From Argument to Evidence: The Importance of Expert Declarations in Responding to a Non-Final Office Action in Ex Parte Reexamination

Executive Summary: In ex parte reexamination, a patent owner’s response to a non-final Office Action is often a critical opportunity to build the factual record, particularly when the rejection depends on technical assumptions about prior-art references, claim mappings, system architecture, protocols, or an obviousness rationale. The article argues that attorney argument can frame the legal dispute, but expert declarations supply the evidentiary foundation needed to show why the examiner’s technical premise is unsupported. Using case studies involving protocol-level, architecture-level, and abstraction-level expert testimony, the article shows how targeted declarations can rebut overbroad prior-art characterizations while helping preserve original claim language in settings where amendment may create litigation, intervening-rights, or prosecution-history risks. The practical lesson is that expert declarations should be used selectively and carefully: they should explain the technology, the references, the skilled artisan’s understanding, and the factual flaws in the rejection, while avoiding conclusory opinions, unnecessary admissions, and inconsistency with parallel litigation or PTAB positions

I.             Introduction

A non-final Office Action in an ex parte reexamination is not merely another prosecution event. For the patent owner, it is often the best opportunity to build a factual record before the Office—particularly where the rejection turns on how a person of ordinary skill in the art would understand a technical reference, claim term, protocol, system architecture, or proposed obviousness combination.

That distinction matters because attorney argument and evidence serve different functions. Attorney argument frames the legal dispute. Expert testimony supplies technical facts. In a reexamination response, the strongest position is often not simply that the examiner is wrong, but that the examiner’s technical premise is unsupported as a matter of evidence.

The USPTO’s own rules recognize this distinction. Under 37 C.F.R. § 1.132, when a claim in an application or a patent under reexamination is rejected or objected to, evidence submitted to traverse the rejection on a basis not otherwise provided for must be submitted by oath or declaration. The MPEP further states that the primary examiner must personally review such evidence and determine whether it is responsive to the rejection and presents sufficient facts to overcome it. MPEP § 716.

Timing is also important. Evidence traversing a rejection, including a Rule 132 declaration, is considered timely when submitted before final rejection, and timely evidence must be considered by the examiner. MPEP § 716.01. In ex parte reexamination, ordinary application-extension practice under 37 C.F.R. § 1.136 does not apply; extensions are governed by 37 C.F.R. § 1.550(c), and reexamination proceedings are conducted with “special dispatch.” MPEP § 2265; 35 U.S.C. § 305; 37 C.F.R. § 1.550(a).

II.         Why Expert Declarations Matter

The core value of an expert declaration is that it converts a patent owner’s response from advocacy into evidence. The MPEP states that arguments by an applicant cannot take the place of evidence in the record. MPEP § 716.01(c); In re Schulze, 346 F.2d 600, 602 (C.C.P.A. 1965); In re De Blauwe, 736 F.2d 699, 705 (Fed. Cir. 1984).

That principle has special force in reexamination. Many reexamination rejections depend on technical equivalence: the examiner or requester may contend that a prior-art “request” is the claimed request, that a prior-art “cache” is the claimed cache, that prior-art “partitioning” is the claimed partitioning, or that a combination would have been a predictable use of known elements. In those circumstances, attorney argument can identify the issue, but expert testimony can explain why the asserted equivalence is technically wrong.

A declaration is most useful when it focuses on the factual predicates of the rejection. The expert should not merely state that the claims are patentable or that the prior art does not render the claims obvious. Instead, the declaration should explain what the references teach, what they do not teach, how a skilled artisan would understand them, and why the examiner’s mapping or combination rationale does not hold together technically.

The MPEP recognizes that factual evidence is preferable to opinion testimony, but also states that opinion testimony is entitled to consideration and some weight when it is not merely an ultimate legal conclusion and when the underlying factual basis is persuasive. MPEP § 716; In re Lindell, 385 F.2d 453, 456 (C.C.P.A. 1967); In re Oelrich, 579 F.2d 86, 91–92 (C.C.P.A. 1978).

III.      Obviousness: The Declaration as a Challenge to the Articulated Rationale

The role of expert testimony is especially important in obviousness. After KSR, obviousness remains flexible, but it is not conclusory. The rejection must still identify why the claimed invention would have been obvious to a person of ordinary skill in the art. KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 418 (2007); In re Kahn, 441 F.3d 977, 988 (Fed. Cir. 2006).

MPEP § 2143 makes this point explicit. The “key” to a proper § 103 rejection is clear articulation of the reasons why the claimed invention would have been obvious, and KSR requires the obviousness analysis to be made explicit. MPEP § 2143. Office personnel must make appropriate factual findings and provide a reasoned explanation as to why the claimed invention would have been obvious at the relevant time. Id.

That is where a declaration can be doctrinally powerful. Expert testimony can attack the factual predicates of the examiner’s articulated rationale. For example, the expert may show that the proposed combination assumes technical compatibility that the references do not support; that a stated “efficiency” rationale is generic and untethered to the claim-required architecture; that the proposed modification would change a reference’s principle of operation; that the alleged predictable result was not predictable in the actual system context; or that a person of ordinary skill would not have had a reasonable expectation of success.

This is different from merely saying “there is no motivation to combine.” The stronger declaration explains why the articulated motivation is technically incomplete, why the proposed integration path is missing, and why the Office Action’s theory depends on hindsight rather than record evidence.

IV.      Case Study One: The ’785 Patent — Protocol-Level Expert Evidence

The first case study involves U.S. Patent No. 7,949,785, titled “Secure Virtual Community Network System,” in Ex Parte Reexamination Control No. 90/019,519. The patent owner submitted the Declaration of Dr. Guevara Noubir in response to the non-final Office Action. The Office later issued a Notice of Intent to Issue Ex Parte Reexamination Certificate (NIRC) confirming claims 30 and 34. Notice of Intent to Issue Ex Parte Reexamination Certificate, Ex Parte Reexamination of U.S. Patent No. 7,949,785, Control No. 90/019,519, at 1–2 (Mar. 6, 2025) [hereinafter ’785 NIRC].

The rejection relied on Leung, Forslöw, and Mehta. The disputed limitation required, among other things, a DNS server configured to receive a DNS request and return multiple pieces of address information, including a network address associated with a route director, a private network address associated with a second device, and a virtual network address associated with that second device.

Dr. Noubir’s declaration attacked the rejection at the protocol level. He explained that Forslöw’s mobile service manager retrieved information using HTTPS requests, but that an HTTPS request is fundamentally different from a DNS request. A DNS request resolves a domain name into an IP address, while the HTTPS request relied upon from Forslöw lacked the domain-name-resolution function required by the claim. Declaration of Dr. Guevara Noubir ¶¶ 37–40, Ex Parte Reexamination of U.S. Patent No. 7,949,785, Control No. 90/019,519 (Jan. 8, 2025) [hereinafter Noubir Decl.].

The declaration also addressed another technical premise: whether a virtual network or VPN necessarily implies a domain name. Dr. Noubir explained that IP-address assignment and VPN operation do not require a domain name, and that the cited combination did not teach a virtual network defined by a domain name as required by claim 30. Noubir Decl. ¶¶ 35–37.

For claim 34, the declaration distinguished registration from authentication or join/leave activity. Dr. Noubir explained that Leung’s use of “registration” and “access request” related to authentication, while Forslöw and Mehta used registration concepts in different technical contexts. Noubir Decl. ¶ 55.

The lesson is that expert declarations can be especially valuable when the rejection depends on protocol substitution. Attorney argument could say that HTTPS is not DNS. Expert testimony can explain why, in the actual network architecture, the two requests perform different functions, contain different information, and satisfy different claim requirements.

V.          Case Study Two: The ’844 Patent — Architecture-Level Expert Evidence

The second case study involves U.S. Patent No. 8,332,844, titled “Root Image Caching and Indexing for Block-Level Distributed Application Management,” in Ex Parte Reexamination Control No. 90/019,938. Patent Owner submitted the Declaration of Dr. Richard Newman in support of its January 23, 2026 response to the October 23, 2025 non-final Office Action. Declaration of Dr. Richard Newman, Ex Parte Reexamination of U.S. Patent No. 8,332,844, Control No. 90/019,938 (Jan. 23, 2026) [hereinafter Newman Decl.].

The Office Action asserted obviousness rejections over Sapuntzakis in view of Holzmann for claims 1–3 and 6–13, and over Sapuntzakis, Holzmann, and Federwisch for claims 19–27. The patent owner’s response explained that the rejection relied on the Zadok declaration for claim mappings and rationales to combine, but argued that the cited art did not disclose the claimed architecture. Response to Office Action, Ex Parte Reexamination of U.S. Patent No. 8,332,844, Control No. 90/019,938, at 1–2 (Jan. 23, 2026) [hereinafter ’844 Response].

This case illustrates an architecture-level use of expert testimony. The claims required a shared root image, per-node leaf images containing only changed or additional blocks, and a cache configured to cache blocks of the root image previously accessed by at least one compute node. The response argued that the rejection improperly treated generic caching as the claimed cache, without showing that the cached blocks were blocks of the shared root image in the claimed root/leaf architecture. ’844 Response at 55–56.

Dr. Newman’s declaration made the same point technically. He explained that Sapuntzakis’s “remote COW disk” and “shadow COW disk” belonged to Sapuntzakis’s own copy-on-write hierarchy and remote disk caching mechanism, and that the Office Action did not identify evidence showing that the cached blocks were blocks of a shared root image as required by the claims. Newman Decl. ¶¶ 145–47.

The declaration also attacked the motivation to combine. The response characterized the Office Action’s “faster access” rationale as a generic performance rationale that might explain why a skilled artisan would want some caching somewhere, but not why the artisan would restructure Sapuntzakis and Holzmann to arrive at the claimed shared-root/per-node-leaf cache configuration. ’844 Response at 65. Dr. Newman similarly opined that the Office Action did not provide a technically grounded motivation to combine Sapuntzakis, Holzmann, and Federwisch into a coherent architecture that achieved the claim-required cache configuration and indexing-result sharing. Newman Decl. ¶ 220.

For claims 19–27, the declaration and response addressed the claimed “index once/share across compute nodes” limitation. The patent owner argued that Federwisch’s cited indexing concerned snapshot-difference detection and destination-map maintenance, not indexing a root image and providing those indexing results to another compute node. ’844 Response at 77–82. Dr. Newman’s declaration likewise framed Federwisch’s “index” as an internal snapshot-comparison construct, not root-image indexing results distributed across compute nodes. Newman Decl., table of contents and discussion of Ground 2.

The examiner ultimately issued a Notice of Intent to Issue Ex Parte Reexamination Certificate confirming claims 1–27. Notice of Intent to Issue Ex Parte Reexamination Certificate, Ex Parte Reexamination of U.S. Patent No. 8,332,844, Control No. 90/019,938, at 1–2 (May 19, 2026) [hereinafter ’844 NIRC].

The practitioner lesson is that architecture-level declarations can prevent an Office Action from reducing specific claim limitations to generic technical labels. The issue was not whether the prior art had “caching” or “indexing.” The issue was whether the prior art disclosed the claimed cache and indexing operations in the claimed root-image/leaf-image architecture.

VI.      Case Study Three: The ’582 Patent — Abstraction-Level Expert Evidence

The third case study involves U.S. Patent No. 7,257,582, titled “Load Balancing with Shared Data,” in Ex Parte Reexamination Control No. 90/015,337. Patent Owner submitted the Declaration of Dr. Daniel Abadi in support of its response to the non-final Office Action. Declaration of Dr. Daniel Abadi, Ex Parte Reexamination of U.S. Patent No. 7,257,582, Control No. 90/015,337 (Feb. 17, 2026) [hereinafter Abadi Decl.].

This case illustrates a third use of expert testimony: fighting the wrong level of abstraction. The Office Action treated generic partitioning, tiling, and load balancing as sufficient. The patent owner reframed the claimed invention as a specific file-partition allocation architecture involving determination of file allocation, logical partitioning of records of a preexisting input file, distribution of partition descriptions, execution of subtasks on partitions, and repeated allocation of another unprocessed partition on a first-come/first-served basis. Response to Office Action, Ex Parte Reexamination of U.S. Patent No. 7,257,582, Control No. 90/015,337, at 1–3 (Feb. 17, 2026) [hereinafter ’582 Response].

Dr. Abadi’s declaration explained that Reeve’s “partitioning” was iteration-space tiling defined by indices, bounds, and tile size directives, not automatic determination of file allocation or subdivision of records of a preexisting input file into file partitions. Abadi Decl. ¶¶ 121–23.

The declaration sharpened the distinction: the ’582 Patent partitions data, while Reeve partitions computation. Claim 1 required partitions of a preexisting file’s records based on file-allocation information; Reeve described tiling an iteration space for scheduling computation. Abadi Decl. ¶¶ 162–64.

The declaration also addressed the “first come first served” limitation. Dr. Abadi explained that Reeve’s FCFS/GRAB strategy assigned iteration-space tiles to threads, while claim 1 used FCFS for repeated allocation of unallocated file partitions after completion of earlier partition processing. Abadi Decl. ¶¶ 166–67.

On motivation to combine, Dr. Abadi explained that a skilled artisan would not have been motivated to combine Iyer’s parallel sorting and merging technique with Reeve’s iteration-space tiling and GRAB scheduling in the manner proposed because the references addressed different computational problems, operated on different decomposition targets, and embodied different execution models. Abadi Decl. ¶¶ 124–25.

The examiner later issued a Notice of Intent to Issue Ex Parte Reexamination Certificate confirming claims 1–14. Notice of Intent to Issue Ex Parte Reexamination Certificate, Ex Parte Reexamination of U.S. Patent No. 7,257,582, Control No. 90/015,337, at 1–5 (Apr. 6, 2026) [hereinafter ’582 NIRC]. The NIRC interpreted the “logically subdividing” and “repeating” limitations consistently with the patent owner’s technical framing, including the distinction between logical file partitioning and actual reading by subtasks. Id.

The lesson is that expert testimony can prevent an examiner or requester from defining the problem too broadly. “Parallelism” was not enough. The relevant question was whether the references taught the claimed file-partition load-balancing architecture.

VII.   Limits and Risks of Expert Declarations

A sophisticated patent-owner strategy should recognize that expert declarations are powerful but not risk-free. A declaration should be used when it strengthens the factual record, not as a reflexive add-on or as a way to disguise attorney argument as expert testimony.

That point is especially important in reexamination because claim amendments often are not a practical option in the same way they may be during ordinary prosecution. Although a patent owner may propose amendments or new claims in ex parte reexamination, no proposed amended or new claim may enlarge the scope of the patent claims. 35 U.S.C. § 305; MPEP § 2258. Any amended or new claim incorporated into the patent after reexamination also has the effect specified in 35 U.S.C. § 252, including potential intervening-rights consequences. 35 U.S.C. § 307(b); 35 U.S.C. § 252. For an asserted patent, even a narrowing amendment may invite arguments that the amended claim is not “substantially identical” to the original claim, potentially affecting past damages or defenses in pending litigation. See Bloom Eng’g Co. v. N. Am. Mfg. Co., 129 F.3d 1247, 1250–51 (Fed. Cir. 1997); Laitram Corp. v. NEC Corp., 163 F.3d 1342, 1346–48 (Fed. Cir. 1998). Thus, while amendments remain procedurally available, they may be strategically unavailable in a reexamination involving related district-court litigation, licensing disputes, or PTAB proceedings.

That reality makes the expert declaration more important, but it also raises the stakes. A declaration may allow the patent owner to preserve the original claim language while rebutting the examiner’s factual premises with evidence. But the declaration itself becomes part of the prosecution history. Statements made in reexamination may later affect claim construction, disclaimer, infringement positions, or invalidity defenses. A technically persuasive distinction over the prior art can become problematic if it is broader than necessary or inconsistent with how the patent owner needs to read the claims in litigation. See Krippelz v. Ford Motor Co., 667 F.3d 1261, 1266–67 (Fed. Cir. 2012); Tempo Lighting, Inc. v. Tivoli, LLC, 742 F.3d 973, 977–78 (Fed. Cir. 2014).

A weak declaration can hurt more than help. A declaration that is conclusory, over-lawyered, or detached from the claim language may receive little weight and may invite the examiner to dismiss the evidence as unsupported. MPEP § 716 explains that opinion testimony may receive weight when supported by an underlying factual basis, but that legal conclusions themselves are not entitled to weight. MPEP § 716; see also In re Brandstadter, 484 F.2d 1395, 1406–07 (C.C.P.A. 1973).

Declarations also must be coordinated with parallel proceedings. Where the same patent is in district-court litigation, IPR, licensing negotiations, or another post-grant proceeding, the reexamination declaration should be aligned with existing claim-construction positions, infringement theories, invalidity responses, expert reports, and PTAB submissions. A declaration that solves the immediate Office Action but creates inconsistency elsewhere can impose a larger strategic cost than the rejection itself.

The expert should not become a substitute brief. The attorney should make the legal argument; the expert should supply the technical facts and technical opinions that support it. A declaration that reads like attorney argument under an expert’s signature can lose credibility. The better practice is to have the expert explain the technology, the references, the claim language, and the technical consequences of the proposed combination, while the response explains why those facts defeat the rejection.

The counterargument is not that every non-final Office Action requires expert testimony. Some rejections can be overcome through the intrinsic record, a straightforward reference distinction, or a narrow issue of claim interpretation. In ordinary prosecution, a clean amendment may also be the most efficient path. In reexamination of an asserted patent, however, amendment may carry intervening-rights and litigation consequences that make it far less attractive. ( see Same Office Action, Different Stakes: Responding to USPTO Office Actions in Ex Parte Reexamination and Ordinary Patent Prosecution)  Where the rejection depends on a technical characterization of the prior art or an asserted rationale to combine, and where amendment would risk narrowing the claims in a way that affects related litigation, the absence of expert evidence may leave the patent owner arguing facts without evidence while also lacking a practical amendment path.

VIII.                 Practical Drafting Implications

A strong reexamination declaration should be targeted, not encyclopedic. It should identify the expert’s qualifications, the materials reviewed, the level of ordinary skill, the relevant claim-construction assumptions, the technical background, the reference teachings, and the technical flaws in the rejection.

The declaration should also be coordinated with the attorney response. The response should cite declaration paragraphs precisely and explain how the expert’s facts defeat the prima facie case. Where the issue is obviousness, the response should tie the declaration to the examiner’s articulated rationale under MPEP § 2143. The declaration should attack the factual predicates; the response should explain the legal consequence.

The best declarations also avoid unnecessary admissions. They should not define claim terms more narrowly than needed. They should not criticize prior art in a way that conflicts with infringement positions. They should not adopt a litigation expert’s broader report wholesale. And they should not turn into attorney argument under another name.

IX.      Conclusion

In ex parte reexamination, the patent owner is not merely trying to persuade an examiner. The patent owner is building a record that may affect confirmation, appeal, later USPTO proceedings, and parallel litigation. A well-crafted expert declaration can be central to that record.

The expert declaration matters because it supplies evidence where attorney argument cannot. It matters because obviousness requires articulated reasoning, and expert testimony can attack the factual predicates of that reasoning. The strongest expert testimony does not simply say that the references are different. It explains why the differences matter to the claimed invention.

But the tool must be used carefully. A conclusory or over-lawyered declaration may receive little weight and may create prosecution-history risk. The expert should not replace the brief. The expert should provide the technical facts that allow the attorney’s legal argument to carry persuasive force.

In a non-final reexamination response, that distinction can be decisive. The patent owner who builds the factual record early gives the examiner a reason to reconsider the rejection, gives the Board a record to review if appeal becomes necessary, and gives later tribunals a clearer account of what the claims mean and why the prior art does not meet them.