Ex Parte Castellani et al

Patent Trial and Appeal Board

Apr 27, 2016

12781299 (P.T.A.B. Apr. 27, 2016)

UNITED STATES PATENT AND TRADEMARK OFFICE
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
12/781,299 05/17/2010 Stefania Castellani 20100042USNP-XER2455US01 2458
62095 7590 04/28/2016
FAY SHARPE / XEROX - ROCHESTER
1228 EUCLID AVENUE, 5TH FLOOR
THE HALLE BUILDING
CLEVELAND, OH 44115
EXAMINER
KIM, PAUL
ART UNIT PAPER NUMBER
2169
MAIL DATE DELIVERY MODE
04/28/2016 PAPER
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

Ex parte STEFANIA CASTELLANI, NIKOLAOS LAGOS,
MARIA ANTONIETTA GRASSO, FREDERIC ROULLAND, and
NICOLAS HAIRON
____________

Appeal 2014-001236
Application 12/781,299
Technology Center 2100
____________

Before JOHN A. JEFFERY, BRADLEY W. BAUMEISTER, and
DENISE M. POTHIER, Administrative Patent Judges.

POTHIER, Administrative Patent Judge.

DECISION ON APPEAL

STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134(a) from the Examiner’s
Final Rejection of claims 1–10 and 12–24. See App. Br. 9.1 Claim 11 has
been canceled. Mar. 12, 2012 Amendment 4. We have jurisdiction under
35 U.S.C. § 6(b).
We affirm-in-part.

1 Throughout this Opinion, we refer to (1) the Final Rejection (Final Act.)
mailed December 10, 2012, (2) the Appeal Brief (App. Br.) filed May 6,
2013, (3) the Examiner’s Answer (Ans.) mailed August 29, 2013, and (4) the
Reply Brief (Reply Br.) filed October 28, 2013.
Appeal 2014-001236
Application 12/781,299

2
INVENTION
Appellants’ invention helps users search a document collection. See
generally Spec. ¶ 1. Claim 1, reproduced below, is illustrative:
1. A method for guiding a user in building a query
comprising:
providing an interaction space in which a user formulates
a query using a plurality of query building blocks selected from
a predefined set of query building blocks, wherein selectable
ones of the query building blocks are displayed in the
interaction space;
providing a content-based mechanism which limits the
selection of the next building block to those which will yield at
least one result from an indexed collection of documents, a
structure-based selection mechanism which limits the selection
of the next building block to those which meet a predefined
grammar, and a usage-based selection mechanism which
proposes a next building block based on patterns of building
blocks used in building prior queries;
with a processor, guiding the user’s formulation of the
query including, when at least one building block has been
selected by the user for formulating the query, guiding selection
of at least a next building block by at least one of:
the content-based mechanism which limits the
selection of the next building block to those which will
yield at least one result from the indexed collection of
documents,
the structure-based selection mechanism which
limits the selection of the next building block to those
which meet the predefined grammar, and
the usage-based selection mechanism which
proposes the next building block based on patterns of
building blocks used in building prior queries.

Appeal 2014-001236
Application 12/781,299

3
The Examiner relies on the following as evidence of unpatentability:
Appelt
Bourdoncle
US 6,601,026 B2
US 7,152,064 B2
July 29, 2003
Dec. 19, 2006

THE REJECTION
Claims 1–10 and 12–24 are rejected under 35 U.S.C. § 103(a) as
unpatentable over Bourdoncle and Appelt. Final Act. 2–8.

CLAIMS 1, 7–9, AND 12–18
Claim 1 recites, in pertinent part, providing “a content-based
mechanism,” “a structure-based selection mechanism,” and “a usage-based
selection mechanism.” Appellants contend that the prior art fails to teach
each of these mechanisms. Reply Br. 10–18.2
“content-based mechanism”
In the rejection of claim 1 under § 103, the Examiner finds that
Bourdoncle teaches the recited content-based mechanism. Final Act. 3–4.
In particular, the Examiner finds that the recited selectable “query building
blocks” correspond to Bourdoncle’s categories and keywords. Id. at 3.
According to the Examiner, by displaying categories to the user, Bourdoncle
teaches the content-based mechanism that limits the selection of the next
building block to those that will yield a result from an indexed collection of
documents. Id. (citing Bourdoncle 8:7–24).
Appellants argue that Bourdoncle responds to a user’s previous query
by retrieving and displaying keywords and categories. Reply Br. 11. In
Appellants’ view, Bourdoncle’s keywords and categories may refine the

2 We primarily cite to the Reply Brief in this opinion because the previous
arguments presented in the Appeal Brief are included.
Appeal 2014-001236
Application 12/781,299

4
previous query’s results, but these keywords and categories do not limit the
ways in which a user may build a query. Id. Appellants further argue that
Bourdoncle’s queries do not ensure that at least one result exists in the
indexed collection. Id. at 11–12. According to Appellants, Bourdoncle only
normally returns results from the collection, but the claim requires that the
query “will yield” such results. Id.at 12 (emphasis omitted); see also id.
at 15–16.

Issue
Under § 103, has the Examiner erred in rejecting claim 1 by finding
that Bourdoncle and Appelt collectively would have taught or suggested “a
content-based mechanism which limits the selection of the next building
block to those which will yield at least one result from an indexed collection
of documents”?

Analysis
We disagree with Appellants’ argument that Bourdoncle’s categories
fail to limit the selection of the next building block as recited (id. at 11–12).
In the disclosure, Appellants discuss a user creating a question, and the
Appellants’ system launches a query of the document collection after each
additional building block. Spec. ¶ 87. When adding blocks to the question,
the content-based mechanism leads the user to blocks that form results-
producing questions. Id. ¶ 91. For example, editable building blocks allow
the user to type input (e.g., building blocks “what”), and as the user types,
the content-based guidance provides suggestions that will yield at least some
query results. See id. ¶¶ 88–91.
Appeal 2014-001236
Application 12/781,299

5
Similarly, Bourdoncle presents suggested categories based on
previous queries that include building blocks. Bourdoncle 8:4–6. The
Examiner maps these categories to the recited “next building blocks.”
Final Act. 3. Like the recited building blocks, Bourdoncle’s categories are
used to refine and limit the current query. Bourdoncle 8:7–8. For example,
the user can select a category to create a new query that produces new
results. Id. at 8:8–10. Given that these categories have a limiting effect, the
issue becomes whether the selection is limited to blocks that yield “at least
one result from an indexed collection of documents,” as recited in claim 1.
On this record, we agree with the Examiner’s finding (Final Act. 3–4)
that Bourdoncle’s selection is so limited. In particular, Bourdoncle’s
suggested category-based refinements are selected from categories
associated with query results. Bourdoncle 6:25–29; Fig. 1A (element 15).
Because each suggested category came from at least one corresponding
document in the query (id.), one can infer that selecting a category will
produce at least some documents that are part of a collection. See id.; see
also Ans. 5. On this record, the Examiner’s finding that selecting
Bourdoncle’s categories will yield at least one result from the collection
(Final Act. 3–4) is reasonable.
We find Appellants’ argument to the contrary (Reply Br. 11–12, 15–
16) unpersuasive. Bourdoncle expressly states that “when a query is refined
by a category, the resulting documents are a subset of the category.”
Bourdoncle 7:2–4. Appellants do not provide persuasive evidence that a
subset of a category3 is an empty set. To be sure, Bourdoncle states
elsewhere that refining a first query by category “normally returns a subset

3 Bourdoncle’s category is a set of documents. Bourdoncle 6:65–66.
Appeal 2014-001236
Application 12/781,299

6
of the set of hits or results provided by the first query.” Id. at 8:22–24
(emphasis added), cited in Reply Br. 12. But this could mean simply that in
some situations, more documents—as opposed to no documents—are
returned. For example, Bourdoncle discloses that using stemming4 may
cause a category-based refinement to return more documents than those
returned by the original query. Bourdoncle 9:10–15.
Additionally, under an obviousness analysis, Bourdoncle teaches
embodiments where selecting a category limits the results to those in a
collection, as previously stated. That is, Bourdoncle need not teach the
claim limitation is necessarily present in all of Bourdoncle’s embodiments to
establish a prima facie case of obviousness. As such, Appellants’
hypothetical related to Bourdoncle’s initial search query (Reply Br. 11–12)
is not persuasive. Specifically, Appellants argue that Bourdoncle allows a
user to enter a query so narrow that returns no results. Id. Even if true, we
have illustrated scenarios above where the content-based mechanism does
limit the selection of the next building block to those that yield at least one
result. See also Bourdoncle, Figs. 1A (indicating 839 results), 2A (a more
restrictive search path of “greenhouse effect: Issues” that yields 26 results).
Accordingly, a showing that Bourdoncle provides “a content-based
mechanism” and at least in one instance uses the produced categories to
limit the selection of the next building blocks to yield a result from a
document collection, as the Examiner has done here (see Final Act. 3–4),
satisfies the disputed limitation.

4 Bourdoncle’s stemming automatically associates additional words with a
keyword. See, e.g., Bourdoncle, col. 7, ll. 47–63. For example, the stem
“telephon*” covers “telephone” and “telephony,” among others. See id.,
col. 7, ll. 48–51.
Appeal 2014-001236
Application 12/781,299

7
“structure-based selection mechanism”
The Examiner finds that Bourdoncle teaches all of the limitations of
independent claim 1 except for the recited structure-based selection
mechanism. Final Act. 3–4. The Examiner cites Appelt to teach this feature
and proposes to include this feature with Bourdoncle’s system to limit the
selection of certain keywords (i.e., recited “building blocks”) in concluding
that this feature would have been obvious. Id. at 4; Ans. 5–6. In particular,
the Examiner finds that Appelt discloses developing a set of rules that
identify various natural language queries and are stored in grammar files.
Id. at 3–4; Ans. 5–6 (citing Appelt 12:23–35). The Examiner concludes that
it would have been obvious to use Appelt’s grammar files to limit the
selection of Bourdoncle’s keywords. Id. at 4.
Appellants argue that Appelt lacks a structure-based mechanism.
Reply Br. 15. According to Appellants, Appelt generates search parameters
using a predefined grammar. Id. But Appellants contend that Appelt’s
parameters cannot limit building-block selection, as claimed. Id. In
Appellants’ view, Appelt’s rules only identify likely queries, but the rules do
not limit the development of an actual query or the selection of building
blocks. Id. at 16. Appellants further argue that the Bourdoncle-Appelt
combination does not either. Id. at 17.

Issue
Under § 103, has the Examiner erred in rejecting claim 1 by finding
that Bourdoncle and Appelt collectively would have taught or suggested “a
structure-based selection mechanism which limits the selection of the next
building block to those which meet a predefined grammar”?
Appeal 2014-001236
Application 12/781,299

8
Analysis
Appellants’ argument that Appelt does not limit the development of
an actual query (see Reply Br. 15–16) amounts to an individual attack. Such
an individual attack does not show nonobviousness where, as is here, the
rejection is based on a combination of references. See In re Merck & Co.,
Inc., 800 F.2d 1091, 1097 (Fed. Cir. 1986). That is, the rejection relies on a
combination of Bourdoncle and Appelt to address the recited structure-based
selection mechanism. See Final Act. 3–4; Ans. 6, 5 quoted in Reply Br. 17.
Specifically, the Examiner maps Bourdoncle’s keywords and
categories to the recited next building blocks. See Final Act. 3 (citing
Bourdoncle 6:15–23). Furthermore, the Examiner proposes applying
Appelt’s rules to Bourdoncle’s building blocks—namely limiting the
selection of “certain keywords according to the rules of grammar files” to
improve Bourdoncle’s system. Ans. 5–6; Final Act. 4. Because the
Examiner does not rely on Appelt alone for limiting the selection of the next
building block-based feature, Appelt need not teach the recited feature
exclusively, for we consider what the references teach collectively. Thus,
Appellants’ arguments (Reply Br. 15–16) do not fully address the
Examiner’s rejection.
Apart from this individual attack against Appelt (Reply Br. 15–16),
Appellants do not provide a persuasive rebuttal to the Examiner’s
obviousness conclusion. For example, Appellants have not shown that the
proposed combination would have posed a unique challenge or otherwise
have been beyond the level of ordinary skill. See Leapfrog Enters., Inc. v.
Fisher-Price, Inc., 485 F.3d 1157, 1162 (Fed. Cir. 2007). Instead,

5 Appellants mistakenly refer to Ans. 5.
Appeal 2014-001236
Application 12/781,299

9
Appellants allege that the combination would not have resulted in limiting
the selection. Reply Br. 17. Without more, we determine that the
Examiner’s obviousness conclusion (Final Act. 3–4) is reasonable.

“usage-based selection mechanism
In the rejection under § 103, the Examiner finds that Bourdoncle
teaches the recited usage-based selection mechanism. Final Act. 3 (citing
Bourdoncle 10:11–29). In particular, the Examiner finds that Bourdoncle
proposes the next building block by generating additional keywords and
categories. Ans. 6–7 (citing Bourdoncle, Fig. 6). According to the
Examiner, Bourdoncle’s first set of categories and keywords are the recited
patterns in building prior queries. Ans. 7.
Appellants argue that Bourdoncle retrieves keywords based on the
most recently entered query. Reply Br. 17–18. According to Appellants,
generating results from a single query does not suggest a next building block
based on patterns used in building prior queries. Id. at 18.

Issue
Under § 103, has the Examiner erred in rejecting claim 1 by finding
that Bourdoncle and Appelt collectively teach or suggest “a usage-based
selection mechanism which proposes a next building block based on patterns
of building blocks used in building prior queries”?

Appeal 2014-001236
Application 12/781,299

10
Analysis
Bourdoncle’s Figure 6, reproduced below, illustrates its search
process.

Figure 6 showing a process flowchart for searching a document collection.

In this process, Bourdoncle’s user initially inputs a query (step 24).
Bourdoncle, Fig. 6. After the query results are returned (step 26), the user
may select a category (step 34) from those proposed and displayed (steps 28
Appeal 2014-001236
Application 12/781,299

11
through 32). Id. The process then repeats after a new query is defined
(arrow leaving step 36). Id.
Appellants argue that Bourdoncle only returns results based on the
latest query or a single pattern. Reply Br. 17–18. On this record, the
Examiner has shown that Bourdoncle’s proposes a next building block
(step 32) based on the recited patterns in multiple prior queries (step 36), not
just a single query. See Ans. 6–7. Based on Bourdoncle’s iterative nature,
Bourdoncle’s suggested keywords and categories (step 32) are based on
previous query results (steps 26 and 36). See Bourdoncle 10:11–29;
Figs. 1A, 2A. For example, Figure 1A shows an initial query with building
blocks “greenhouse effect” that yields additional building blocks (e.g.,
categories 15 and keywords 17). Bourdoncle, Fig. 1A. Figure 2A shows a
second and refined query (e.g., search path “Greenhouse effect: Issues”)
necessarily builds off of the first query, where the results yield other
resulting keywords (e.g., Global warming). Bourdoncle, Fig. 2A. As such,
the usage-based selection mechanism in Figure 2A (e.g., resulting keywords)
is based on patterns of building blocks (e.g., “greenhouse effect” and
“Issues”) used in previous queries (e.g., two prior queries). Accordingly, we
agree that these previous queries establish the recited patterns. See Ans. 6–7
(citing Bourdoncle, Fig. 6).
To the extent that Appellants argue that “based on” requires that each
building block used in building prior queries must be used in proposing a
next building block (see Reply Br. 17–18), Bourdoncle also discloses this
feature. As illustrated above, Bourdoncle’s user can select categories that
further refine a previous category. See Bourdoncle 10:20–29. When this
Appeal 2014-001236
Application 12/781,299

12
selection is repeated, it is reasonable to interpret this results in a pattern of
categories. See Ans. 6–7.
In particular, Bourdoncle states that “when a query is refined by a
category, the resulting documents are a subset of the category.”
Bourdoncle 7:2–4. Also, these categories can be hierarchical. See id. 7:14–
16. So when a user navigates downward in this hierarchy, the list of
traversed categories forms a pattern that the next suggested category is
“based on” (id.), as recited. That is, each selected category (step 34) in this
traversal forms the basis for the next suggested keywords and categories
(step 28). See id.; see also id., Fig. 6 (showing an iterative process).
For this additional reason, we are unpersuaded by Appellants’
argument that Bourdoncle lacks a “usage-based selection mechanism.”

“guiding the user’s formulation”
Appellants argue that the Examiner has not considered all of claim 1’s
limitations. Reply Br. 10. In particular, Appellants characterize the
Examiner’s interpretation as making optional the recited structure-based and
usage-based selection mechanisms. Id. To the contrary, though, the
Examiner interprets the structure-based and usage-based selecting
mechanisms as being optional in “guiding the user’s formulation” step—not
in the providing step. Ans. 3 (citing to Final Act. 3–4, n.1, 2). Furthermore,
the Examiner does explain how the Bourdoncle-Appelt combination
provides all three mechanisms. Ans. 3; Final Act. 3–4.
Regarding the recited “guiding” step, we agree with the Examiner’s
claim interpretation. Although claim 1 calls for providing all three
mechanisms, the claim only requires the “guiding selection” be performed
Appeal 2014-001236
Application 12/781,299

13
by “at least one” of the recited mechanisms. This interpretation is consistent
with Appellants’ remarks. See Reply Br. 11. For example, Appellants state
that “any one or more of [the three mechanisms] may be used in
combination to help a user formulate a query.” Id. In formulating the
obviousness rejection, the Examiner needs only to show that Bourdoncle and
Appelt collectively guide the selection by at least one of the recited
mechanisms. For the reasons discussed previously, the Examiner did not err
in finding that Bourdoncle guides selection of the recited building blocks by
a content-based mechanism (Ans. 3–4).
For the foregoing reasons, Appellants have not persuaded us of error
in the rejection of claim 1, as well as claims 7–9 and 12–18, not separately
argued with particularity. See Reply Br. 10–18.6

CLAIMS 4 AND 5
Contentions
Claim 4 recites, in part, “the building blocks include building blocks
for result type, entities, relations, and events.” The Examiner relies on
Bourdoncle in concluding claim 4 would have been obvious over the
Bourdoncle-Appelt combination. Final Act. 5 (citing Bourdoncle 8:7–24).
Because Bourdoncle returns categories and keywords with query results, the
Examiner finds that these categories and keywords correspond to the recited
result-type block. Ans. 7.

6 Appellants list claim 11 in a section heading. Reply Br. 10. But claim 11
is not rejected because this claim has been canceled. See id. at 9 (noting that
only claims 1–10 and 12–24 stand rejected); see also id. at 38, Claims
App’x.
Appeal 2014-001236
Application 12/781,299

14
Appellants argue that Bourdoncle lacks a result-type block.
Reply Br. 18–20. In Appellants’ view, it would have been unreasonable to
have a result-type block because categories and keywords are the search
results. Id.
Issue
Under § 103, has the Examiner erred in rejecting claim 4 by finding
that Bourdoncle and Appelt collectively teach or suggest “blocks for result
type”?

Analysis
When describing listing the types of blocks, the Specification states
that the “ symbol” corresponds to one of the building blocks
defining the block type returned by a query. Spec. ¶ 76; see also id. ¶ 74,
Table 1. Although instructive, the Specification does not set forth clearly a
definition for the term “result type” sufficient to rebut the presumption that
this word should be given its ordinary and customary meaning.
Here, Appellants admit that Bourdoncle’s categories and keywords
are “the result of a search query.” Reply Br. 19. That is, Bourdoncle’s
category and keyword blocks are a “result” by the plain meaning of the term.
Appellants, however, argue that these categories and keywords are not
“results type” blocks, because they are not used in formulating an input
query. Id. But Appellants offer no definition or persuasive reason for
result-type that would exclude blocks that are the result of a query. See
id. at 18–20. Moreover, as discussed above, the categories and keywords are
used in formulating queries. So under the plain meaning of the term
Appeal 2014-001236
Application 12/781,299

15
“result,” the Examiner’s interpretation of Bourdoncle’s categories and
keywords as “blocks for result type” (Ans. 7; Final Act. 5) is reasonable.
Appellants further argue that the Examiner’s interpretation would
render all categories and keywords result-type blocks, and none would be
left for the “entities, relations, and events.” Reply Br. 19–20. Here again,
Appellants provide no definition for these types that requires the recited
types to be mutually exclusive. See id. That is, we see no definition
precluding a result type from also being one or more of an entity, relation, or
event. On this record, we determine that the Examiner did not err in
interpreting Bourdoncle’s categories and keywords as “blocks for result
type” as well as “entities, relations, and events.”
For the foregoing reasons, we sustain the rejection of claim 4, as well
as that of claim 5, not separately argued with particularity. See
Reply Br. 10–20.

CLAIM 6
Contentions
Claim 6 recites, in part, “one of the blocks is a connector block for
linking two other blocks with a connector.” The Examiner relies on
Bourdoncle in concluding claim 6 would have been obvious over the
Bourdoncle-Appelt combination. Final Act. 6 (citing Bourdoncle 10:11–29).
According to the Examiner, the keywords and categories retrieved by
Bourdoncle correspond to the connector blocks. Ans. 8.
Appellants argue that the Examiner’s interpretation of “connector
block” is unreasonable. Reply Br. 20. In Appellants’ view, the Examiner
cannot argue that keywords and categories are both connector blocks in
Appeal 2014-001236
Application 12/781,299

16
claim 6 and result-type blocks in claim 4. Id. Furthermore, Appellants
contend that Bourdoncle’s keywords and categories do not link two other
blocks. Id.

Issue
Under § 103, has the Examiner erred in rejecting claim 6 by finding
that Bourdoncle and Appelt collectively teach or suggest “a connector block
for linking two other blocks”?

Analysis
We begin by construing the key disputed limitation of claim 1, which
calls for, in pertinent part, “connector block.” To this end, we look to
Appellants’ Specification. The Specification is “always highly relevant” and
the “single best guide to the meaning of a disputed term.” In re Abbott
Diabetes Care Inc., 696 F.3d 1142, 1149 (Fed. Cir. 2012) (quoting Phillips
v. AWH Corp., 415 F.3d 1303, 1315 (Fed. Cir. 2005) (en banc)). Although
lacking a definition, the Specification is consistent with the Examiner’s
interpretation of the recited connector block as Bourdoncle’s categories.
Ans. 8. Specifically, the Specification states that “[e]xamples of the
composition block types may be as follows”: a “connector” contains “AND”
and “OR” subtypes. Spec. ¶ 40; see also id. ¶ 74, Table 1.
Like the disclosed connector blocks (id. ¶ 40), Bourdoncle’s blocks
also link two blocks by the term “and”—e.g., the “Issue, Religion, and
Spirituality” category. See Bourdoncle, Fig. 1A (emphasis added). On this
record, the Examiner’s finding that Bourdoncle teaches “a connector block
for linking two other blocks” (Ans. 8) is reasonable.
Appeal 2014-001236
Application 12/781,299

17
Appellants further argue that the Examiner’s interpretation of
categories as result type blocks precludes their interpretation as connectors.
Reply Br. 20. But we see nothing on the record that precludes categories
from including multiple types (e.g., result type and connector type). Nor do
Appellants provide a definition or a sufficient reason for the recited block
types that requires mutual exclusivity. See id.
For the foregoing reasons, we sustain the rejection of claim 6.

CLAIM 10
Claim 10 recites, in part, “the interaction space includes a question
editing area and wherein the method comprises providing for the user to
move selected blocks into the question editing area in formulating the query”
(emphasis added). The Examiner relies on Bourdoncle in concluding claim
10 would have been obvious over the Bourdoncle-Appelt combination.
Final Act. 6 (citing Bourdoncle 10:11–29). Specifically, the Examiner finds
that Bourdoncle’s user interface is the question editing area. Ans. 9.
According to the Examiner, presenting the keywords returned with
Bourdoncle’s results corresponds to the recited block movement. Id.
Appellants argue that Bourdoncle lacks the recited movement
provided to the user in the claim. Reply Br. 22. According to Appellants,
Bourdoncle’s user has no reason to move categories and keywords into the
question editing area. Id. We agree.
Bourdoncle’s categories are used to limit the current query.
Bourdoncle 8:7–8. For example, the user can select the category to create a
new query that produces new results. Id., 8:8–10. But the Examiner has not
Appeal 2014-001236
Application 12/781,299

18
shown that Bourdoncle provides the user with a way to move these
categories into the question editing area as recited. See Final Act. 6.
Furthermore, the Examiner’s explanation that Bourdoncle’s search
results provide the recited movement (Ans. 9) is unsupported. For example,
Bourdoncle’s Figure 5A, reproduced below, shows the user interface after
the user selects the category “Climate Change.” See Bourdoncle 9:39–45.

Figure 5 showing the user interface after “Climate Change”
is selected.
Appeal 2014-001236
Application 12/781,299

19

As shown in Bourdoncle’s Figure 5, the category “Climate Change”
has been selected, but the question editing area only displays “greenhouse
effect.” See id. In other words, contrary to the Examiner’s finding
(Final Act. 6), Bourdoncle does not display “Climate Change” in the user-
editable area, such that selected blocks can be considered to have moved into
a question editing area. Bourdoncle, Figure 5A.
Accordingly, we do not sustain the rejection of claim 10.

CLAIM 24
Claim 24 recites, in part, “providing for the user to start to build the
query by selecting a first of the query building blocks, and moving it into the
editing area.” The Examiner finds that the recited providing “is an intended
use [that] lacks patentable weight” and thus, “no prior art has been applied
to” this feature. Final Act. 8 n.4.
We, however, agree with Appellants’ argument that the providing step
further limits the claim. Reply Br. 23. In particular, claim 24 affirmatively
recites providing particular functions to the user. Namely, the claimed
method must provide two functions: (1) “selecting a first of the query
building blocks” and (2) “moving it into the editing area.” By not
addressing these features (see Final Act. 8), the Examiner has not made a
prima facie case of obviousness.
In the Answer, the Examiner responds to Appellants’ argument with a
rationale similar to the one supporting claim 10’s rejection. Compare
Ans. 10 with Final Act. 9. To the extent that the Examiner finds that
Bourdoncle’s generation of keywords and categories corresponds to the
Appeal 2014-001236
Application 12/781,299

20
moving blocks into the editing area (see Final Act. 8), we disagree for the
reasons stated above for the rejection of claim 10.
Accordingly, we do not sustain the rejection of claim 24.

CLAIMS 2 AND 3
Contentions
As with the rejection of claim 1, the Examiner finds that Bourdoncle
teaches all of the limitations of independent claim 2 except for the recited
structure-based selection mechanism. Final Act. 4–5. The Examiner cites
Appelt in concluding that this feature would have been obvious. Id. In
particular, the Examiner finds that Appelt discloses grammar files. Id. at 4.
The Examiner concludes that it would have been obvious to use Appelt’s
grammar files to limit the selection of Bourdoncle’s keywords. Id. at 4–5.
Appellants specifically argue that one of ordinary skill would not have
combined Appelt with Bourdoncle. Reply Br. 24. According to Appellants,
Appelt’s natural language inputs and outputs are unlike Bourdoncle’s
keywords and categories. Id. Appellants contend that the references do not
suggest how to modify Appelt to produce the recited structure-based
selection mechanism. Id. at 25. In Appellants’ view, Appelt uses grammar
files, but these files are not integrated into a structure-based selection
mechanism. Id. Instead, Appellants argue that Appelt’s grammar files are
used in generating a search in a computer language. Id. Appellants further
contend that the Examiner’s rejection does not address the specific types of
blocks recited in claim 2. Id. at 26.
Lastly, Appellants argue that the Examiner’s interpretation of a
result-type block is unreasonable. Id. at 27. In Appellants’ view, the
Appeal 2014-001236
Application 12/781,299

21
Examiner has not shown the Bourdoncle’s user is required to select one of a
set of result-type blocks. Id.

Issues
Under § 103, has the Examiner erred in rejecting claim 2 by finding
that Bourdoncle and Appelt collectively would have taught or suggested
I. “a structure-based selection mechanism which limits the selection
of the next building block to those which meet a predefined grammar”?
II. “the grammar of the structure-based selection mechanism requiring
that the user selects one of a set of result type building blocks, and at least
one other block, which is not a result type block, in formulating the query”?

Analysis
I
The Examiner has provided a reason with some rational underpinning
to combine Appelt and Bourdoncle. Namely, the Examiner states that using
Appelt’s grammar files with Bourdoncle’s search technique would have
predictably resulted in limiting keyword selection according to those rules.
Final Act. 5. In support of this conclusion, the Examiner points to the
similarities between Appelt’s and Bourdoncle’s search systems. Id. Namely,
that both references operate on keywords. Id. In essence, the Examiner
determines that applying Appelt’s grammar files to Bourdoncle’s system
improves one known process operating on keywords for another yielding no
more than the identified predictable results. See KSR Int’l Co. v. Teleflex
Inc., 550 U.S. 398, 416 (2007); Final Act. 5; Ans. 6.
Appeal 2014-001236
Application 12/781,299

22
Appellants have not rebutted the Examiner’s specific findings
adequately. “A suggestion, teaching, or motivation to combine the relevant
prior art teachings does not have to be found explicitly in the prior art . . . .”
In re Kahn, 441 F.3d 977, 987–88 (Fed. Cir. 2006). That is, an obviousness
inquiry “can take account of the inferences and creative steps” of an
ordinarily skilled artisan. See KSR, 550 U.S. at 418. As discussed above,
the Examiner relies on the predictable and resulting improvement of
Bourdoncle’s search technique by including Appelt’s teaching within
Bourdoncle’s framework to restrict the keywords to those that are in
accordance with grammar file rules and thus limit the structure-based
selection mechanism’s selection of next building blocks to those that meet a
predefined grammar. See Final Act. 5. This reasoning and conclusion has
not been squarely addressed by Appellants’ argument. See Reply Br. 24–25.
Instead, Appellants argues there is no suggestion to modify Appelt’s system
to produce “a structure-based selection mechanism” as recited, rather than
address the proposed inclusion of Appelt’s technique to Bourdoncle’s
system. See id.
Apart from these arguments, Appellants point out the differences
between Appelt and Bourdoncle. See Reply Br. 24 (arguing that Appelt’s
natural language inputs and outputs are different from Bourdoncle’s full
integration with keywords and categories). But “[t]he test for obviousness is
not whether the features of a secondary reference may be bodily
incorporated into the structure of the primary reference.” In re Keller, 642
F.2d 413, 425 (CCPA 1981). Here, the Examiner proposes combining the
references to add Appelt’s natural language inputs and outputs to
supplement Bourdoncle’s keywords and categories. Final Act. 5.
Appeal 2014-001236
Application 12/781,299

23
Appellants have provided no evidence of incompatibility. See
Reply Br. 24–25. For example, Appellants argue that Appelt’s grammar
files are used in generating a search in a computer language. Id. at 25. But
Appellants have not shown that repurposing Appelt’s grammar files in
Bourdoncle to assist in generating keywords and categories according to
particular grammar rules would have posed a unique challenge or otherwise
have been beyond the level of ordinary skill. See Leapfrog, 485 F.3d at
1162.
Accordingly, we are unpersuaded by Appellants’ arguments regarding
the combinability of Appelt and Bourdoncle to arrive at “a structure-based
selection mechanism.”

II
We are also unpersuaded by Appellants’ argument that the
Examiner’s rejection does not address the specific block types recited in
claim 2. Reply Br. at 26. The Examiner’s articulated position is based on
the combination of Bourdoncle and Appelt. Although Appellants allege that
the combination does not disclose the recited features generically,
Appellants do not explain why the Examiner’s findings and conclusions are
deficient. Id. at 26. In fact, Appellants’ argument that the Examiner
provides an unreasonable interpretation of a result type block (id. at 27)
acknowledges that at least some findings by the Examiner were made. On
this point, we refer back to our discussion of claim 4.
Likewise, we are unpersuaded by Appellants’ argument that the
Examiner has not shown the Bourdoncle’s user is required to select one of a
set of result type blocks. Id. Here, Appellants attack Bourdoncle
Appeal 2014-001236
Application 12/781,299

24
individually when arguing that the art does not teach requiring that the user
select one of a set of result type building blocks, as recited. Id. But the
rejection relies on a combination of Bourdoncle and Appelt to address this
disputed portion of the recited structure-based selection mechanism. See
Final Act. 5. Specifically, the Examiner proposes applying Appelt’s rules to
Bourdoncle’s building blocks, yielding building blocks to those that meet a
predefined grammar and, thus, the resulting blocks for selection (e.g.,
keywords) are those in accordance with the grammar rules. Id. at 4–5;
Ans. 6. Additionally, as discussed earlier when addressing claim 4, these
building blocks are result type. Because the Examiner does not rely on
Bourdoncle alone for limiting the selection of a block-based query (see id.),
Bourdoncle alone need not teach this feature.
For the foregoing reasons, we sustain the rejection of claim 2, as well
as claim 3, not argued separately with particularity (see Reply Br. 24–27).

CLAIM 19
We, however, do not sustain the Examiner’s obviousness rejection of
claim 19 (Final Act. 3–4). Here, the “rejection is so uninformative that it
prevents the applicant from recognizing and seeking to counter the grounds
for rejection.” See In re Jung, 637 F.3d 1356, 1362 (Fed. Cir. 2011)
(internal citations and quotation marks omitted).
Unlike claim 1, claim 19 requires “contemporaneously displaying at
least a subset of the query building blocks,” as well as guiding a user “after
receiving the user’s selection.” These recitations have not been addressed
by the Examiner. See Final Act. 3–4. Instead, without discussing claim 19’s
specific limitations, the Examiner states that this claim is rejected as obvious
Appeal 2014-001236
Application 12/781,299

25
over Bourdoncle and Appelt and groups the claim with claim 1. See id.
Because the Examiner does not discuss several of the limitations in claim 19,
or otherwise explain how the applied references are relevant to these
limitations, the Examiner failed to make a prima facie case of obviousness
for claim 19. See id. Consequently, the Examiner’s rejection of claim 19
(id.) is erroneous for at least that reason.

CLAIMS 20–23
For the rejections of claims 20–23, Appellants refer to the arguments
presented for the patentability of claim 1. See Reply Br. 31. In particular,
Appellants argue that the three recited components are not optional
(compare id. with id. at 10–11), and the Bourdoncle-Appelt combination
lacks a content-based selection component (compare id. at 31 with id. at 11–
14) and a usage-based selection component (compare id. at 32–33 with
id. at 17–18). The issues before us, then, are similar to those in connection
with claim 1 in many aspects, and we refer Appellants to our previous
discussion.
We note, however, that the “guidance module for guiding a user in
formulating a query” of claim 20 does not include the phrase “at least one
of,” as recited in claim 1’s step of “guiding selection of at least a next
building block.” Compare App. Br. 33 with App. Br. 28–29. Even so, we
have illustrated above that Bourdoncle and Appelt collectively teach
providing “a content-based selection” mechanism or component, “a
structure-based selection” mechanism or component, and “a usage-based
selection” mechanism or component. Because the prior art provides these
mechanisms and these mechanisms guide the user in formulating subsequent
Appeal 2014-001236
Application 12/781,299

26
queries as discussed with claim 1, the combination also teaches a “guidance
module for guiding a user in formulating the query,” as recited in claim 20.
Therefore, we will sustain the Examiner’s rejection of claims 20–23
for similar reasons.

CONCLUSIONS
The Examiner did not err in rejecting claims 1–9, 12–18, and 20–23
under § 103.
The Examiner erred in rejecting claims 10, 19, and 24 under § 103.

DECISION
The Examiner’s decision rejecting claims 1–10 and 12–24 is affirmed-
in-part.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a). See 37 C.F.R.
§ 1.136(a)(1)(iv).

AFFIRMED-IN-PART