US 5931919

I'll search for this specific patent number as instructed, then summarize.

US Patent 5,931,919 — Verification Summary

I searched for the literal identifier 5931919 (US 5,931,919 / application 08/966,405). Findings below; I flag where I lack authoritative confirmation.

Bibliographic data

Note: the EP family publication (EP 0917051 A3) lists the same four inventors with Southern-California addresses, which is consistent with the US filing and corroborates the assignment record (REEL/FRAME 008809/0206, executed 1997-11-07).

Uncertainty flagged: the examiner of record is inconsistently reported across secondary databases ("Shahid Alam" and "Wayne Amsbury" both appear on the Unified Patents portal). I cannot authoritatively confirm a single examiner name, so I treat this as unverified.

Abstract (verbatim)

"Temporary states are used transitionally in run-time situations and are unknown to the object database. A temporary state is created if, when an object is performing a requested event, interim work needs to be performed before the object reaches a permanent destination state. Use of a temporary state is transparent to the caller of the requested event. Within a single flow of control, an unlimited number of nested locks may be acquired. A request to unlock an object will only truly unlock the object if the unlock requester is at the highest level of locking. For recursive or embedded functions, no knowledge of the previous locking condition is necessary. Attribute-based locking provides a mechanism for allowing only certain members of the system to lock an object for writing while the object is in a predetermined state. However, any thread may obtain a read lock on an object in order to examine its attributes. This may be expanded to state that the system may restrict locking of an object or entity based upon any attribute in the database or upon any combination or logical expression of these attributes."

The patent has 5 claims, of which claim 1 and claim 5 are independent. Claims 2–4 depend from claim 1.

Plain-language overview of the independent claims

Claim 1 — Attribute-based locking method (the operative claim). A first object receives a request to perform an action. It write-locks itself, then moves from a first state into a transition state. While in that transition state it (a) unlocks itself and (b) spawns a first thread to carry out the work needed for the requested action. That spawned thread then asks for a write-lock and is granted one with "special permission" — i.e., it is permitted to write-lock the object while the object is still in the transition state, which ordinary threads are not. On completion, the object moves from the transition state to a non-transition state and unlocks. Crucially, threads that are not part of the spawned thread are denied locks on the object while it remains in the transition state, and must wait until the object reaches the non-transition state.

Claim 5 — Broad method claim. An object moves from a first state to a transition state to perform predetermined operations. The transition state locks the object against access by threads not related to the transition operations (note: this is stated as an effect of the state itself, not explicitly as a write-lock acquired by a thread). The object then transitions to a second state and unlocks when the transition-state operations complete.

Dependent claims 2–4 add: (2) actions along the spawned thread may request additional locks on the object while in the transition state; (3) unrelated threads are blocked "until the object is in the second state"; (4) the "state" may be any combination or logical expression of attributes.

Drafting/scope observations

• Claim 5 is substantially broader and more loosely worded than claim 1 — it omits the write-lock-by-itself, the spawn step, and the special-permission mechanism, and instead asserts that the transition state itself "takes" a lock. This makes claim 5 the more expansive (and potentially more vulnerable) claim.
• Claim 3 refers to "the second state" although claim 1 recites a "non-transition state" and no "second state" — a possible antecedent-basis/consistency issue worth noting for any validity analysis.
• The specification repeatedly describes the broader principle that locking decisions may be keyed to any attribute or logical expression of attributes, which claim 4 captures for claim 1 only.

Docket / litigation check

I searched specifically for implementation before the CAFC in 2026 referencing 5931919 and found no CAFC 2026 docket, appeal, or oral-argument entry for this patent.

Two cautions:

1. One search returned a secondary-source mention of "the '593 patent" in a ParkerVision/MediaTek–related prospectus filing (with a June 1, 2026 CAFC oral argument in ParkerVision v. Qualcomm). That reference is not to 5,931,919 — it is a different patent whose last three digits happen to be 593. Per your instruction not to auto-correct or conflate identifiers, I did not merge these.
2. The 5,931,919 patent expired 2017-11-07, so it is unlikely to be the subject of 2026 appellate activity; anything you see cited as "the '593 patent" in 2026 litigation should be independently verified against its full number.

Related family (for context)

Continuation/related Xerox filings sharing the 1997-11-07 priority include US 6,546,412 B1 ("State-based object transition control and nested locking") and US 6,971,100 B2, plus published application US 2002/0107998 A1. These appear to be separate patents covering the nested locking aspects described in this specification's FIG. 4.

Sources: the authoritative full text at https://patents.google.com/patent/US5931919/en; Unified Patents portal https://portal.unifiedpatents.com/patents/patent/US-5931919-A; EP family publication https://patentimages.storage.googleapis.com/15/10/76/29ae9e2c77cb38/EP0917051A3.pdf; USPTO.report grant page https://uspto.report/patent/grant/5931919.

I'll search litigation databases for this specific patent number.

Let me search the litigation-specific databases directly.

Let me run a few more targeted searches for litigation records.

Litigation search results for US Patent 5,931,919

Finding: No litigation involving US Patent 5,931,919 was identified. I searched litigation-oriented sources specifically for the literal number 5931919 and found no district court case, ITC investigation, PTAB proceeding, or CAFC appeal naming this patent.

What I searched and what came back

Important negative results (near-misses — do NOT conflate)

My searches surfaced several cases involving different patents with "919" in the number. Under your instruction to interpret identifiers literally, I am flagging these as not US 5,931,919:

• RE43,919 ("Baby Crib") — Wonderland Nurserygoods v. Baby Trend, C.D. Cal. No. 5:14-cv-01153-JWH-SP (Baby Trend summary judgment of invalidity granted on the original-patent requirement, Feb. 13, 2023); also IPR2015-00841. Different patent (reissue of 6,859,957).
• 7,012,919 and 8,243,593 — asserted by Sable Networks/Sable IP against Cloudflare, W.D. Tex. No. 6:21-cv-00261. Note the "5919"-adjacent digit strings and the "'919"/"'593" short names — these are not US 5,931,919.
• 8,032,919 — Sony/ARRIS, IPR2017-01699; 9,522,919 — Purdue Pharma/OxyContin ANDA litigation. Both different patents.
• Unistroke "593"-style short cites — the ParkerVision/MediaTek-type "'593" references are to a different patent entirely, consistent with the caution already flagged in the previously generated summary.

Why no litigation is expected

Two structural reasons corroborate the null result:

1. Expiration. Per the authoritative family/legal-status data, 5,931,919 expired 2017-11-07 (anticipated expiration; "Expired – Lifetime"). Assertion of an expired patent is foreclosed prospectively, so any infringement action would have had to be filed before late 2017 — and none appears.
2. Ownership profile. The patent was held throughout by Xerox Corporation (with security interests recorded to Bank One, N.A. and JPMorgan Chase, later released 2022). It was never transferred to an NPE/assertion entity, which is consistent with the absence of an assertion campaign. The related continuation family — US 6,546,412 B1, US 6,971,100 B2, and published app US 2002/0107998 A1 ("State-based object transition control and nested locking") — is likewise listed under Xerox as assignee; I found no litigation asserting those either.

Limitations and confidence

• I have high confidence there is no currently pending or historically reported litigation on 5,931,919, given expiration and the absence of any docket-level hit across the sources searched.
• Caveat: I could not query PACER directly, and full historical district-court dockets (especially pre-2010 cases that predate broad online docket coverage) are not exhaustively indexed by the free sources I reached. It remains possible but unconfirmed that a short-lived, quietly resolved pre-2017 action exists in a court's records that did not surface. I am not aware of any, and I am stating that as an absence of evidence rather than a positive assertion of no-case-ever.
• I did not auto-correct any near-miss identifiers; every "919" case above is reported under its own distinct patent number.

Bottom line: Based on the sources searched, no litigation is known involving US Patent 5,931,919 — no plaintiff, defendant, jurisdiction, case number, filing date, or outcome can be reported because no such case was located.

I'll verify the structured "no proceedings" default against live sources before writing anything.

Nothing surfaced yet tying 5931919 to any AIA trial. Let me probe a few more angles — CBM/reexam, litigation assertions, and any third-party challenger.

One more pass to make sure I'm not missing a challenge indexed under the application number or on a PTAB aggregator.

Proceedings overview

Zero AIA trial proceedings. The canonical USPTO Open Data Portal block returns no IPR, PGR, or CBM for US 5,931,919, and independent web searches corroborate that: 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denials. Bottom line for a defendant: nobody ever filed a PTAB challenge against this patent — and that is not a "hardened patent" signal, it is an expiration signal. The patent issued 1999-08-03, expired 2017-11-07 (20-year term from the 1997-11-07 filing), and the § 286 six-year damages lookback from today (2026-10-01) reaches back only to 2020-10-01 — after expiration. If a demand letter cites claims 1–5, your primary problem is not invalidity, it is that the patent is expired and the recoverable-damages window is closed.

---

No proceedings to enumerate

The per-proceeding template does not apply, because there is nothing to fill it. I state that plainly rather than manufacturing structure for a null set.

What I checked (2026-10-01):

Uncertainty, stated explicitly: I could not execute a direct query against PTAB E2E or the full Docket Alarm PTAB index from this session. My negative finding rests on the ODP structured block plus multiple corroborating web searches. That is strong evidence for "no AIA trials," but it is not the same as having paginated the entire E2E docket. Also note that pre-AIA inter partes reexamination (a CRU proceeding, not an AIA trial) is a different animal and I found no evidence of one either — but I cannot fully exclude it from the sources searched. Treat "no reexam" as unverified.

Adjacent proceedings — ⚠️ NOT this patent, do not conflate

Searches for PTAB activity involving Xerox returned a large volume of 2021–2025 IPR and Federal Circuit activity. None of it involves 5,931,919. Anyone assembling a docket picture for this patent must keep these separate:

• IPR2021-01264 (Meta/Facebook v. Xerox) — U.S. Patent 7,043,475; FWD 2023-01-23, all challenged claims unpatentable; CAFC No. 2023-1714 / 2023-1913; affirmed 2025-06-17 (Xerox Corp. v. Meta Platforms, Inc., CourtListener opinion).
• IPR2021-01294 / IPR2021-01458 / IPR2021-00987 (Meta, X/Twitter, Snap v. Xerox) — U.S. Patent 8,489,599; consolidated FWD 2023-03-24, claims 1–25 unpatentable; CAFC Nos. 2023-1983/1985/1988/2028, affirmed 2025-07-24.
• IPR2021-01434 — U.S. Patent 8,732,584; FWD 2023-03-15, all challenged claims unpatentable.
• IPR2021-01461 — U.S. Patent 9,208,439; instituted 2022-03-24.

These are Xerox's content-delivery/context patents asserted against social-media defendants in C.D. Cal. They share only an assignee with 5,931,919. The "'599 patent" in that litigation is U.S. 8,489,599 — not 5,931,919. (My earlier summary flagged the same trap in a ParkerVision filing; same discipline applies here.)

Family note: the continuation/related filings sharing the 1997-11-07 priority — US 6,546,412 B1 and US 6,971,100 B2 ("State-based object transition control and nested locking") and published app US 2002/0107998 A1 — likewise are separate patents. I found no PTAB proceedings on those either, but I did not exhaustively verify them; that would require a targeted E2E pull per patent number.

---

Strategic summary

Claim status: all of claims 1–5 are UNTESTED at the PTAB — and all are expired. There is no FWD to quote, no claim canceled, no claim sustained in a trial. Claims 1 and 5 (the independents) and dependent claims 2–4 stand exactly as issued on 1999-08-03, never having been through an AIA trial. The validity posture of this patent is therefore whatever a district court would make of it, unassisted by any PTAB record. As noted in my earlier claim analysis, claim 5 is considerably broader and more loosely drafted than claim 1, and claim 3 has a visible antecedent-basis gap ("the second state" versus claim 1's "non-transition state") — if this patent were live, those would be the attack points. It is not live.

Estoppel landscape: there is none, and it doesn't matter. Because no IPR/PGR was ever instituted, § 315(e)(2) estoppel never attached to anyone. Any defendant facing an assertion today can raise any prior-art ground — § 102, § 103, § 112 — in district court without restriction, including grounds that would have been "reasonably could have raised" in an IPR. That is a theoretical advantage only, because the real defense is threshold: the patent is expired as of 2017-11-07. Concretely, a complaint filed today (2026-10-01) can recover damages only for infringement on or after 2020-10-01 under § 286 — a window in which the patent was already dead. Pre-suit damages for the entire life of the patent are time-barred. (This is a general statutory principle I'm applying to the known expiration date, not a case-specific holding; confirm with a litigation partner and check for any tolling agreement or earlier-filed complaint that could change the calculus.)

Pattern signals: absent, and the absence is explained by economics, not by patent strength. No petitioner filed serially; no repeat-challenger pattern; no defensive aggregator (Unified Patents shows a bibliographic page but no filing on this number) ever stepped in. The patent's forward citations are mostly Xerox's own continuations and Flexiworld/Chang mobile-printing patents — a 1990s internal locking/state-machine disclosure, not a commercially asserted asset. The reason no IPR exists is the same reason you should not see an IPR: filing a $500K+ IPR against a patent that expired in 2017 and yields no recoverable damages is irrational. Contrast the Xerox social-media campaign above, where live patents drew four-plus IPRs within four months of suit.

---

Recommended next steps

If you are a defendant and a demand letter or complaint cites 5,931,919:

1. Lead with expiration, not invalidity. Check the face of the patent: filed 1997-11-07, issued 1999-08-03, expired 2017-11-07. The 20-year term is confirmed by the Google Patents legal-events entry ("2017-11-07 Anticipated expiration") and the Unified Patents expiry field. No term extension, no PTA, no continuation of the term applies here.
2. Close the § 286 damages window. Today is 2026-10-01; the six-year lookback opens 2020-10-01, after expiration. Unless there is an earlier-filed complaint or a tolling agreement resurrecting exposure, recoverable damages are effectively zero. Ask plaintiff's counsel in writing for the filing date of the first complaint asserting this patent and for any tolling agreement.
3. There is no FWD to cite — say so explicitly if opposing counsel implies the patent has been "tested." It has not been tested at the PTAB, and it also cannot be. An IPR filed now would be pointless (no amendability, no live commercial exposure), so do not budget for one.
4. Do not let the "'599 patent" trap flip this. If the demand cites "the '599 patent," verify the full number. U.S. 8,489,599 (Xerox content-delivery) is invalidated — claims 1–25 unpatentable, affirmed by the Federal Circuit on 2025-07-24 in CAFC No. 23-1983 — but it is a different patent from 5,931,919. The invalidation in that case gives you no leverage on this one, and vice versa.
5. If any live assertion is threatened in the next 12 months, the only remaining defensive work is (a) a § 286 limitations defense, (b) an on-sale/public-use or § 101 eligibility challenge if the plaintiff somehow survives the damages bar, and (c) a check for prior art on this patent's own claims — none of which has ever been adjudicated. That art would have to be assembled from scratch, since no petitioner ever did it.

If no proceeding or assertion exists — which is the current state as of 2026-10-01 — the correct answer is the one above: no PTAB activity on file, and no reason to expect any. For a patent that expired in 2017, silence at the PTAB is the expected outcome, not a signal of strength.

I'll verify the assignment record against the authoritative source data I already have, then cross-check with search.

Inventors

All four are named as assignors on the original assignment, executed on the filing date itself (1997-11-07) — the standard "hereby assign all right, title and interest" employment assignment. There is no evidence of inventor departure or subsequent inventor-side re-recordation; none of the four appears again in the chain. As noted in the prior section, the EP family publication lists Southern-California addresses for the inventors, which is consistent with a Xerox facility outside the Palo Alto/Norwalk core (Xerox historically operated document-systems engineering in the El Segundo, CA area). I cannot verify individual California addresses from the assignment record itself, so treat the location as unconfirmed.

Original assignee

Xerox Corporation (Connecticut corporation; the assignment recites "XEROX CORPORATION, CONNECTICUT"). Xerox is a workplace-technology company — historically copiers, printers, and document management, now a services-led, software-enabled print/IT-solutions business per its latest Form 10-K. It is still operating (NASDAQ: XRX via parent Xerox Holdings Corporation); the most recent filings show the July 1, 2025 Lexmark acquisition and continued restructuring ("Reinvention"), with no bankruptcy filing, no Chapter 7/11, and no patent estate sale.

Whether Xerox shipped a product embodying these claims is not determinable from the sources I have. The patent claims are directed to object state-model transition and locking (job/document/printer objects in FIG. 4 strongly suggest a print-workflow / job-management system), so an internal Xerox document-processing system is plausible — but "plausible" is not evidence, so I mark this unclear rather than assert it.

Assignment timeline

The USPTO Assignment Center record (as surfaced through the Google Patents legal-events mirror of the same reels/frames) contains four recorded post-issuance events. Correspondent-of-record data is not exposed in the sources available to me — the Google Patents legal-events table surfaces reel/frame, conveyance, parties, and execution date, but not the recording attorney/correspondent. I therefore cannot populate the "Repeat correspondent" signal below and flag it explicitly as a data gap rather than guess.

• 1997-11-07 (executed) / recorded 1997-11-07 — Reel 008809/0206

  • Conveyance: Assignment
  • Assignor: Jennifer D. Thomas; Sang W. Kim; David P. Nesbitt; Steve A. Okamoto (individually)
  • Assignee: Xerox Corporation (Connecticut)
  • Correspondent: not available in source data
  • Context: Initial employment assignment of inventors' rights to the original assignee, executed on the filing date.

• 2002-06-21 (executed) / recorded 2002-06-28 — Reel 013153/0001

  • Conveyance: Security Interest (Assignment record type)
  • Assignor: Xerox Corporation
  • Assignee: Bank One, NA, as Administrative Agent (Illinois)
  • Correspondent: not available in source data
  • Context: Securitization — Xerox granting a security interest over IP as collateral, not a transfer of ownership.

• 2003-06-25 (executed) / recorded 2003-10-31 — Reel 015134/0476

  • Conveyance: Security Agreement
  • Assignor: Xerox Corporation
  • Assignee: [JPMorgan Chase Bank, as Collateral Agent (Texas)](/asserters/jpmorgan-chase-bank-as-collateral-agent-texas)
  • Correspondent: not available in source data
  • Context: Securitization — successor collateral-agent security agreement over the same Xerox IP collateral; again a lien, not an ownership change.

• 2022-08-22 (executed) / recorded 2022-09-07 — Reel 066728/0193

  • Conveyance: Release By Secured Party
  • Assignor: JPMorgan Chase Bank, N.A. (as successor-in-interest administrative/collateral agent to JPMorgan Chase Bank)
  • Assignee (owner of record): Xerox Corporation (Connecticut)
  • Correspondent: not available in source data
  • Context: Release/lien discharge — the security interest is terminated and Xerox's record title is cleared. No third party ever became owner.

Key finding: there is no ownership transfer away from Xerox anywhere in the recorded history. The only true assignment of title is the 1997 inventor→Xerox assignment; the 2002 and 2003 entries are liens, and the 2022 entry discharges them. No LLC, no holding company, no asserter, and no defensive aggregator ever appears.

Timeline diagram

timeline
    title Ownership of US 5931919
    1997 : Inventors assign rights to Xerox
    1999 : Patent issues to Xerox
    2002 : Security interest to Bank One
    2003 : Security agreement with JPMorgan Chase
    2017 : Patent expires by term
    2022 : JPMorgan releases security interest

NPE / troll-pattern signals

1. Shell-entity transfer — Not present. No transfer to any "IP/Patents/Licensing/Holdings/Ventures" entity appears in any of the four recorded events (Reels 008809/0206, 013153/0001, 015134/0476, 066728/0193). Title remains with Xerox Corporation, an operating company.

2. Known asserter in the chain — Not present. Neither Xerox Corporation, Bank One, nor JPMorgan Chase appears on any NPE/asserter list (Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, etc.). The two non-Xerox names are banks acting as collateral agents — a securitization role, not an assertion role.

3. Repeat correspondent across the chain — Unclear / data gap. The correspondent of record is the single most diagnostic field for this analysis, and it is not exposed in the sources available to me; I will not invent attorney names or firms. To close this, the analyst should pull each reel/frame directly in USPTO Assignment Center (search patent 5,931,919) and record the "Correspondent" field. Given that three of the four events are ordinary Xerox corporate finance recordings, a repeat correspondent here would most likely reflect Xerox's outside IP-recording counsel rather than an NPE handler — so even if recurrence exists, it is unlikely to be an NPE tell in this chain.

4. Cascading transfers — Not present. No chained LLC-to-LLC transfers; the two transfers within roughly 24 months (2002-06-28 and 2003-10-31 recordings) are two successive collateral-agent liens on the same grantor's collateral, not a shell cascade.

5. Pre-litigation transfer — Not present. I found no infringement suit naming US 5,931,919, and in any event the patent expired 2017-11-07, so no pre-suit transfer could exist. The chain shows no assignment within six months before any asserted suit.

6. Bankruptcy fire-sale — Not present. Xerox Holdings / Xerox Corporation is not in bankruptcy (latest FY2025 Form 10-K filed under a going-concern basis; debt-reduction and restructuring activity, but no Chapter 7/11 and no patent estate sale). The patent's 2017 expiry also predates any recent Xerox financial stress.

7. Privateering — Not present. No transfer to an NPE asserting on Xerox's behalf; Xerox retained title throughout.

8. Defensive aggregator — Not present. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN. It terminates at the original operating assignee, Xerox.

Verdict

Insufficient data (effectively only the original assignment). Justify: the recorded chain consists solely of the inventor→Xerox assignment on 1997-11-07 (Reel 008809/0206), two collateral liens to Bank One and JPMorgan Chase (Reels 013153/0001, 015134/0476), and a 2022 lien discharge back in favor of Xerox (Reel 066728/0193) — no third party ever acquired title, so there is no NPE pattern to assess. The patent also expired 2017-11-07, a decade before any current assertion, which independently removes it from NPE/assertion relevance. The one item that could sharpen this call — the correspondent of record on each reel/frame — could not be extracted from my sources and should be confirmed directly at the USPTO Assignment Center: https://assignmentcenter.uspto.gov/ (mirror: https://assignment.uspto.gov/patent/index.html), searching by patent number 5931919.

I'll verify the patent and its cited references against live sources before assessing each one.

Prior-Art Analysis — US 5,931,919

Patent under analysis: US 5,931,919 A ("State-based object transition control and attribute-based locking"), App. 08/966,405, filed 1997-11-07, granted 1999-08-03, Xerox Corp. Expired – Lifetime (anticipated expiration 2017-11-07).

Scope of this memo and an important threshold caveat

The user asked which cited reference "potentially anticipates" each claim under 35 U.S.C. § 102. A § 102 anticipation requires a single reference to disclose every limitation of a claim, arranged as claimed. As my own earlier analysis established, the operative claims are narrow and combine three distinct concepts:

Result of the review: None of the twelve examiner-cited U.S. patents discloses the conjunction of a runtime-only transition/transient state with a special-permission write-lock granted to a spawned thread while the object sits in that state. Every reference below is therefore best characterized as § 103 (obviousness) material or § 102 background rather than a true single-reference anticipation of claims 1–5. I state element-level overlap where it exists and flag it explicitly where it does not. (This contradicts nothing in my prior sections — it refines the "relevant prior art" question that was not yet answered.)

All twelve U.S. references were cited by the examiner on the 5,931,919 face. Dates and abstracts below were re-verified against Google Patents / Justia / USPTO.report records.

---

A. Examiner-cited U.S. patent references (12)

1. US 5,008,853 A — Xerox Corp.

• Filed/issued: 1987-12-02 / 1991-04-16
• Statutory basis: § 102(a)/(b) (pre-filing publication)
• Description: "Representation of collaborative multi-user activities relative to shared structured data objects in a networked workstation environment." Shared structured data objects with concurrency/access control at the workstation level.
• § 102 relevance: Background only. Discloses shared objects and multi-user access control broadly. Does not disclose a transition state, self-unlock-while-transitional, or thread-spawned special-permission write-locks. No complete claim is anticipated.
• Potential claim overlap: none meaningful (background for the "object" environment of claims 1/5).

2. US 5,161,227 A — IBM Corp. (Dias et al.)

• Filed/issued: 1989-11-13 / 1992-11-03
• Statutory basis: § 102(a)/(b)
• Description: "Multilevel locking system and method." A two-level (LE1/LE2) data-access manager: a global table records the states of all lock entities; LE1 checks compatibility of a request against the lock-entity state and grants/denies; denied requests are queued at LE2 while the requesting task is suspended; a "global table also includes an indication of exclusive lock status."
• § 102 relevance: This is lock-state and lock-hierarchy art, not object-state-transition art. It does disclose the concept of a lock entity that itself has a state and of compatible/incompatible request modes — element-level overlap with the "state" and "write-lock request" concepts of claims 1 and 5.
• Potential claim overlap: Claim 1/5 only weakly (multi-level lock management; shared vs. exclusive lock state). No anticipation — no transition state, no spawn, no special-permission exemption.

3. US 5,226,143 A — IBM Corp.

• Filed/issued: 1990-03-14 / 1993-07-06
• Statutory basis: § 102(a)/(b)
• Description: "Multiprocessor system includes operating system for notifying only those cache managers who are holders of shared locks on a designated page by global lock manager." Global lock manager tracks holders of shared locks and notifies only interested cache managers.
• § 102 relevance: Tracks which threads/entities hold which lock mode and selectively notifies them — peripheral element-level overlap with the "certain members only may lock for write" idea of claim 1 and with read-lock availability in claim 5.
• Potential claim overlap: Claim 5 tangentially (selective lock participation). No anticipation.

4. US 5,452,447 A — Sun Microsystems, Inc.

• Filed/issued: 1992-12-21 / 1995-09-19
• Statutory basis: § 102(a)/(b)
• Description: "Method and apparatus for a caching file server." Distributed file-server cache coherence with locking of cached data.
• § 102 relevance: General file/object locking background. No relation to transition states, spawned-worker locking, or attribute-conditional locking.
• Potential claim overlap: none meaningful. No anticipation.

5. US 5,454,108 A — IBM Corp. (Devarakonda & Mohindra) — high relevance

• Filed/issued: 1994-01-26 / 1995-09-26 (EP counterpart EP 0 665 495 A2/A3)
• Statutory basis: § 102(a)/(b)
• Description: "Distributed lock manager using a passive, state-full control-server." Local lock managers must acquire a token from a lock-control server that is the owner of the resource and is "passive and has all the control information for a resource it owns." If no conflicting token exists, the server grants it; if conflicts exist, the server returns the list of conflicting holders, the requester negotiates give-up/acknowledge, then the server updates its table and grants the token.
• Verified detail: the specification expressly describes a lock state table with a "mode field" and a notion that a lock "is in transition" (e.g., read→write, acquiring), per the uspto.report full text of 5,454,108.
• § 102 relevance: This is the closest cited art on the "state-full" / state-of-the-lock / in-transition axis and on conditional grant based on state compatibility. It maps to the idea of a special state during which ordinary grants are withheld.
• Potential claim overlap: Claim 1 and claim 5 most plausibly (a lock that is granted/denied conditional on the lock's own "state," including an "in transition" state, held by a designated owner). Still not a complete anticipation — 5,454,108 governs lock-state, not object-state lifecycle transitions, and has no "object unlocks itself in the transition state, spawns a thread, and re-grants that thread a special-permission write-lock" sequence. Best cited in a § 103 combination with a transient-state reference (see #7).

6. US 5,481,721 A — Next Computer, Inc.

• Filed/issued: 1991-07-17 / 1996-01-02
• Statutory basis: § 102(a)/(b)
• Description: "Method for providing automatic and dynamic translation of object oriented programming language-based message passing into operation system message passing using proxy objects." Proxy-object message translation.
• § 102 relevance: Object-oriented event/message background only. No locking, no state transitions, no threads.
• Potential claim overlap: none. No anticipation.

7. US 5,727,203 A — Sun Microsystems, Inc. — high relevance

• Filed/issued: 1995-03-31 / 1998-03-10
• Statutory basis: § 102(e) (granted after the 1997-11-07 filing date, but filed well before it)
• Description: "Methods and apparatus for managing a database in a distributed object operating environment using persistent and transient cache." Distinguishes persistent object data from transient (runtime-only) representations.
• § 102 relevance: This is the best cited art on the "runtime-only / not-in-the-object-database" character of the temporary state of 5,931,919's FIG. 3 and its abstract ("used only during runtime and is unknown to the object database").
• Potential claim overlap: Claim 5 (a transition state that exists to perform predetermined operations, distinct from the persistent states). No complete anticipation — 5,727,203's transience concerns cache residency, not object state-machine transition states with an unlock/spawn/special-permission lock protocol.

8. US 5,729,710 A — IBM Corp.

• Filed/issued: 1994-06-22 / 1998-03-17
• Statutory basis: § 102(e)
• Description: "Method and apparatus for management of mapped and unmapped regions of memory in a microkernel data processing system." Memory-region mapping.
• § 102 relevance: Remote. Possibly background for "transient/mapped vs. unmapped" state duality, but no locking or object state model.
• Potential claim overlap: none meaningful. No anticipation.

9. US 5,758,184 A — [Microsoft Corp.](/litigations/by-plaintiff/Microsoft%20Corp.)

• Filed/issued: 1995-04-24 / 1998-05-26
• Statutory basis: § 102(e)
• Description: "System for performing asynchronous file operations requested by runnable threads by processing completion messages with different queue thread and checking for completion by runnable threads." A requesting thread issues an async operation; a different queue thread processes completion messages; the requesting thread later checks for completion.
• § 102 relevance: This is the cited art closest to the "spawning a first thread … to perform operations in response to the first action" element of claim 1, and to the asynchronous worker/completion pattern.
• Potential claim overlap: The spawn step of claim 1 (and its dependent claims 2–3, which concern activity along the spawned thread). No anticipation — 5,758,184 has no object state model, no write-lock-by-self, no transition state, and no special-permission write-lock.

10. US 5,761,659 A — Sun Microsystems, Inc. (Bertoni et al.) — high relevance

• Filed/issued: 1996-02-29 / 1998-06-02
• Statutory basis: § 102(e)
• Description: "Method, product, and structure for flexible range locking of read and write requests using shared and exclusive locks, flags, sub-locks, and counters." Range locking via sub-locks; each sub-lock has a mode (shared/exclusive), a reader counter (number of processes holding a read lock), a waiting flag, and a queue of waiting processes; a shared region may have many readers while an exclusive region has one writer; requests either block/wait or fail per a block flag.
• § 102 relevance: Strong element-level overlap with (a) read vs. write lock modes, (b) multiple/nested lock acquisition on the same resource (sub-locks, counters) — i.e., the nested-locking concept — and (c) the "wait until the conflicting lock is removed" behavior underpinning claim 1's blocking of unrelated threads and claim 3.
• Potential claim overlap: Claims 1, 2, and 5 at the element level (multiply-held/counted locks; blocking waiting requesters). No anticipation — 5,761,659 is a lock-data-structure invention, with no object state machine, no transition state, and no special-permission grant.

11. US 5,771,388 A — National Instruments Corp.

• Filed/issued: 1994-05-04 / 1998-06-23
• Statutory basis: § 102(e)
• Description: "System and method for mapping driver level event function calls from a process-based driver level program to a session-based instrumentation control driver level system." Event-function mapping between process-based and session-based driver models.
• § 102 relevance: Background on event-based modeling (the patent's "Description of Related Art" contrasts event-based and state-based models). No locking or transition states.
• Potential claim overlap: none. No anticipation.

12. US 5,681,619 A — Mfactory, Inc.

• Filed/issued: 1995-04-03 / 1997-10-21
• Statutory basis: § 102(a)/(b) (issued 1997-10-21, days before the 1997-11-07 filing)
• Description: "Hierarchical encapsulation of instantiated objects in a multimedia authoring system." Hierarchical object containment/instantiation.
• § 102 relevance: Background on hierarchical object models; no locking, states, or threads.
• Potential claim overlap: none. No anticipation.

---

B. Additional reference cited in the family (not on the U.S. face) — the single most relevant reference

13. US 5,317,737 A — NCR Corp. (Richard R. Barton) — "Method and apparatus for controlling a re-entrant synchronization lock tenure in a multiprocessor system"

• Filed/issued: 1991-07-29 / 1994-05-31
• Statutory basis: § 102(a)/(b)
• Why it matters: This reference is cited on the EP 0 917 051 A3 search report (the EP family member of 5,931,919) with an "A" category against claims 1–5, even though it does not appear on the U.S. face of 5,931,919.
• Disclosure (verified from the patent text): A method for changing the tenure of the owner of a re-entrant synchronization lock. If the requesting process is the same as the current owner, the tenure level is incremented by one and the method exits (no deadlock, no new lock). If a non-owner requests, the method acquires exclusive access to state data and either assigns ownership with tenure = 1, or performs a contention process. Release decrements the tenure and only truly releases ownership when tenure reaches zero. The stated objects are to prevent deadlock and avoid the inefficiency of releasing/re-acquiring a lock held by the same owner.
• § 102 relevance: This maps almost element-for-element onto 5,931,919's nested/re-entrant locking disclosure (the specification's "unlimited number of nested locks," "unlock will only truly unlock … if the unlock requester is at the highest level of locking," and the FIG. 4 false-lock/false-unlock scheme), and onto claim 2 ("actions performed along the spawned first thread may request additional locks on the first object").
• Potential claim overlap: Claims 1 (re-entrant lock on the same object by the same owner), 2, and 5. This is the strongest single § 102 candidate for the nested-locking subject matter — but note the claims as issued are drafted around attribute-based/transition-state locking, so 5,313,737 anticipates the nested-locking disclosure, not the claim language of claim 1 or 5 as a whole.

---

C. Non-patent literature cited (and the EP search-report additions)

These are cited on the U.S. face and/or the EP 0 917 051 A3 search report and are directly on the nested-locking / nested-transaction axis:

---

D. Bottom line — ranked relevance and how each maps to claims

Most relevant (recommended primary § 103 combination art):

1. US 5,317,737 A (Barton/NCR) — re-entrant lock tenure counter → nested/false-lock scheme (claims 1, 2, 5 at the element level).
2. US 5,454,108 A (IBM) — state-full lock server with a lock-state mode field and an "in transition" state, conditional grant (claims 1, 5).
3. US 5,727,203 A (Sun) — transient (runtime-only) vs. persistent data → the "temporary/transition state unknown to the object database" concept (claim 5).
4. US 5,761,659 A (Sun/Bertoni) — shared/exclusive sub-locks with counters and waiting queues → multiply-held/nested locks and blocking waiters (claims 1, 2, 3, 5).
5. US 5,758,184 A (Microsoft) — spawning a separate worker thread to complete an async operation → the spawn element of claim 1.

Secondary/background: US 5,161,227 (lock hierarchy), US 5,226,143 (selective holder notification), US 5,008,853 / US 5,681,619 (shared/hierarchical objects), US 5,771,388 (event-based modeling), US 5,452,447, US 5,481,721, US 5,729,710.

On strict § 102: No single cited reference anticipates any of independent claims 1 or 5, because none discloses the combination of (i) an object's runtime-only transition state, (ii) the object unlocking itself in that state, and (iii) a spawned thread being granted a special-permission write-lock while the object remains in that state, with non-spawned threads blocked until the non-transition state. The cited art is best used to show that the individual building blocks (re-entrant/nested locks, transient state, state-full conditional lock grants, async worker threads, multi-attribute locks) were each known — i.e., a § 103 case against the loose wording of claim 5 and against the dependent claims 2–4 — rather than a § 102 anticipation of claim 1.

Two housekeeping flags carried forward (not new contradictions):

• Claim 3 recites "the second state" although claim 1 recites only a "non-transition state" — an antecedent-basis concern relevant to any validity/relevance ranking.
• US 5,317,737 A is a family-cited reference (it appears on EP 0 917 051 A3, not the U.S. face). I have kept it in a separate section rather than merging it into the twelve U.S. examiner citations, per the instruction not to conflate identifiers.

Confidence note: I verified the 12 U.S. citations and the Barton reference against Google Patents, Justia, Unified Patents, and USPTO.report. I could not independently retrieve the full text of every cited reference in this session, so the descriptions above are drawn from the verified abstracts/titles and, where noted, quoted specification passages; the § 102 claim mappings are my technical assessment, not a legal conclusion (and are moot for infringement purposes given the 2017-11-07 expiration).

I'll ground the obviousness analysis by pulling details on the key prior-art references cited on the patent page before mapping them to the claims.

Good grounding on the lock-manager references. Let me pull the remaining cited references — especially the ones that would supply the "temporary/transient state" and "state-dependent lock permission" elements.

Obviousness Analysis — US 5,931,919 under 35 U.S.C. § 103

Scope note: This analysis builds on the prior verification summary. It does not repeat the claim-by-claim plain-language overview already given, but it does resolve one point left open there and adds a threshold issue that materially changes how the reference list must be used.

---

1. Threshold issue the earlier section did not address: which cited references are actually available as prior art

US 5,931,919 has a filing/priority date of 1997-11-07, so pre-AIA § 102/§ 103 govern (the AIPA amendments do not apply). Several of the most useful references in the examiner's citation list issued after the applicant's filing date and therefore cannot be § 102(a) or § 102(b) art — but they remain available as pre-AIA § 102(e) art because their US filing dates precede 1997-11-07. This is likely exactly why the examiner cited them.

Two consequences. (a) An invalidity contention cannot simply assert the 1998-issuing references as "prior art patents"; it must plead them as § 102(e) art and rely on their filing dates. (b) Pre-AIA § 103(c) disqualifies § 102(e)/(f)/(g) art from use in a § 103 combination where the art and the application were commonly owned at the time the invention was made. This does not help Xerox here for the examiner-cited list, because none of the § 102(e) references are Xerox's. It does matter for the sibling Xerox filings (US 6,546,412, US 6,971,100, US 2002/0107998), which share the 1997-11-07 priority date: they are not "by another," and in any event § 103(c) would bar their use. They must not be used as § 103 art against 5,931,919.

Separately, note US 5,008,853 is Xerox's own 1991 patent. Because it is § 102(b) art (not § 102(e)), § 103(c) does not reach it — it remains fully available in a § 103 combination.

NPL discrepancy to flag (literal): the patent page lists Bodorik at pp. 309–318, while the IDEAS '97 program listing reports pp. 45–52. I am not auto-correcting either; the actual pagination should be verified from the proceedings before the reference is cited.

---

2. Claim 1 limitation-by-limitation mapping

Result: every element of claim 1 is met by the combination US 5,758,184 + US 5,161,227 + US 5,761,659. No single reference meets all elements, so the case must be a § 103 combination case, not § 102.

---

3. The principal combinations and the motivation to combine

Combination I (claims 1, 3)

5,758,184 + 5,161,227 + 5,761,659

Motivation. All three are in the same field of endeavor (pre-AIA § 103 / KSR "same field") and address the same problem: permitting concurrent access to a shared resource without blocking the requesting context and without deadlock.

• 5,758,184 supplies the reason to spawn the thread the claim recites: it is addressed to the inefficiency of the calling application stalling while a long-running (I/O) operation completes, and it teaches doing that work on a different thread while the queue object "manages multiple threads and their concurrency." The patent's own stated motivation for spawning a thread — so that a long task can be completed out-of-line and the object's state changed when the work finishes — is the same problem 5,758,184 solves.
• 5,161,227 supplies the state-conditioned grant/deny mechanism: the manager "checks this table for compatibility between the request and the current state of the lock entity," grants if compatible, and otherwise queues the request and suspends the task. That is, word for word, the mechanism of claim 1 step (i) plus the "special permission" step (f).
• 5,761,659 supplies the per-portion mode and wait mechanism that lets some requesters be served while others are made to wait: shared vs. exclusive sub-lock modes, a block flag that forces a request into a wait state, a "wanted" indicator, and a queue of waiting processes.

A POSITA presented with the problem of keeping a long-running object readable while still reserving write access for the thread that is doing the work would combine these three known lock-manager primitives. The combination is a predictable use of known elements according to known methods (KSR), requires no new hardware, and no unforeseen interaction arises — all are software lock-manager behaviors in a single runtime.

Combination II (claims 1–2)

Combination I + US 5,317,737. Claim 2 requires that "actions performed along the spawned first thread may request additional locks on the first object in the transition state." US 5,317,737 is directed precisely at this: a re-entrant synchronization lock whose tenure level is incremented when the same owner re-requests the lock. A POSITA combining a spawned worker thread with a state-conditioned lock manager would use a re-entrant lock so that the worker's internal helper calls can re-acquire the object without deadlock — the reference's own stated purpose. The § 102(b)/issue date (1994) also removes any § 103(c) common-ownership complication.

Combination III (claim 4)

Combination II + Bodorik & Jutla (MVM) and/or Scheerer et al. Claim 4 generalizes the lock condition to "any combination or logical expression of attributes." Bodorik & Jutla expressly describe enforcement of access-control protocols through FSM specification on units of data, with "lock unit state information" determining access — i.e., the lock decision is a function of the unit's state, defined per unit, and can differ from region to region once held as an arbitrary specification. Scheerer et al. is directed to locks parameterized along multiple dimensions with online allocation, which is a multi-attribute lock test. Generalizing a state-attribute test to a multi-attribute/boolean test is a simple substitution of one known element for another yielding predictable results, and the specification itself concedes the generalization ("This concept may be expanded beyond the state attribute"). This is the strongest sub-case in the patent.

Combination IV (claim 5) — the strongest obviousness case

5,161,227 + Bodorik & Jutla + 5,008,853 (+ 5,454,108). Claim 5 requires only: a first state → a transition state to perform predetermined operations; the transition state locks the object from threads not related to those operations; transition to a second state and unlock on completion. That is the ordinary exclusive-access-during-a-critical-section pattern (5,161,227's exclusive mode; 5,761,659's exclusive sub-lock) applied to an object that performs predetermined operations in a named state (Bodorik's FSM-specified lock units; 5,008,853's shared structured data objects with multi-user concurrency management). Claim 5 does not recite thread spawning, self-unlocking, or special permission at all — so the reference set needed to defeat it is materially smaller and more clearly analogous. I would expect claim 5 to be the most vulnerable claim in the set and claim 1 to be the most defensible.

---

4. Where the obviousness case is weak, and what the applicant would argue

1. The "unlock-then-write-lock" inversion. Claim 1 requires the object to release its write lock while in the transition state and yet still deny other threads' write locks. That is not how a conventional lock manager enforces exclusivity — normally you keep the write lock. The applicant will argue that neither 5,161,227 (which holds/queues) nor 5,761,659 (which grants by mode/range) teaches granting exclusive write access on the basis of the requester's identity as a designated thread while the object's own state variable excludes everyone else. The counter is Bodorik's per-unit FSM-specified access protocols (which do key permission to unit state) plus 5,161,227's lock-owner field, from which requester-identity-based permission is a predictable design choice — but this is the limitation to attack most carefully, and it is where a petitioner would want a reference that expressly ties lock permission to a thread identity attribute. I did not find such an express teaching in the references I verified.
2. "Special permission" is a result-oriented recitation. It is not defined with any structure in the claim; a POSITA would read it as merely "the lock manager grants this requester a lock it would deny another requester," which is exactly what 5,161,227's mode/state compatibility test and 5,761,659's mode field do. This cuts in favor of obviousness.
3. Method claims vs. system/apparatus disclosures. Most of these references are drafted as systems/apparatus. A reference describing a system programmed or adapted to carry out steps can teach a claimed method, but the applicant may raise the method/apparatus distinction. (I am flagging this as an argument point; the supporting case law should be verified rather than quoted from memory.)
4. Examiner already considered all these references and allowed the case. Because the applicant overcame these exact references during prosecution, a § 103 attack built from the same references bears a heavy burden under the § 282 presumption of validity and must explain a specific error in the examiner's reasoning (e.g., that the examiner treated 5,758,184 as limited to I/O completion ports rather than as a general teaching of spawning a distinct worker thread).
5. Non-analogous-art challenge is weak. All references sit in G06F 9/44–9/52 or 17/30 — the same field of endeavor (concurrent access to shared computing resources). The only arguably distant reference is US 5,771,388 (instrumentation driver events), and nothing in the claim-1 combination depends on it.
6. § 112 interaction worth re-flagging (already noted in the prior section): claim 3 recites "the second state" although claim 1 recites only "a non-transition state." For § 103 purposes, claim 3 must be construed as adding nothing beyond claim 1's final limitation. If instead it is held indefinite, the appropriate vehicle is § 112(b), not § 103.

---

5. Secondary considerations (Graham factor 4)

The materials contain no evidence of nexus, unexpected results, commercial success, long-felt-but-unsolved need, failure of others, copying, or industry praise — and no licensing activity. The patent ran its full 20-year term and expired 2017-11-07; fee payments at years 4, 8, and 12 are recorded. I found no litigation or PTAB proceeding. Absent such evidence, the secondary-considerations factor does not rebut the combinations above.

---

6. Explicit uncertainty — references I could not fully verify in this session

I retrieved full or near-full text and confirmed the specific teachings cited above for: US 5,317,737, US 5,161,227, US 5,761,659, US 5,758,184 (plus its continuation US 6,223,207), and the Bodorik & Jutla MVM abstract/record. My tool budget was exhausted before I could retrieve US 5,454,108, US 5,008,853, US 5,680,619, US 5,727,203, US 5,771,388, US 5,729,710, US 5,452,447, US 5,226,143, US 5,481,721, Scheerer et al., and Threads Primer pp. 140–143. My characterizations of those are title/abstract-level and are offered as candidate support, not as verified disclosures. Any of them must be pulled and read in full before being asserted in an invalidity contention or petition — in particular Threads Primer pp. 140–143, whose specific content I cannot vouch for, and US 5,008,853 and US 5,727,203, which are the references I would most want to check for an express transient/temporary-state teaching.

Net assessment: claims 2, 4, and especially 5 are, on this record, clearly obvious over the cited art with the motivations above. Claim 1 is obvious over 5,758,184 + 5,161,227 + 5,761,659, but its novelty-defining feature — releasing the write lock in the transition state while still denying write locks to all but a designated thread — is the limitation on which the defense would focus, and the strongest record against it would come from a reference expressly keying lock permission to requester identity or to an arbitrary object attribute.