag sources. Nothing in the specification precludes such an implementation.
`
`12. Second, a person of ordinary skill in the art would not have found any
`
`clear description in the specification suggesting that the exemplary “tag sources” are
`
`stored physically or logically separate from one another, an assumption that appears
`
`critical to Dr. Surati’s reasoning. The Surati Declaration relies primarily on a high-
`
`level block diagram of Figure 1 to suggest that three of the exemplary tag sources
`
`are stored by or within different software applications 34, i.e. the photo tag selection
`
`
`
`7
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`009
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`module 148B, Internet browser 138, and address book 142. (Surati Decl., ¶¶97-98.)
`
`But the specification does not describe how these software applications actually
`
`store or make tags available for display in a tag list, let alone suggest that the
`
`software applications store tags separately or independently from one another in a
`
`way that would preclude a single search. In fact, Figure 1 suggests that each of the
`
`software modules share the same flash memory 108 and random access memory
`
`(RAM) 106. (’173. Fig. 1.) A person of ordinary skill in the art would have
`
`understood that, at the time of a search, data from the tag source must at least be
`
`loaded into such memory to evaluate each tag and identify those that match.
`
`13. Dr. Surati implies that the free-form text tag entries are stored within
`
`photo selection module 148B, but the specification does not actually say this. It
`
`instead says that photo selection module 148B “may create a new free-form tag entry
`
`and add it to a free-form text cache as a new tag entry,” but the specification does
`
`not say where this “free-form text cache” is stored. (’173, 5:56-59.) Thus, while a
`
`particular software application might be responsible for a particular tag source, that
`
`does not tell us anything about the actual storage of the tags for that particular tag
`
`source. Nor does it preclude the tags from the various sources from being collected
`
`into a common memory buffer or cache (such as a file or table located in random
`
`access memory 106) at the time a search for matching tags is conducted.
`
`14. There are multiple examples in which Dr. Surati himself appeared to
`
`
`
`8
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`010
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`acknowledge the uncertainty with respect to where or how exemplary “tag sources”
`
`in the specification are actually stored. For example, the Surati Declaration suggests
`
`that a particular exemplary tag source – “a list of friends from an online service like
`
`Facebook™” – corresponds to a remote database accessible over a network. (Surati,
`
`¶99 (quoting ’173, 5:43-44).) But at his deposition, he acknowledged that the list of
`
`Facebook friends could be also stored on locally on the user’s device. (Surati Depo.,
`
`150:16-151:21.) Similarly, the Surati Declaration implies that the “list of contacts
`
`from the user’s address book 142” is stored locally on the user’s device. (Surati
`
`Decl., ¶98 (quoting Fig. 1).) But at his deposition, he admitted that the address book
`
`could be stored in a local database, or could have been replicated from something
`
`“stored in the cloud.” (Surati Depo., 152:2-9.)
`
`15. All of this confirms my point above, that there simply is not enough
`
`information in the specification for a person of ordinary skill in the art to draw any
`
`reliable conclusions with respect to about how the exemplary tag sources are stored
`
`in relationship to each other, such that each tag source is “separately searchable.” A
`
`person of ordinary skill in the art would have found nothing in the specification
`
`expressing any “clear and unmistakable disclaimer” that would warrant importing a
`
`separately searchable limitation into the definition of tag sources.
`
`16. Finally, even if one could infer that the tags in the various tag sources
`
`are stored in separate software or hardware modules, it does not necessarily follow
`
`
`
`9
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`011
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`that each tag source must be “separately searchable,” because a search of the tags
`
`can be implemented in a number of ways regardless of how the tags sources are
`
`arranged in relationship to one another.
`
`17. As I explained previously, the ’173 specification imposes no limits on
`
`how a search of tag sources could be conducted. The claim language itself does not
`
`even require a search of the actual tag sources themselves. Claim 1 for example
`
`recites “displaying a tag list including tags from one or more tag sources matching
`
`a search string,” which does require that the tags in the list be “from” the one or
`
`more tag sources, but does not require that the matching tags be obtained directly
`
`from the tag sources themselves, or a result of searches applied directly to the tag
`
`sources themselves. For example, a person of ordinary skill in the art could have
`
`implemented the tag list of claim 1 by using the well-known technique of pre-
`
`fetching and collecting data from the “one or more tag sources” into a single
`
`common cache or buffer in memory, which could later be searched (in lieu of the
`
`underlying “tag sources”) upon entry of a search string. This technique would allow
`
`a single search of the common cache or buffer to identify matching tags that were
`
`pre-fetched from the various tag sources. The resulting tag list, when displayed,
`
`would still qualify as “a tag list including tags from one or more tag sources
`
`matching a search string,” because while the claim requires that the tags be “from”
`
`the tag source, it does not require that the tag list be compiled from a search applied
`
`
`
`10
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`012
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`directly to the tag sources themselves. And this would not involve any “separately
`
`searchable” tag sources because the search itself applied to common cache or buffer
`
`memory in which data from all tag sources was collected.
`
`18. Nothing in the specification preludes such an implementation, and these
`
`types of pre-fetching implementations were commonplace to persons of ordinary
`
`skill in the art. In fact, it was well-known to persons of ordinary skill in the art that
`
`one benefit of a “pre-fetching” system is that it reduces overhead by avoiding a need
`
`to access the underlying data source (such as a database). (See Alan Jay Smith,
`
`Sequentiality and Prefetching in Database Systems (1978) [Ex. 1026], at 001-002
`
`(“One method used in some systems to reduce the frequency of I/O operations is to
`
`maintain in a main memory buffer pool a number of blocks of the database. Data
`
`accesses satisfied by blocks found in this buffer will take place much more quickly
`
`and with much less computational overhead.”).)
`
`19.
`
`In the case of the tag sources of the ’173 patent, as noted, one of the
`
`exemplary tag sources is “a list of friends from an online service like Facebook™”
`
`(’173, 5:43-44), and a person of ordinary skill in the art would have recognized that
`
`a search of this tag source would be much faster if the friends were pre-fetched so
`
`they could be searched locally by the device rather than having to access the
`
`underlying “tag source” remotely over a network. There is nothing in the ’173 patent
`
`that would suggest that the inventors intended to foreclose existing pre-fetching and
`
`
`
`11
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`013
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`caching techniques in connection with any search of tag sources.
`
`20. Dr. Surati relies on a statement from the patent specification that the
`
`“photo tag selection module 148B may be configured to search one or more selected
`
`‘tag sources’ for tags.” (Surati Decl., ¶95 (quoting ’173, 5:39-42) (emphasis mine).)
`
`Dr. Surati claims that the fact that photo selection module 148B can “search just
`
`one” tag source supports his “separately searchable” construction. (Id., ¶¶95, 96.)
`
`But it does not. The fact that photo selection module 148B can search a single tag
`
`source says nothing about how the photo selection module actually performs that
`
`particular search, let alone performs a search across multiple tag sources. The
`
`specification nowhere states that each tag source is separately searched.
`
`21. And this brings me to a final problem with Dr. Surati’s proposed
`
`construction – it is not clear if it imposes any requirements at all. During his
`
`deposition, Dr. Surati was repeatedly asked whether or not, when his construction of
`
`“tag sources” was applied to claim 1, the claim would actually require that the
`
`claimed “tag sources” be separately searched. Dr. Surati testified that separate
`
`searching need only be a capability, not a requirement, or his construction. (Surati
`
`Depo., 173:12-174:25.) For example, in connection with claim 1 he testified:
`
`Q.
`
`Just to make sure I understand, for a system that
`uses multiple tag sources, Claim 1 does not require
`that each of those tag sources be actually searched
`separately?
`
`
`
`12
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`014
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`
`A.
`
`I don’t think it necessarily requires that they be
`searched separately, but the result had better be the
`same as if they were.
`(Surati Depo., 174:16-25.) With respect, this statement does not make sense to me
`
`and contradicts other parts of Dr. Surati’s opinions. For example, the Surati
`
`Declaration states that “[t]he ’173 Patent itself confirms the plain meaning of ‘tag
`
`sources’ as describing distinct and separate processes for obtaining tags” (Surati
`
`Decl., ¶88), but apparently the actual performance of those “distinct and separate
`
`processes” is not meaningful to the claim. Under Dr. Surati’s construction of “tag
`
`sources,” this states nothing more than an abstract technical capability that has no
`
`bearing on how a practicing system actually accesses or searches the claimed “tag
`
`sources.”
`
`II. RESPONSE TO DR. SURATI’S OPINIONS RE ZUCKERBERG
`REFERENCE (GROUNDS 1-5)
`A. Zuckerberg Discloses Distinct Tag Sources
`22. The Surati Declaration argues that Zuckerberg fails to disclose two
`
`distinct “tag sources” because the text list 544 and the friend lists 546 are part of a
`
`single tag source – a list of previously used tags. (Surati Decl., ¶¶115-125.) I
`
`respectfully disagree. As I explained in my opening declaration, it would have been
`
`obvious that the collection of tags in the text list 544 is separate from the collection
`
`of tags in the friends list 546. (Ex. 1002, ¶77.)
`
`23. Dr. Surati relies heavily on the following statement in Zuckerberg: “In
`
`
`
`13
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`015
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`some embodiments, the list of previously used tags includes a text list 544 and a
`
`friends list 546.” (Zuckerberg, 8:56-58 (underlining added).) The gist of Dr.
`
`Surati’s argument is that the “list of previously used tags” constitutes a single “tag
`
`source,” and as such, the text list 544 and friends list 546 do not qualify as separate
`
`“tag sources” under his proposed construction. To the extent Dr. Surati’s arguments
`
`rely on his position that the term “tag sources” requires separate searchability, those
`
`arguments fail because his claim construction position is incorrect as explained.
`
`24. More fundamentally, Dr. Surati ignores the fact that Zuckerberg makes
`
`plain that the “list of previously used tags includes a text list 544 and a friends list
`
`546.” (Zuckerberg, 8:56-58 (underlining added).) Zuckerberg thus discloses that
`
`the “list of previously used tags” includes at least two sub-lists. The items in the text
`
`list 544 and the items in the friends list 546 clearly represent two distinct collections
`
`of tags – friends list 546 contains a collection of contacts or contact addresses (8:62-
`
`66), and text list 544 contains a collection of text strings (8:52-58). These two
`
`collections clearly cover different categories of information and can be readily
`
`recognized and distinguished by the Zuckerberg system, as evidenced by the fact
`
`that the user interface in Figure 5 visually displays “friends list” and “text list” in
`
`two groups separated by a dividing line. (Zuckerberg, Fig. 5.)
`
`25. The fact that Zuckerberg can recognize and distinguish “friends list”
`
`tags from “text list” tags confirms the existence of two distinct collections of tags.
`
`
`
`14
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`016
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`As I mentioned above in my critique of Dr. Surati’s definition of “tag sources,”
`
`nothing in the claim properly construed impose requirements on how the tag sources
`
`must be arranged in computer memory.
`
`26. But even if the claim were construed to require some physical
`
`separation between distinct “tag sources,” this would not render the claims non-
`
`obvious over Zuckerberg.
`
` As noted,
`
`Zuckerberg expressly states that the “list of
`
`previously used tags includes a text list 544
`
`and a friends list 546.” (Zuckerberg, 8:56-58
`
`(underlining added).) This description would
`
`have suggested, to a person of ordinary skill
`
`in the art, to store the “list of previously used
`
`tags” as a larger list with two distinct and
`
`physically separate sub-lists, i.e., a “text list” and “friends list,” as shown in the
`
`conceptual diagram at right. Under this view, each sub-list stored in the computer’s
`
`memory qualifies as a separate “tag source” even under Dr. Surati’s definition;
`
`because the two lists are physically separated, the system could access and search
`
`the text list without having to access or search the friends list, or vice versa. A person
`
`of ordinary skill in the art would have found this implementation obvious, as it is
`
`suggested directly by Zuckerberg’s statement that the “list of previously used tags
`
`
`
`15
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`017
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`includes a text list 544 and a friends list 546” (8:56-58), and the way the two sub-
`
`lists are rendered separately as shown in Figure 5. A further benefit is that this
`
`implementation would allow the developer to display the text list 544 and a friends
`
`list 546 as shown in Figure 5 by simply retrieving items from the two distinct lists
`
`in memory, avoiding the need to collate or sort items were they stored in memory as
`
`a single list of tags.
`
`27. But this would not be required in order to meet the claim properly
`
`construed. Even if one organized the list of
`
`previously used tags as a single list, with items from
`
`the “text list” and “friends list” interleaved together
`
`as shown in the image at right, there are still two
`
`distinct collections of tags – the collection of friends
`
`list tags (i.e. Bob White and Brian Bathurst shown in
`
`red), and the collection of text list tags (i.e. everything else as shown in blue). These
`
`represent two distinct collections of tags, regardless of the manner in which they are
`
`physically stored. This is because, as noted, Zuckerberg can readily recognize and
`
`distinguish tags corresponding to the “text list” from tags corresponding to the
`
`“friends list,” display these collections as two distinct lists as shown in Figure 5.
`
`(Ex. 1002, ¶¶76-79.) Nothing more is required by the claims.
`
`28. Dr. Surati states that “[t]here is no disclosure in Zuckerberg regarding
`
`
`
`16
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`018
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`how tags in ‘text list 544’ or ‘friend list 546’ are obtained, much less that they are
`
`retrieved from ‘separately searchable collections of tags,’ as reflected in
`
`BlackBerry’s proposed construction of tag sources.’” (Surati Decl., ¶119.) But the
`
`same criticism could be levied against the ’173 patent which, as explained above,
`
`does not provide any definitive details regarding storage or searching of tag sources.
`
`29. Dr. Surati also provides an illustration to show how the user interface
`
`of Zuckerberg allegedly responds if the user enters the name of a “friend” (“betty
`
`jo”) for the first time. (Surati Decl., ¶¶123-124.) But this progression does not show
`
`that the two lists are actually part of a single tag source, or suggest that they are
`
`searched as part of a single search. Zuckerberg merely states that “the list of
`
`previously used tags may be culled to include only those that match the text in some
`
`manner” (Zuckerberg, 8:53-56), and like the ’173 patent, does not specify the
`
`mechanics of how searching of the previously-used tag list is performed. But those
`
`details are not relevant to the claim limitations as explained.
`
`B.
`Zuckerberg “Dividing Line” Argument
`30. My opening declaration explained that the “dividing line” of
`
`Zuckerberg (shown in Figure 5) qualified as a “tag type indicator for each tag in the
`
`tag list,” under the theory that the claim did not require the display of a visually
`
`separate tag type indicator adjacent to each tag in the list. (Ex. 1002, ¶85.)
`
`31.
`
`I understand that after the IPR petition and my opening declaration were
`
`
`
`17
`
`Facebook's Ex. 1023
`IPR2019-00516
`
`019
`
`

`

`Reply Declaration of Sandeep Chatterjee, Ph.D.
`
`filed, the Petitioner and the Patent Owner in connection with the underlying
`
`litigation agreed that the claims of the ’173 patent do require display of a separate
`
`tag type indicator for each tag in the list. I also understand that the Board reached a
`
`similar conclusion in its Institution Decision. Accordingly, I understand that
`
`Petitioner is no longer arguing in this proceeding that the dividing line of Zuckerberg
`
`alone qualifies as a “tag type indicator” as articulated in Ground 1, but I understand
`
`that Petitioner maintains its position that the challenged claims are obvious over
`
`Zuckerberg in view of Rothmuller (Grounds 2-3) and Plotkin (Grounds 4-5).
`
`III. RESPONSE TO DR. SURATI’S OPINIONS ABOUT COMBINATION
`OF ZUCKERBERG WITH ROTHMULLER AND PLOTKIN
`(GROUNDS 2-5)
`A. Dr. Surati’s Argument That Rothmuller and Plotkin Do Not
`Disclose Distinct “Tag Sources” Is Irrelevant to Grounds 2-5
`32. Dr. Surati asserts that the tag categories in Rothmuller and Plotkin do
`
`not constitute distinct “tag sources” and thus, the category icons in Rothmuller are
`
`not “indicative of a tag source.” (Surati Decl., ¶¶131-143 (Rothmuller), ¶¶144-149
`
`(Plotkin).) I respectfully disagree that this argument provides any basis for
`
`challenging the combination of Zuckerberg with Rothmuller or Plotkin.
`
`33. My proposed combinat