UNITED STATES pATENT AND TRADEMARK OFFICE
`
`UNITED STATES DEPARTMENT OF COMMERCE
`Unit~d Scaces Pac~n l and Tradrmark Offic~
`Adctn:.s: COMMISSIONER FOR PATENTS
`P.O. Box 14SO
`.
`Aluandria. Virtinia l2JIJ.J4~
`www .u.spco,gov
`
`APPLICATION NO.
`
`FILING DATE
`
`FIRST NAMED INVENTOR
`
`ATTORNEY DOCKET NO.
`
`CONFIRMATION NO.
`
`90/011,014
`
`05/2612010
`
`5,832,494
`
`23373-00IOR.Xl
`
`6009
`
`08/0612010
`
`1590
`ALDONOTO
`DORSEY & WHITNEY
`I 330 CONNECTICUT AVE. NW, SUITE 200
`WASHINGTON, DC 20036
`
`EXAMINER
`
`ARTVNIT
`
`PAPER NUMBER
`
`DATE MAILED: 08/06/2010
`
`Please find below and/or attached an Office communication concerning this application or proceeding.
`
`PT0-90C (Rev. 10/03)
`
`001
`
`Facebook Ex. 1002 Part 2
`
`

`
`UNiTED STATES PATENT AND TRADEMARK OFFICE
`
`•
`
`comm;u•~, '" '"'""
`
`United Stales Patents and Trademark Office
`P.O .Box 1450
`Alexandria. VA 22313-1 450
`www.uspto.gov
`
`THIRD PARTY REQUESTER'S CORRESPONDENCE ADDRESS
`FISH & RICHARD SON, PC
`P.O. BOX 1022
`MINNEAPOLIS, MN 55440-1022
`
`Date:
`MAILED
`AUG 0 6 2010
`
`CIN'nW. ~INAT10N UNIT
`
`EX PARTE REEXAMINATION COMMUNICATION TRANSMITTAL FORM
`
`REEXAMINATION CONTROL NO.: 90011014
`PATENT NO.: 5832494
`ART UNIT : 3992
`
`Enclosed is a copy of the latest communication from the United States Patent and Trademark
`Office in the above identified ex parte reexamination proceeding (37 CFR 1.550(f)).
`
`Where this copy is supplied after the reply by requester, 37 CFR 1.535, or the time for filing a
`reply has passed, no submission on behalf of the ex parte reexamination requester will be
`acknowledged or considered (37 CFR 1.550(g)).
`
`002
`
`Facebook Ex. 1002 Part 2
`
`

`
`Order Granting I Denying Request For
`Ex Parte Reexamination
`
`Control No.
`
`90/011 ,014
`Examiner
`
`COLIN M. LAROSE
`
`Patent Under Reexamination
`
`5,832,494
`Art Unit
`
`3992
`
`--The MAILING DATE of this communication appears on the cover sheet with the correspondence address-·
`
`The request ,for ex parte reexamination filed 26 Mav 2010 has been considered and a determination has
`been made. An identification of the claims, the references relied upon, and the rationale supporting the
`determination are· attached.
`
`Attachments: a)D PT0-892,
`
`b)0 PTO/SB/08,
`
`c)[gl Other: PT0-1449
`
`1. [g) The request for ex parte reexamination is GRANTED.
`
`RESPONSE TIMES ARE SET AS FOLLOWS:
`
`For Patent Owner's Statement (Optional): TWO MONTHS from the mailing date of this communication
`(37 CFR 1.530 (b)). EXTENSIONS OF TIME ARE GOVERNED BY 37 CFR 1.550(c).
`
`For Requester's Reply (optional): TWO MONTHS from the date of service of any timely filed
`Patent Owner's Statement (37 CFR 1.535). NO EXTENSION OF THIS TIME PERIOD IS PERMITTED.
`If Patent Owner does not file a timely statement under 37 CFR 1.530(b}, then no reply by requester
`is permitted.
`
`2. D The request for ex parte reexamination is DENIED.
`
`T~is decision is not appealable (35 U.S. C. 303(c)). Requester may seek review by petition to the
`Commissioner under 37 CFR 1.181 within ONE MONTH from the mailing date of this communication (37
`CFR 1.515(c}). EXTENSION OF TIME TO FILE SUCH A PETITION UNDER 37 CFR 1.181 ARE
`AVAILABLE ONLY BY PETITION TO SUSPEND OR WAIVE THE REGULATIONS UNDER
`37 CFR 1.183.
`
`In due course, a refund under 37 CFR 1.26 ( c ) will be made to requester:
`a) D by Treasury check or,
`b) D by credit to Deposit Account No.
`c) o ·by credit to a credit card account, unless otherwise notified (35 U.S.C. 303(c)).
`
`, or
`
`I
`cc: Reauester I if third oartv reauester )
`U.S. Patent and Trademarlt Offoce
`PTOL-471 (Rev. 08-00)
`
`I
`
`I
`
`I
`
`Office Action in Ex Parte Reexamination
`
`Part of Paper No. 20100803
`
`003
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page2
`
`ORDER GRANTING REEXAMINATION OF U.S. PATENT 5,832,494
`
`Request for Reexamination
`
`1.
`
`A third-party Request for Reexamination of claims 1-3, S, 7-16, 18-21,23-25, and 31-33
`
`was received on S/26/2010 based on the following prior art patents and/or publications:
`
`APPENDfX.C
`
`APPENDIXD
`
`APPENDIXE
`
`APPENDIXF
`
`APPENDIXG
`
`APPENDIX I
`
`R.A. Botafogo and B. Shneiderman, !~Identifying Aggregates in Hypertext
`Structures/ 1 Hypertext '9.} ·Pr.o~eeding&, December 199-f, pp.~3-74
`("JJ.ota[Qgp_ l991'').
`
`R.A. Botafogo, "Cluster Analysis for Hypertext Systems;'' ACM-·
`SIG1R'9:3., Vol. 6, pp: 1 i 6-'125, 1'993' (''Botafogo 1993';)
`
`.HJ>. Frei and D. Stieger, '~Making Use ofHyperte.xt Links when·
`Retrieving Information," ACM, 1992 ("Frei & Sti.eger·J992")
`
`S. Baase, .Computer Algorithms: Introduction to Design and Analysis, '2nd
`Edition, Addison-Wesley Publishing Co .. , 1988 ("Baase 1988")
`
`D, Lucar:ella, "A Model for'Hypertext-Based Information Retrieval,"
`.Pro~eedings of the ECHT'9.0, Cambridge UniverSity Press,. N. Streitz, A .
`. Rizk and J. Andre, eds., ~P· 81-94, November· 1990 ("Lucarel/a 199.0")
`
`_E. Fox, "E~~ending t~e J3ool~an ~np Vectqr Sp~ce Mod_eiS· Qflnform~t~on
`Retrieval with P-Nonn Queries .and Multiple Concept Types," Comell
`University, 1983' (''Fox 1983")
`
`B.R. Schatz: and J.I3-. Hardin, ''NCSA Mosaic and. the Wor.ld Wide Web:
`·d.Iobal Hypermeqia PJ;"otoco'ts for the Internet," Sqience, Vol. 265·, Aug.
`12, 1.994 (''Schatz & Hardin 1994")
`
`APPENDIXJ
`
`U.S. Patent No. 5,838,906 to Michael D. Doyle, et al., Issued November
`17 ,. 1998 C'Doyle')
`
`004
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page 3
`
`Priority and Effective Date of the '494 Patent
`
`2.
`
`The '494 patent has a U.S. non-provisional filing date of5/17/1996 and is a continuation-
`
`in-part of application serial no. 08/076,658, filed 6/14/1993.
`
`Upon issuing a Notice of Allowability for the '494 patent, the Examiner indicated that the
`
`claims were not entitled to benefit of the 6/14/ 1993 filing date:
`
`do·not appear, and the tenn cluster. appears once in.pass.ing .. Therefore this group or'claims:.is
`
`considered to· have 5/l7/96 as their priority date for purposes. of examination in terms ofprior:art ..
`
`·Independent claims 1, 12, 14, 23, and 33 each recite limitations related to link(s), node(s),
`
`and/or cluster(s). Accordingly, these claims and all claims that depend t.herefrom have an
`
`effective date of 5/17/1996.
`
`Claip)s 18-21 do not recite such limitations and are assumed to be supported by the
`
`parent i658 application. Accordingly, claims 18-21 are given an effective date of6/14/1993 .
`
`Prosecution History of the '494 Patent
`
`3.
`
`On 9/9/1997, the Examiner issued a Restriction Requirement that broke all pending
`
`claims 1-98 into five groups ..
`
`4.
`
`On 1/9/1998, the Applicant elected Group I (claims 1-38), with traverse.
`
`005
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011 ,014
`Art Unit: 3992
`
`Page4
`
`5.
`
`On 2/3/1998, the Examiner allowed claims 1-21 and 27-38 and formed an Examiner's
`
`Amendment that canceled claims 22-26 and 39-98. [Claims 22-26 had been mistakenly included
`
`in Group I.]
`
`In addition, the Examiner issued the following reasons for allowance:
`
`The independent claims: are .addressed, in ~i$ht of the disclosure, to terms such as "indir.ect
`
`relati-onship", '~cruster· litiks", and "searching li'nkS", which are. interpreted lo refer to an explicit
`.
`.
`.
`
`use ofmuJtiple links between nod.es; The prior art of record analyzes relationshi'ps, Clusters,
`
`displays candidates for conformation, and so on, but does not anticipate. or suggest doing so with
`
`•
`
`I
`
`an explicit use of multiple links.
`
`Substantial New Question of Patentability
`
`6.
`
`The request for ex parte reexamination ("Request") alleges the following substantial new
`
`questions of patentability (SNQs) based on the above-identified prior art:
`
`l. ·claims 18-21 are unpatentable under 35 ·u.s.c .. § 102 as
`anticipated by Fox 1983.
`
`2. ClaimsJ, 2, 3, 5, 12.:-16, 18-=21, 23-25 and '31-33 are unpatentable under
`35 U.S.C. § 102 as anticipated by Botafogo 1.991.
`
`3. Claims 1, 2; 3,;5, 12-16, 23~25· and 31-33 are unpatentable under
`3S U.S.C. §.1'02 as ·anticipated by./Jotafogo J.9.93.~
`
`4. Claims.l., .2, 3,·5., 12~16, 18-21, 2·3-25 and 31.-33 are .u·npatentable ·under
`35 U.S.C. §.102 as anticipated. by Fret& Stieger1992.·
`
`5. Claims .1:, 2,.3, 5, 12-i6, 18-21, 23-25 and 31-33 are unpatentable under
`35 U.S.C. ·§ 102' ~s a.nt.i¢ip~ted by Baas.e 1988.
`
`6. .Claims :1., 2, 3,. 5,_12-16, 18-=21, 23-25 and 33 :ar.e u~patentable under.
`35 U.S.C. § 102 as anti.cipated.by Lucarella 1990 •. ·
`
`006
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page 5
`
`7; Clai.ms 7-11 are unpatentabre under 35;.u.s.c. § l03.(a) as·obvious over (i)
`!J.Q(pfogl) 1 P!Jl i11 yi~,r Qf Schatz ~· /fqrdin .1994., '(i~) Botafogo 1993. ·in y.iew o.t
`Sc#a.tz & /far4.in 19~4, (iii) .Frel &- $(ieger 1,9.92 i.n view of Sc~a!z· ~ lf..t!.rd.in.
`1994, (iv) Baase 1988 in view. of Schatz & Hardin 1994, and (v) LucareUa 1990
`in view of.Schatz & Hardin: i994.
`
`8: Claims 7-ll are unpatentable under 35 U.S;C. § l03(a).as obvious over (i)
`Botafogo.1991 in view Qf Doyle, (ii) Botafogo 1 .. 99~:in vie~ of Doyle, (Ui) .Frei
`& Stieger 19.92 in view·of Doyle, (iv.) Baase Z.988 in view of Doyle, and.(v:)
`Lucarella J-990 in view of Doyle.
`
`9. ·Claim zs·is unpatentable under 3s· u·~s.c. § l03(a) as.obvious over
`Botafogo 1993 io -view o.f Luc.arella 199.0.
`
`10. Claims 18-21 are unpatentable in view of the '352 patent based on the
`ground of double. patenting.
`
`7.
`
`·sNQ 1-Fox 1983 for claims 18-21
`
`The Request (pp. 18-25) alleges that Fox 1983, which is directed to extending the
`
`Boolean and vector space models of information retrieval with p-norm queries and multiple
`
`concept types, raises a substantial new question of patentability for claims 18-21. Specifically,
`
`the Request identifies Fox 1983 as teaching the limitations that previously distinguished the
`
`claims from the prior art during the initial examination (i.e., Fox 1983 allegedly anticipates
`
`claims 18-21 ).
`
`Since Fox 1983 appears to provide new technological teachings that were missing in the
`
`prior art during the initial examination ofthe '494 patent and was not previously considered
`
`during the initial examination, Fox 1983 does raise a substantial new question of patentability for
`
`claims 18-21.
`
`007
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page6
`
`8.
`
`SNQ2-Botafogo 1991 forclaims 1,2,3,5, 12-16, 18-21 ,23-25,and31-33
`
`The Request (pp. 2 5-41) alleges that Botafogo 1991 , which is directed to identifying
`
`aggregates in hypertext structures, raises a substantial new question of patentability for claims 1,
`
`2, 3, 5, 12-16, 18-21 ,23-25, and 31-33. Specifically, the Request identifies Botafogo 1991 as
`
`teaching the limitations that previously distinguished the claims from the prior art during the
`
`initial examination:
`
`Claim 1
`
`In allowing the claims of the '494 patent, the Office asserted that the prior art of re·c.ord
`
`did not disclose "an explicit use· of muitiple links between nodes·," noting the terms "indirect
`
`relationship," "cluster links," and "searching links," in the claims. Thus, the basis for allowance
`
`of issued claim l appears ·to have stemmed principally from lhe "generating" and "deriving"
`limitations. These LimitationS of.claim 1, however, fail to patentably distingUish over Botafogo
`1991. Indeed, Botafogo 1991 discloses these claimed features at, for example; pages 65-66, 6g,
`70 and 72. At page 72. for example, Bora fogo 1991 discloses "analyzing the structure of a
`.
`'
`hypertext, i .e., how thenodes,are linked" and."[identifying] groups of nodes that had a high
`semantic relation.'! Borafogo 1:9.9.1 further explains that "tliose. [groups of] nodeS ·should be
`
`aggregated to fonn a more abstract node." Botafogo 1991 at 72. Botafogo 1.9.91 discloses two.
`gr~ph algorithms to create aggregates; biconncctcd·and.strongly connected components:· .See
`Botafogo 1991 at, e.g., pp. 6'8-72. The:Botafogo 1991 reference thus clearly teaches all of the
`
`claimed features, including those:related to generating candidate cluster links, analyzing indirect
`
`relationships and deriving actual cluster links.
`
`008
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page 7
`
`Claim 12
`
`Notably; Bdtafogo ./991 disCloses. the, fe.arur.es of clai.m 12 ·related to generating candidate
`
`·cluster links, analyzing indirect relationships and deriving actual. cluster links, which the.PTO
`
`beli<;ved to be.m.issing from the prior. art.ofrecord during examination of the claims at issue. As
`
`.explained above in connection with claim I, Botafogo 1991 discloses analyzing how nodes are
`
`linked, identifying groups of node8 having a relatioi1, and aggregating nodes using graph
`
`algorithms. See·Botafogo 1991 at, e.g., 65-66, 68, 70, 72.
`
`Claim 14
`
`In particular, Bolafogo 1991 discloses the features of clai.m 14 related to "an explicit .use
`
`ofrnultipl~ liriks," which the PTO believed to be rnis~ing from the prior art of record during
`examination of the claims at issue. As explained above, Borafogo 1991 discloses analyzing how
`nodes are liti.Ked, identifying groups· of nodes having a relatio~ and aggregating nodes to foi'rn an
`aggregate ·node using graph algorithms. See Botafogo 1991 at, e.g., pp: 65-66, 68, 70·, 7~.
`
`Botafogo 1991 includes discussion regarding "the number of nodes and links and ... their ratio"
`
`and a "compactness" metric that indicates the interconnectedness of a hypertext, noting that "a
`
`too ·high cbmpactnes.s means that each node has inimy links." See 11otafogo 1991 at 65. Bo'tafogo
`
`1991 also discusses "pa·ths" between nodes, stating that ''the more paths there are .between .any
`
`two nodes the stronger their refation." I d. at 66.
`
`009
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Claim 18
`
`Page 8
`
`As detailed in Exhibit 2, Botajogo 1991 discloses each and every element of claim 18 and
`thus ant.icipates·thtsdaim. Botafogo 199i dis.clos.es a method.ofllll!ilyzing a database having
`objects (e.g:., an ·article) and a fir-st .numerical representation "(the hypertext •'waph")""of direct
`telatiorisbi:ps ill the database (see, e.g., pp. 64-66),. comprising the st~ps of:
`generating a second numerical representation. (representation· of the "aggregate")
`·using the jim numerical representation (the "graph"), wlterem the second numerical
`representation. accounts for indirect relationships in the database (As discussed above,
`Botafogo 19.9i determines indirect.relationships. See, e.g., pp. 64-66.);
`storing the second·numer.ical representation (see, e.g., pp. 64-66);
`
`identifying at leastone·object (e.g., an ar.ticle) in the database~ ;vherein the Stored
`.i),;me.ric(lJ representation. is u.sed to identify objects (see, e.g., pp; 64-66, 68, 70); apd
`4ispla)!inf( one ·or -more i(fentifred objects ft:om the datpb{lSe (see, e.g., pp, 71-72,
`inCluding.Botafogo 1991's discussion·regarding a "user interface").
`
`Claim 23
`
`In pauicular, Botajogo 1991 discloses features related to· "an .explicit usc of multiple
`links," which .the PTO bcl.i~vcd to QC missing frolll .the. prior ~rt. of record during c.xa~&tion of
`the cl~ims at issuG . . As prcvi.o~~ly notc"d, Bo{ajogQ J 991 disclo$cs at:talyzing 4ow no·dcs :a.rc:
`lirikcd, identifying woups·of.nodcs having a relation, and aggrcgatin~ nodes to form an
`abs~action using graph algorithms. See 'Botafogo 1991 at, e.g., pp. 65-66, .. 68, 70, 72. Botafogo
`1991 includes discussion regarding "the number·of nodes and Links and . , , their ratio" and a
`"cpmp.actness11.metric th~t. indicates the interconn~ctedness of a hypertext? no.ting th!lt "a too high
`compactnes~ m~ans that each node has.many links." SeeBotajogo 1991 at 65 . .Botajqqo 199.f
`
`also discusses "paths" between nodes,. stating that ''the more.paths there are between any two
`r:10d.cs. th.c &trp~gcnhcir r¢ia~ion;" f(J.. at Q6 ..
`
`010
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011 ,014
`Art Unit: 3992
`
`Page 9
`
`Claim 33
`
`As detaile9 in Exbibit.2, BotajQgo 1991 cj.isclo$es a method of representing data in a
`
`computer database and for. computerized searching of the data, wherein relationships exist ·in the
`database (see, e.g., pp. 64-66), comprising:
`assigning links to represent relmionships·in the dalabase (For example·, Botafogo 1991
`discloses a hypertext "graph:" Each link in the grapii_represents a rel(ltionship betwee.n two
`nodes. See, e.g., pp. 64-66.);
`generating node identijicalions based upon the assigned links, wherein node
`identifications. are g(!nerated so thai each link represents a relationship between two identified
`nodes (As discussed above, Borafogo 1991 discloses generating node i~eqtifications tha~ are
`based on the links between the nodes. See, e.g., pp. 64-66.);
`storing the links and node identifications, wherein the links and nodes may be retrieved
`(As discussed above; Botafogo I 99 I analyzes links and nodes. These links and nodes are stored
`for analysis and retrievaL See, e.g .• pp. 64-66.);
`
`.'iearchingjor node identifiCations using the stored links (For example, Botafogo 1991
`teaphes "identify[ing) semuntic ~lusters in a hrJ>ertext:" Further, Botafogo 1991 explai~ that
`"[b ]y analyzing the structure of a hypertext ... it was possible to identify groups of nodes."
`
`Borafogo 1991 further teaches clustering of nodes and-discloses a clustering algorithm to locate a
`subsetofno.des. See,e,g.;pp.64-66;68, 70, 71-72.);and
`displaying node identifications, wherein the displayed node identifications are located
`in the .'iearching step (see, e.g., pp. 71-72, including Botafogo 1991's discussion regarding a
`
`"user interface").
`
`As indicated above, Botafogo /991 discloses each and every element of claim 33,
`inCluding those the PTO believed to be missing from the· prior urt of record-during examination
`ofthe· ~pplication that matured i"nb,) th~ '494 patent: The Botafogo 1.991 ref~rence therefore
`
`anticipates claim 33 .
`
`. Since Botafogo 1991 appears to provide new technological teachings that were missirig in
`
`the prior art during the initial examination of t~e '494 patent and was not previously considered
`
`011
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011 ,014
`Art Unit: 3992
`
`Page 10
`
`during the initial examination, Botafogo 1991 does raise a substantial new question of
`
`patentability for claims I, 2, 3, 5, 1.2-16, 18-21,23-25, and 31-33.
`
`9.
`
`SNQ 3- Botafogo 1993 for claims 1, 2, 3, S, 12-16,23-25, and 31-33
`
`The Request (pp. 41-55) alleges that Botafogo 1993, which is directed to cluster analysis
`
`for hypertext systems, raises a substantial new question of patentability for claims 1, 2, 3, 5, 12-
`
`16, 23-25, and 31-33. Specifically, the Request identifies Botafogo 1993 as teaching the
`
`limitations that previously distinguished the claims from the prior art during the initial
`
`·examination:
`
`Claim 1
`
`A~ noted above, the basis for the PTO 's allowance of issued claim 1 appears to have
`
`stemmed principally from the "gen·era.ting" and "deriving" limitations. These limitations of
`claim 1, however, fail to patentably distinguish over Botafogo 1993. Indeed, Botafogo 1993
`
`disclos~s these claimed features at, for example, pages I 17-IIR, 12 1-122. More.panicularly,
`
`Botafogo 1993's "~apn" corre5ponds to candidate cluster links for the selected nodes. Botafogo
`1993 teaches analyzing indirect.relationships fork> 1, where k is the distance between nodes.
`These relationships are used to define k-components and correspond to "a maximal induced
`subgrapb <A> ... with· th.e property that for every partition A1, A2 o( <A> at least k links of <A>
`are each incident with an object of A1 and of A2." Botafogo 1993 at 117-118. Botafogo 1993
`
`teaches deriving an "aggregate" from the candidate cluster links. Page 118 of the reference
`
`explains bow the aggregate can be id~tiJied by removing links (see Definition 4· reproduced
`
`below).
`
`012
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page 11
`
`Claim 12
`
`Notably,_ Botafogo 1993 discloses the features of claim 12 related to generating candidate
`clusw liriks,. ana~yzing indirect relati!)nships an~ ~~riVing actual ch,tster links,.which 'the PTO
`believed to be missing from the prior urt of record during ex~tm.imdion oftlie cl~tims ut.issue. As
`cxpl~tined ubovc, Botafogo 1993's "gmph" corresponds to ctmdid~ttc cluster links. Botafogo·
`1993 teaches analyzing indirect relationships for k.:<>l, Where k is the distance between nodes.
`These rcla~ipnsh.lps arc .used to define k-<:o,mponcnls ·and correspond to ''a maxiQ'lal indu~cd
`subgraph -<A> ... with the. property that.for every partition A 1, A2 of <A'? at least k links of <:A>
`are each incident with an object of A1 and of A2." Botafogo 11/93 at 177-11 H.
`
`Claim 14
`
`As dct~tilcd in Exhibit3', Botafogo 1993 discloses ll method for representing the
`relationship ~tween nodes ~sing ~tored direct links, paths, and c~ndi@t.e cluster links (see, e.g.,
`
`pp. 116-118, 1-21-122),. comprising the steps of:
`a) initializinx a :,·et ofcandidate Clu!iter links (see, e.g., pp. 117-118, 121-122- As
`discussed above, Botafogo 1993 initializes candidate cluster links.);
`b) selecting the destination node of a path as the selected node to.analyze (see, e.g., pp.
`116-118, 121-122- As discussed ~tbovc, Botafogo 7993. selects ll node to ~tnlllyze. Further,
`Botafogo 1.993 explains that "the property we wiil Q&e to identify the ch~sters is th~ .nwnber of
`independent paths between nodes.");
`
`c) retrieving the set of direct/inks from the sefected node to any other node in the
`.database (see, e.g ..• pp. 118-ri9~ .IZl-.122 -- Bota/ogo .1993 discusses apply.ing.its method on
`:bypcrtc~ts ·stru~rurcd. as a grap}:l, wi tb )'i.nk.s COD!lC~t~g ~odes. See. pag~ 1'17 .. Fut:tper, Botafogo
`1993 discloses identifying liriks when determining k-components . . See, e.g., pp . .118 ("A k(cid:173)
`componeht of a graph G as a maximal -induced subgraph <A> with the property that for every
`partition A1. A2 of <A>, at. least k links of{.t\) an, each incident with an object of A-1 and af A2•
`. . IdentitY all th·c k-componcnts of the graph. Figure 2 shows an :example"); sec also pp.-119-
`1 21, including dis~ussion of aggregates that have hypertext nodes anq links; Fig. 2. Botafogo
`1993 also explains that determining the links are.necessary clustering because "links indicate a
`strong relationship bet)Neen the nodes they connect.");
`
`013
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page 12
`
`d) determining the weight oftlie path using the retrieved. direct links (§ee,. e.g., pp. 117-
`
`111:! ("In hypertext systems, in order to ·identify clusters we will use a measure of similarity
`
`oetween.nodes that is readily available: .links indicate a strong relationship between the no4es
`they connect. Also a path betw.een nodes ·indicates a relationship between them. Intuitively, the
`
`more there are independent p·athS between nodes, the more those nodes are related, Many of the
`existing cllLstcring algorithms, such as single-linkage, weak and strong k-linkagc, and k-ovcrlap
`[21, 14j, arc ba~~d qn the ngttpn.-o(thc i,ntc.nsHy of qmncctign bc~~9Jl no.~~,:s jn !l b~~h,
`Hypertext systems, given their s·tructure ·as· directed graphs, lend .themselves very naturally to be
`analyzed in· this way .... First, in a. hypertext represented as a ·directed graph, the strength. of
`connection as indicated by the number of independent paths between nodes is a natural
`
`indication of the strength of relationship between those nodes.");
`
`repeating steps b through d/or·eachpi!fh (see, e.g., pp. 117-119, 121-122); and
`
`c) storing the determined weights as candidate cluster links (see, e.g., pp. 117-118).
`
`As indicated above, Sotafogo .1993 discloses each and every element of claim 14,
`iitcluding those the PTO believed to be missing from the prior ltrt of-record during cxumi'nution
`ofthe application that maturet;i in.to the.' 494 pa~ent. Bntafogo 1993 thus anticipate~ claim 14
`under section I 02.
`
`014
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Claim 23
`
`Page 13
`
`As detajle~t.ln Exhjbit 3, Boiafogo 199 3 discloses a method pf representing· data in a
`computer. database with relationships (see, e.g., pp. 117-118, 12i-.122), comprising the steps of:
`
`assigning nodes node identifications [see, e.g., page 116 ("Using the graph theoretic
`def!rdtiowofk-edge-co1J1po.nen~ we pres~nt an am,egat.i.on tecbni.qu,e to cluster tbe n.odes, We
`then use this technigues to cluster the. nodes."), page 117 ("First, in a. hypertext represented as a
`di~cctcd graph, the. strength of connection as indicated by the number of independent paths
`b.etween nodes is· a natural indication of the strength of relationship between those nodes."), and
`pages 119-21);
`
`ge~er(l#ng l~k.'l'~ wherein each link represent!f a relationship fleljt•ef?n hv(} nod~s and it{
`identified by the two nodes.in which the relationship exists [see, e.g., pp. 117-118 (''Jn
`hypertext systems, in order to identify clusters we will use a measu~e of similarity between nodes
`th~t is readily available: links indicate a. strong relationship between .the nodes they colinect. Also
`a path between no.des indicates a relationship between them. , .. Fit:st, in a_ hypertext represented
`as a directed graph, the strength of connection. as iQdicated by the number of' independent paths
`between nodes is a natural indication of the strength of relationship between those rio des.''), 1.21-
`
`122.);
`
`015
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/01 1,0 14
`Art Unit: 3992
`
`Page 14
`
`allocating a weight to each link, w.herein the weight signifies.the strength of the
`relationship represented by tlte link reiative· to 'lite. strength ofoiher relationships represented
`by. other links [st'e, e.g., pp. 117-118 (Page 11'7: "In.liypertext systems, in order to 'identify
`
`cll!Ste.rs we will use a measur:e of similari,ty between nodes that is ~adily available: links indicate
`a stror;tg reia~ionstlip between the n,Qd~ tijey ~o~~ct. Also a p~th between nodes indicate.s a
`relationship between them. Intuitively, the more there are ind~pendent paths between nodes, the
`
`mor.c those nodes arc rchitcd, Many·ofthc·cxisting clustering algorithms, such as single-linkage,
`
`weak and strong k-linkage, and k-o.verlap.[21, 14], are based on the.notionofthe··intensity of
`oonncction between nodes in a graph. Hypertext systems, given their structure as .dircctcd graphs,
`
`leQd themselves very naturally to be analyzed in this way . ... First, in a hypertext represept~d as
`
`a directed gr~ph, the strength of connection as indicated by the number of independent paths
`
`'between nodes is a natural indication of the strength of relationship between those nodes.") (Page
`118: "Defmition 4. An aggregate in .the graph is a set of connected nodes which are in a k-
`
`componeni but in no (k+ i)-component. ... ln order to find the aggregates practicaUy, the
`following st~ps are done ... The graph theory literature provides alf the tools to find the k(cid:173)
`components with reasonable efficiency."), pp. 121- 122.]; and
`
`displaying a node identification (see, e.g., pp. 1 1'6, Iil-122) (Botafogo 1993 teaches
`disp laying a ·node identification in the context of an information retrieval system, ie., to the user.
`
`Botafogo /.993 explains that is methods are used to "improve data display, browsing and
`retrieval." See Botafogo 1993, e.g., pp. 116.).
`As indi~ated abl;)ve, Botafogo .J 993 discloses each and every element.of claim 23 .•
`
`including those the PTO believed to be missing from the. prior-art of record during examination
`
`of the application that'rh atufed into the ?494 patent. Botafogo 1993 thu.S anticipates claim 2J
`
`under section 102.
`
`016
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page 15
`
`Claim 33
`
`A$ ~etail~d ~ -E_x.l:tib1t 3, Bc!,i{ljogo 199.-3 di_s<;)Qo$eS ~ach:_and ~Yt<ry element of ci.~l111 ~3 and
`thus anticipates this· claim. Indeed, Bot~(ogo i993 discloses a method of representing ~ala in a
`eoniputer database ·and for computerized searching ofthe data, wherein relationshi_ps exist- in· the
`database (see; e.g., pp.117--118, 121-122), comprising:
`assigning links to represent 'r.e/ationships in the=database.(Botafogo 199.3 teaches
`assigning links based on the structure of a hypertext, and it.is implicit in authoring of"hypertext.
`Page 117 of the reference explains: "In hypertext systems, in order to identify c-lusters-we will
`
`use a measure ofsimilarity between nodes that is readily available: links indicate a strong
`relationship between the nodes they connect" .and ·that "a. link needs to represent a relationship
`bet\\(een nodes." See, e._g., pp . .ll7~ 118, 121-122.);
`
`gener.atihg no.de identifications based upon the assigned links, wlterein-node
`identifications are generated so tit at each 1ink represent's a ·relationship between two identified
`nodes (ln Figure '2 of Botafogo 1993, each node is identified by a letter, and the letters are
`
`ordered based oh the layout of the-link graph-beginning of the alphabet on the left, and end of
`
`.the alphabet on the right. See, e.g., pp. 117-118, 121-122);
`
`storing the links and node.identifications, wlterein tlte links and nodes may be retrieved
`
`-(Bo_tafogo 1993 teaches storing link,s and nodes and using these links· and. nodes for its·
`algorithms. See, e.~:, pp. 117-118, '121-122.);
`~;earc:hingfor node identijicat.ioiti' using the.tJ·tored lin/is (Botafogo 1993 at page 122
`
`teaches that its algorithms can "help .in an information tcfricval system." Botafogo. /9.93 te!lches
`
`that, when a user is at node 'h, link information is used to, for. example, ''retumi] nodes ~- f, g, d.""
`See, e.g., pp. 117-118, 121-122-.); and
`
`displaying node identifications, wherein the displayed node identifications are located
`in ·tile sear.c:hing step (JJoiafogo 199 3 teaches returning the node identifications in the contex't of
`an information re_trieval system, i.e., to the use~:. Botafogo 19,93 explains that is methods are
`used. to "improve data-display, browsing and retrieval.'_' See, e.g., pp: 1'16, '121-.122.).
`As· indicated ·aoo:ve, Bota]ogo 199 3 discloses each and. every element of-claim 33,
`inCluding-those the PTO believed to be missing: fro·m the _prior art. of record during examination
`
`of the applica~i~n t}lat-ml).tured !nto .the '494 patent. Bo_tafogo 1:99.3 'thus a.ntjcip~es cl~im 33
`under section 102.
`
`017
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page 16
`
`.Since Botafogo 1993 appears to provide new technological teachings that were missing in
`
`the prior art during the initial examination of the '494 patent and was not previously considered
`
`during the initial examination, Botafogo 1993 does raise a substantial new question of
`
`patentability for claims 1, 2, 3, 5, 12-16, 23-25, and 31-33 .
`
`10.
`
`SNQ 4-Frei& Stieger 1992 for claims 1, 2, 3, 5,12-16, 18-21,23-25, and 31-33
`
`The Request (pp. 56-76) alleges that Frei & Stieger 1992, which is directed to making use
`
`of hypertext links when retrieving information, raises a substantial new question of patentability
`
`for claims 1, 2, 3, 5, 12-16, 18-21, 23-25, and J 1-33. Specifically, the Request identifies Frei &
`
`Stieger 1992 as teaching the limitations that previously distinguished the claims from the prior
`
`art during the initial examination:
`
`Claim 1
`
`As previously noted, the basi·s for the PTO's ullo\vuilcc of iSsued cluirri .I .upp·e~rs fo hllvc
`stemmed principally from. the "generating" and. "denv.ing1' limitations. These. limitations of
`claim 1, however, fails to patentably distinguish over Frei & Stieger 19.92. Indeed, Frei &
`
`.Stieger 19C)2 disCloses these claimed features at sections 2 and 3 of the reference and, for
`example, at pages 102-106. At p·~ge 102, Frei & Stiegef 1992 discu:~ses links and·node.i:o, a.-;· well
`
`as "exploiting the links when specific content-related infonnation is to. be re.trieved." Fig. 1 of
`
`the reference (reproduced below) shows. an example of a hypermedia collection, which includes
`
`nodes (e.g., .nl) , Sl,.lb-nodes (e.g., nl_. l), and .links (sho~n by the .lin~ and arrows) connecting·the
`
`various nodes.
`
`018
`
`Facebook Ex. 1002 Part 2
`
`

`
`Application/Control Number: 90/011,014
`Art Unit: 3992
`
`Page 17
`
`(Fliii & Stieger /992 - .Page 103)
`Frei .& Stieger 1992 de~cribes indexing nodes and.link~, including indexing functions
`that "take informat