Trials@uspto.gov
`571-272-7822
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
` Paper 18
`
`
` Entered: April 27, 2015
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`HUGHES NETWORK SYSTEMS, LLC and
`HUGHES COMMUNICATIONS, INC.,
`Petitioner,
`
`v.
`
`CALIFORNIA INSTITUTE OF TECHNOLOGY,
`Patent Owner.
`_______________
`
`Case IPR2015-00068
`Patent 7,116,710 B1
`____________
`
`
`Before KALYAN K. DESHPANDE, GLENN J. PERRY, and
`TREVOR M. JEFFERSON, Administrative Patent Judges.
`
`JEFFERSON, Administrative Patent Judge.
`
`
`DECISION
`Denying Institution of Inter Partes Review
`37 C.F.R. § 42.108
`
`
`
`
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`I.
`INTRODUCTION
`Hughes Network Systems, LLC and Hughes Communications, Inc.
`(collectively, “Petitioner”) filed a Corrected Petition requesting an inter
`partes review of claims 1, 3, 4, 5, 6, 15, 16, 20, 21, and 22 of
`U.S. Patent No. 7,116,710 B1 (Ex. 1001, “the ’710 patent”). Paper 4
`(“Pet.”). California Institute of Technology (“Patent Owner”) timely filed a
`Preliminary Response. Paper 13 (“Prelim. Resp.”). We have jurisdiction
`under 35 U.S.C. § 314(a), which provides that an inter partes review may
`not be instituted “unless . . . there is a reasonable likelihood that the
`petitioner would prevail with respect to at least 1 of the claims challenged in
`the petition.” After considering the Petition, the Preliminary Response, and
`associated evidence, we conclude that Petitioner has not demonstrated a
`reasonable likelihood that it would prevail in showing unpatentability of all
`the challenged claims. Thus, we deny institution of an inter partes review of
`claims 1, 3, 4, 5, 6, 15, 16, 20, 21, and 22 of the ’710 patent.
`A. Related Proceedings
`Petitioner indicates that the ’710 patent is the subject of the
`
`proceedings in California Institute of Technology v. Hughes
`Communications, Inc. et al., No. 13-cv-07245 (C.D. Cal.). Pet. 1–2.
`The ʼ710 patent is also the subject of IPR2015-00067. Additionally,
`Petitioner indicates that the ʼ710 patent is related to U.S. Patent No.
`7,421,032, U.S. Patent No. 7,916,781, and U.S. Patent No. 8,284,833, which
`are the subject of IPR2015-00059, IPR2015-00060, IPR2015-00061, and
`IPR2015-00081. Paper 7, 1–2.
`
`
`
`2
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`B. The ʼ710 Patent
`The ’710 patent describes the serial concatenation of interleaved
`convolutional codes forming turbo-like codes. Ex. 1001, Title. It explains
`some of the prior art with reference to its Figure 1, reproduced below.
`
`
`Figure 1 is a schematic diagram of a prior “turbo code” system. Id. at 2:14–
`15. The ’710 patent Specification describes Figure 1 as follows:
`
`A standard turbo coder 100 is shown in FIG. 1. A block
`of k information bits is input directly to a first coder 102. A k
`bit interleaver 106 also receives the k bits and interleaves them
`prior to applying them to a second coder 104. The second coder
`produces an output that has more bits than its input, that is, it is
`a coder with rate that is less than 1. The coders 102,104 are
`typically recursive convolutional coders.
`Three different items are sent over the channel 150: the
`original k bits, first encoded bits 110, and second encoded bits
`112. At the decoding end, two decoders are used: a first
`constituent decoder 160 and a second constituent decoder 162.
`Each receives both the original k bits, and one of the encoded
`portions 110, 112. Each decoder sends likelihood estimates of
`the decoded bits to the other decoders. The estimates are used to
`decode the uncoded information bits as corrupted by the noisy
`channel.
`
`
`
`3
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`Id. at 1:38–53 (emphasis omitted).
`A coder 200, according to a first embodiment of the invention, is
`described with respect to Figure 2, reproduced below.
`
`
`Figure 2 of the ’710 patent is a schematic diagram of coder 200. Id. at2:16–
`17.
`
`The Specification states that “coder 200 may include an outer coder
`202, an interleaver 204, and inner coder 206.” Id. at 2:34–35 (emphasis
`omitted). It further states as follows.
`
`The outer coder 202 receives the uncoded data. The data
`may be partitioned into blocks of fixed size, say k bits. The
`outer coder may be an (n,k) binary linear block coder, where
`n>k. The coder accepts as input a block u of k data bits and
`produces an output block v of n data bits. The mathematical
`relationship between u and v is v=T0u, where T0 is an n x k
`matrix, and the rate1 of the coder is k/n.
`The rate of the coder may be irregular, that is, the value
`of T0 is not constant, and may differ for sub-blocks of bits in the
`data block. In an embodiment, the outer coder 202 is a repeater
`that repeats the k bits in a block a number of times q to produce
`a block with n bits, where n=qk. Since the repeater has an
`irregular output, different bits in the block may be repeated a
`
`1 The “rate” of an encoder refers to the ratio of the number of input bits to
`the number of resulting encoded output bits related to those input bits. See
`Ex. 1010 ¶ 19.
`
`
`
`4
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`different number of times. For example, a fraction of the bits in
`the block may be repeated two times, a fraction of bits may be
`repeated three times, and the remainder of bits may be repeated
`four times. These fractions define a degree sequence or degree
`profile, of the code.
`The inner coder 206 may be a linear rate-1 coder, which
`means that the n-bit output block x can be written as x=TIw,
`where TI is a nonsingular n x n matrix. The inner coder 210 can
`have a rate that is close to 1, e.g., within 50%, more preferably
`10% and perhaps even more preferably within 1% of 1.
`Id. at 2:41–64 (emphasis omitted and footnote added). Codes characterized
`by a regular repeat of message bits into a resulting codeword are referred to
`as “regular repeat,” whereas codes characterized by irregular repeat of
`message bits into a resulting codeword are referred to as “irregular
`repeat.” The second (“inner”) encoder 206 performs an “accumulate”
`function. Thus, the two step encoding process illustrated in Figure 2,
`including a first encoding (“outer encoding”) followed by a second encoding
`(“inner encoding”), results in either a “regular repeat accumulate” (“RRA”)
`code or an “irregular repeat accumulate (“IRA”) code, depending upon
`whether the repetition in the first encoding is regular or irregular.
`Figure 4 of the ’710 patent, reproduced below, shows an alternative
`embodiment in which the first encoding is carried out by a low density
`generator matrix.
`
`
`
`5
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`
`
`Figure 4 of the ’710 patent is a schematic of an irregular repeat and
`accumulate coder using a Low density generator matrix (LDGM)2 coder. Id.
`at 2:20–21, 3:25. The LDGM coder “performs an irregular repeat of the k
`bits in the block, as shown in FIG. 4.” Id. at 3:52–54. LDGM codes are a
`special class of low density parity check codes that allow for less encoding
`and decoding complexity. LDGM codes are systematic linear codes
`generated by a “sparse” generator matrix. No interleaver (as in the Figure 2
`embodiment) is required in the Figure 4 embodiment because the LDGM
`provides scrambling otherwise provided by the interleaver.
`C. Illustrative Claim
`Petitioner challenges claims 1, 3, 4, 5, 6, 15, 16, 20, 21, and 22 of the
`’710 patent. Pet. 3–4. Claim 1 is illustrative of the claims at issue and is
`reproduced below:
`1. A method of encoding a signal, comprising:
`obtaining a block of data in the signal to be
`encoded;
`
`
`2 A “generator” matrix (typically referred to by “G”) is used to create
`(generate) codewords. A parity check matrix (typically referred to by “H”)
`is used to decode a received message.
`6
`
`
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`partitioning said data block into a plurality
`of sub-blocks, each sub-block including a plurality
`of data elements;
`first encoding the data block to from a first
`encoded data block, said first encoding including
`repeating the data elements in different sub-blocks
`a different number of times;
`interleaving the repeated data elements in
`the first encoded data block; and
`second encoding said first encoded data
`block using an encoder that has a rate close to one.
`
`
`
`D. The Alleged Grounds of Unpatentability
`The information presented in the Petition sets forth proposed grounds
`of unpatentability of claims 1, 3, 4, 5, 6, 15, 16, 21, and 223 of the ’710
`patent as follows (see Pet. 14–43):
`
`References
`Divsalar3 and Luby4
`Divsalar, Luby, and Hall5
`Divsalar, Luby, and Ping6
`Divsalar, Luby, Ping, and
`Hall
`
`Basis
`§ 103(a)
`§ 103(a)
`§ 103(a)
`
`§ 103(a)
`
`Claim(s) Challenged
`1, 3, 4, 5, 6, 15, 16, 21, and 22
`15, 16, 21, and 22
`20
`
`20
`
`
`3 Dariush Divsalar, et al., Coding Theorems for “Turbo-Like” Codes,
`THIRTY-SIXTH ANNUAL ALLERTON CONFERENCE ON COMMUNICATION,
`CONTROL, AND COMPUTING, Sept. 23–25, 1998, at 201–209 (Ex. 1011,
`“Divsalar”).
`4 U.S. Patent No. 6,081,909, issued June 27, 2000 (Ex. 1016, “Luby”).
`5 Eric K. Hall, et al., Stream-Oriented Turbo Codes, 48TH IEEE VEHICULAR
`TECHNOLOGY CONFERENCE 1998 at 71-75 (Ex. 1013, “Hall”).
`6 L. Ping, et al., Low Density Parity Check Codes with Semi-Random Parity
`Check Matrix, 35 ELECTRONIC LETTERS 38–39, 1999 (Ex. 1014, “Ping”).
`7
`
`
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`II. ANALYSIS7
`A. Claim Construction
`The Board will interpret claims of an unexpired patent using the
`broadest reasonable construction in light of the specification of the patent in
`which they appear. See 37 C.F.R. § 42.100(b); Office Patent Trial Practice
`Guide, 77 Fed. Reg. 48756, 48766 (Aug. 14, 2012). Under the broadest
`reasonable construction standard, claim terms are given their ordinary and
`customary meaning, as would be understood by one of ordinary skill in the
`art in the context of the entire disclosure. In re Translogic Tech. Inc.,
`504 F.3d 1249, 1257 (Fed. Cir. 2007).
`Petitioner discusses several claim terms and the District Court’s
`construction of those terms, but does not offer proposed constructions. Pet.
`12–14. We determine that no claim construction is necessary for the
`purposes of this decision.
`B. Divsalar (Ex. 1011) and Luby (Ex. 1016)
`Petitioner contends that claims 1, 3, 4, 5, 6, 15, 16, 21, and 22 are
`unpatentable under 35 U.S.C. § 103(a) over Divsalar and Luby (Pet. 14–33)
`and that each of the remaining challenged claims 15, 16, 20, 21, and 22 are
`unpatentable under 35 U.S.C. § 103(a) as obvious over Divsalar and Luby in
`combination with other asserted prior art (Pet. 34–43).
`
`
`7 Patent Owner argues that, as a threshold matter, the Petition should be
`dismissed because Petitioner fails to identify all real parties in interest.
`Prelim. Resp. 3. Because we have determined that Petitioner has not
`demonstrated a reasonable likelihood of prevailing, we need not address the
`real parties in interest issue in this Decision.
`8
`
`
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`1. Divsalar (Ex. 1011) as a Printed Publication
`Petitioner states that Divsalar was “published no later than April 30,
`1999 at the University of Texas library.” Pet. iii; see also Pet. 2 (stating that
`Divsalar was “published at least by April 30, 1999 and available as prior art
`under 35 U.S.C. § 102(b)”). In support, Petitioner proffers the declaration
`testimony of a Univ. of Texas librarian (Ex. 1064) including an acquisition
`record pasted into an email. According to Petitioner’s expert, Dr. Henry D.
`Pfister, The Allerton Conference is generally regarded as one of the main
`conferences in the field of information theory and communications. Ex.
`1010 ¶ 28.
`Patent Owner argues that Petitioner has not established that Divsalar
`is a printed publication within the meaning of § 311(b). Prelim. Resp. 21–
`23. Patent Owner states that the acquisition record of the University of
`Texas library does not state that the paper was actually shelved or otherwise
`displayed and accessible to those of “ordinary skill.” Prelim. Resp. 22–23.
`According to the Divsalar cover page, it was presented at the Allerton
`Conference held on September 23–25, 1998. The acquisition record of the
`University of Texas indicating acquisition in April, 1999 lends credence to
`the actual presentation and publication of the paper at the September 1998
`Allerton Conference and dissemination of the paper to the interested public.
`See Ex. 1064 (Declaration of Robin Fradenburgh) 4–6. Given Dr. Pfister’s
`testimony that the Allerton Conference is the premier conference for
`information theorists, we find sufficient evidence to establish Divsalar as a
`printed publication within the meaning of the AIA statute.8
`
`
`8 We also note that Divsalar is listed as being of record among the
`“References Cited” in U.S. Patent No. 7,916,781 (the ’781 patent) that is the
`9
`
`
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`Accordingly, we are persuaded by Petitioner and the supporting
`evidence that Divsalar is prior art for the purposes of this Decision. Patent
`Owner may rebut Petitioner’s contentions and supporting evidence with
`evidence that Divsalar was not presented and published at the Allerton
`Conference.
`2. Divsalar (Ex. 1011)
`Divsalar discloses “turbo-like” coding systems that are built from
`fixed convolutional codes interconnected with random interleavers,
`including both parallel concatenated convolutional codes and serial
`concatenated convolutional codes as special cases. Ex. 1011, 3. With fixed
`component codes and interconnection topology, Divsalar demonstrates that
`as the block length approaches infinity, the ensemble (over all possible
`interleaves) maximum likelihood error probability approaches zero, if the
`ratio of energy per bit to noise power spectral density exceeds some
`threshold. Id.
`The general class of concatenated coding systems is depicted in
`Figure 1 of Divsalar as follows:
`
`
`
`subject of IPR2015-00059. It was cited by the Applicant in an IDS
`apparently filed with the application on June 30, 2008.
`10
`
`
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`Figure 1 illustrates that encoders C2, C3, and C4 are preceded by
`interleavers (permuters) P2, P3, and P4, except C1, which is connected to an
`input rather than an interleaver. Id. at 4–5. The overall structure must have
`no loops and, therefore, is called a “turbo-like” code. Id.
`Divsalar further discloses that “turbo-like” codes are repeat and
`accumulate (RA) codes. Id. at 7. The general scheme is depicted in Figure
`3 as follows:
`
`
`Figure 3 illustrates that information block of length N is repeated q times,
`scrambled by interleaver of size qN, and then encoded by a rate 1
`accumulator. Id. The accumulator can be viewed as a truncated rate-1
`recursive convolutional encoder. Id. Figure 3 further illustrates a simple
`class of rate 1/q serially concatenated codes where the outer code is a q-fold
`repetition code and the inner code is a rate 1 convolutional code with a
`transfer function 1/(1+ D). Id. at 3, 7.
`3. Luby (Ex. 1016)
`Luby discloses a technique for creating loss resilient and error
`correcting codes having irregular graphing between the message data and the
`redundant data. Ex. 1016, 1:5–9. Luby teaches a technique for creating
`encoded messages, which when decoded, facilitate recover and/or correcting
`
`
`
`11
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`of message data that has been lost or corrupted during transmission or
`storage. Id. at 2:56–60.
`4. Analysis – Divsalar and Luby (independent claims 1 and 15)
`Independent claim 1 recite encoding that required “repeating the data
`elements in different sub-blocks a different number of times.” In addition
`irregular repeating, claim 1 further recites partitioning the data block into
`sub-blocks.
`Although Divsalar teaches obtaining a block of data (Pet. 14),
`Petitioner admits that “Divsalar teaches a method that includes repeating
`each input bit the same number of []times” and not an irregular or different
`number of times. Pet. 16 (citing Ex. 1011 at 5; Ex. 1010 ¶ 118); see also Pet.
`24 (stating that Divsalar “does not ‘repeat said stream of bits irregularly,’ as
`required by claim 15”). Petitioner relies on Luby to teach the irregular
`repeating recited in claims 1 and 15. Pet. 24–25. According to Petitioner,
`this is demonstrated by Luby’s Figure 17, reproduced below.
`
`
`Figure 17 depicts an irregular graphing of the edges between node layers in
`an error correcting cascading encoding structure. Ex. 1016, 5:42–46.
`According to Petitioner, the circles (nodes) in the left column represent
`
`
`
`12
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`information bits to be encoded and the circles (nodes) in the right column
`represent parity bits computed from the information bits. Pet. 17 (citing
`Ex. 1010 ¶ 119). Each parity bit on the right is computed by summing
`together (modulo 2) all of the information bits connected to that parity bits
`by an edge. Id.
`According to Petitioner, if d_i is the degree (or the number of adjacent
`edges) of the i-th information bit, then some information bits are connected
`to two parity bits (i.e., have a degree of two) and other information bits are
`connected to three parity bits (i.e., have a degree of three). Id. Petitioner
`argues that a person of ordinary skill in the art would understand that Luby’s
`assignment of a first group to a degree of two and a second group of input
`bits to a degree of three is “partitioning said data block into a plurality of
`sub-blocks, each sub-block including a plurality of elements,” as recited in
`claim 1. Pet. 17; see also Pet. 24. Petitioner contends that in Figure 17 of
`the Luby, “the input bits with a degree of two are one sub-block and the
`input bits with a degree of three are a second sub-block” of at least two input
`bits. Id.
`With respect to the parity bits of Figure 17 of Luby, Petitioner argues
`that “the parity bits are computed by first repeating the i-th information bit
`d_i times and then interleaving the repeated bits based on the edge
`connection in the graph.” Pet. 17–18 (citing Ex. 1010 ¶ 146). “Next, each
`parity bit with degree d is computed as the modulo-2 sum of d repeated
`bits.” Id. Petitioner contends that Luby teaches “partitioning said data block
`into a plurality of sub-blocks, each sub-block including a plurality of data
`elements” because some bits have different degrees, either two or three. Id.
`
`
`
`13
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`
`Based on the record before us, we agree with Patent Owner that
`Petitioner has not sufficiently shown that Luby discloses “partitioning said
`data block into a plurality of sub-blocks, each sub-block including a plurality
`of data elements,” as recited in claim 1. Prelim. Resp. 26. In addition, the
`Petition fails to adequately show how the combination of Luby and Divsalar
`teach “said first encoding including repeating the data elements in different
`sub-blocks a different number of times” as recited in claim 1.
`Petitioner has neither provided sufficient citation to evidence, nor
`adequately explained how one of ordinary skill in the art applies irregular
`graphing of the edges between node layers as disclosed in Luby to partition
`the block of data in Divsalar into “a plurality of sub-blocks, each sub-block
`including a plurality of data elements,” as recited in claim 1. Indeed, the
`Petition claim chart does not adequately explain how parity bits are
`computed based on the edge graphing disclosed in Luby.9
`With respect to claim 15, Petitioner has failed to provide sufficient
`evidence and testimony to support the combination of the Luby and
`Divsalar. Although Luby discusses that “sparse graph codes” can be
`improved by use of “‘irregular graphing’” (Pet. 24–25 (citing Ex. 1016, Fig.
`17; 11:23–49)), Petitioner has not shown that the irregular graphing of Luby
`
`
`9 Even assuming that Petitioner intended to cite Dr. Pfister’s declaration
`testimony on information bits and parity bits in Figure 17 to explain the
`teaching of Luby (see Prelim. Resp. 29), we are not persuaded. The Pfister
`declaration states that Figure 17 of Luby describes the functional
`relationship between the information bits and the parity bits without any
`reference to Luby. Ex. 1010 ¶ 120. Petitioner failed to cite this testimony,
`which does not adequately show how Figure 17 of Luby teaches one of
`ordinary skill in the art the partitioning and irregular encoding of sub-blocks
`limitations recited in claim 1.
`
`
`
`14
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`combined with the regular repeat codes of block of length N in Divsalar
`teaches the limitations of claim 15.
`In sum, Petitioner has not provided sufficient evidence or argument to
`support the contention that “irregular graphing” of Luby teaches the
`irregular stream of bits as required in claim 15. In addition, we note that
`Petitioner provides only conclusory statements that Divsalar has inputs
`configured to receive a stream of bits as recited in claim 15. Pet. 24.
`On the record before us, we are not persuaded by Petitioner’s
`contentions that Luby in combination with Divsalar teaches irregular
`repeating as required in independent claims 1 and 15. Petitioner’s
`arguments and evidence with respect to claim limitations for partitioning and
`assignment and irregular encoding in claim 1 and the limitation for an
`irregular stream of bits in claim 15 are presented confusingly and are not
`persuasive. See Pet. 15–18; 23–25. Petitioner relies on the same arguments
`presented for independent claims 1 and 15 to support the unpatentability of
`dependent claims 3, 4, 5, 6, 16, 21, and 22. Pet. 14–33.
`For the reasons stated above, we conclude that on this record
`Petitioner has not demonstrated a reasonable likelihood that Petitioner would
`prevail in showing that claims 1, 3, 4, 5, 6, 15, 16, 21, and 22 are
`unpatentable as obvious over Divsalar and Luby.
`5. Analysis—Remaining Grounds based, in part, on Divsalar and
`Luby (claims 15, 16, 20, 21, and 22)
`Petitioner relies on the same arguments presented for claim 15 with
`respect to Divsalar and Luby (Pet. 23–25) in the grounds of unpatentability
`asserted for claims 15, 16, 21, and 22 based on obviousness over Divsalar,
`Luby and Hall (Pet. 35–36); and the grounds of unpatentability asserted for
`
`
`
`15
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`claim 20 based on (i) Divsalar, Luby, and Ping and (ii) Divsalar, Luby, Ping
`and Hall (Pet. 37–43).
`For the reasons discussed above, we conclude that on this record
`Petitioner has not demonstrated a reasonable likelihood that Petitioner would
`prevail in showing that claims 15, 16, 21, and 22 are unpatentable for
`obviousness over Divsalar, Luby, and Hall; claim 20 is unpatentable for
`obviousness over Divsalar, Luby, and Ping; and claim 20 is unpatentable for
`obviousness over Divsalar, Luby, Ping, and Hall.
`III. CONCLUSION
`For the foregoing reasons, we determine that the information
`presented in the Petition does not establish a reasonable likelihood that
`Petitioner would prevail in establishing the unpatentability of: (1) claims 1,
`3, 4, 5, 6, 15, 16, 21, and 22 as obvious over Divsalar and Luby; (2) claims
`15, 16, 21, and 22 as obvious over Divsalar, Luby, and Hall; (4) claim 20 as
`obvious over Divsalar, Luby, and Ping; and (5) claim 20 as obvious over
`Divsalar, Luby, Ping, and Hall.
`IV. ORDER
`
`Accordingly, it is
`ORDERED that pursuant to 35 U.S.C. § 314, an inter partes review is
`hereby denied as to all grounds raised in the Petition for the reasons stated
`above and no trial is instituted.
`
`
`
`
`
`
`
`
`16
`
`

`

`Case IPR2015-00068
`Patent 7,116,710 B1
`
`For PETITIONER:
`
`Eliot Williams
`eliot.williams@bakerbotts.com
`
`G. Hopkins Guy III
`hop.guy@bakerbotts.com
`
`
`
`For PATENT OWNER:
`
`Michael Rosato
`mrosato@wsgr.com
`
`Matthew Argenti
`margenti@wsgr.com
`
`
`
`
`
`17
`
`