US 5491798

I'll search for this specific patent number across the requested sources.

Let me run a few more targeted checks to confirm litigation/PTAB activity status.

US 5,491,798 — Verified Bibliographic Record

Sources consulted: Google Patents (https://patents.google.com/patent/US5491798/en), Justia Patents (https://patents.justia.com/patent/[5491798](/patent/5491798)), plus the full specification/claims text supplied in the task (treated as authoritative). Searches against Federal Circuit 2026 docket material and PTAB/USPTO document collections returned no case, IPR, or appeal bearing this patent number.

1. Bibliographic data

Cited prior art (examiner): US 4,544,804 (AT&T Bell Laboratories, "Conference control transfer arrangement"); US 4,555,781 (Reuters Ltd., "Conversational video system having local network control"); US 4,754,478 (Siemens AG, "Method of adding additional subscribers and signals to existing connection"); US 4,811,334 (U.S. Philips Corp., "Method of operation of a nodal telephone and data communication network…").
Non-patent literature cited: Aldred et al., "An Application Programming Interface for Collaborative Working," Fourth IEE Conference on Telecommunications, London, Apr. 18–21, 1993, pp. 146–151; Conti et al., "Requirements for a Multiparty Data Communication Protocol," 1992 Int'l Zurich Seminar on Digital Communications, pp. E6/1–8.
Notable later citing documents: US 5,973,724 (Apple, "Merging multiple teleconferences"); US 6,304,652 B1 (Nokia, "Merging of calls"); US 5,805,846 (IBM, dynamic application sharing among conference devices).

2. Abstract (as issued)

"A network comprises a set of nodes, each comprising a computer workstation in a computer network and including means responsive to messages from other computers in the network to join or leave a call. Each node maintains a list of calls in which that node is participating, an entry in the list for a call including the identities of all the other nodes in that call. In order to split a single call one node sends messages to a subset of nodes to join a second call, and leave the original call. The remaining parties in the original call then continue in a call of reduced scope. In order to merge two calls together, a node that is common to both calls sends messages to all the nodes in both calls, instructing them to leave the existing call and join a new call."

3. Technical context (from the specification)

The invention sits above a "Communications Control Layer" (CCL) and a per-node Call Manager application. Calls are built from two primitives, share_app(X, A, Y, B, N, U) and unshare_app(X, A, N, U), where N names a "sharing set" and U is an opaque user-information string. Two design constraints drive the invention:

• Call IDs are immutable — "there is no facility for changing the name of a sharing set." Names are made unique by concatenating a globally unique node name with an incrementing number.
• Unshare is strictly local — "it is not possible to issue a remote 'unshare' on behalf of another node," so the only way to ask a peer to leave a call is to embed an instruction (e.g. "MERGE Y02 Z03" or "SPLIT X02") in the U parameter of a share message. This is explicitly recited in dependent claims 2, 5, 6, 8 and 17.

Merging is therefore implemented as creating a third call and migrating both existing groups into it, rather than inviting one call's members into the other — the specification states this asymmetry "may result in some inequality between old and new members of the call (this may surface, for example, in resource allocation in a token passing environment)."

4. Independent claims — plain-language overview

Claim 1 — Splitting a call. Splits one existing call across an initial set of nodes into a first and a second call across two disjoint subsets.
Plain language:

1. Each participating node is a networked computer workstation that can join/leave a call on instruction from another node, and each keeps a call list recording the identities of all other nodes in each call.
2. One node in the original call draws up the list of which nodes go into the first (breakaway) call; the second call is implicitly whatever remains — the nodes of the original call not in that first subset.
3. That node then sends each member of the breakaway subset a message telling it to leave the original call and join the first call (two actions in one instruction).
4. Each recipient carries out those actions.

Literal-reading caveat: the issued claim text reads "…between nodes in the initial set for but in said first subset." This appears to be an OCR/printing artifact for "but not in"; I have not auto-corrected it, and the description consistently says the remnant is "nodes in the initial set but not in said subset."

Claim 9 — Merging two calls. Merges first and second calls across two initial node sets into a single call across their superset.
Plain language:

1. Same node/workstation architecture and same per-node call list with peer identities (preamble).
2. One node sends every member of the first call a message instructing it to leave that call and join the single (new, merged) call.
3. The same node sends every member of the second call an analogous message to leave the second call and join the same single call.
4. Each recipient then leaves whichever original call it was in and joins the merged call.

5. Dependent claims — what they add

Depend on claim 1 (splitting):

• 2 — Nodes cannot be remotely compelled to leave a call, but will act on another node's request to do so.
• 3 — Calls are named; a node may not be in two calls with the same name; the initiating node generates a name for the new first call and puts it in the messages.
• 4 — The name of the original call is also included in those messages.
• 5 — Combination of 3 + 2.
• 6 — Combination of 4 + 2.
• 7 — Each node of the subset sends a confirmation back to the initiating node, which forwards it to the other nodes that have joined so far (so every node's call list accurately reflects current participants).
• 8 — Combination of 7 + 2.

Depend on claim 9 (merging):

• 10 — Named calls; the initiating node generates a name for the merged call and includes it in both sets of messages.
• 11 — Preliminary step: pick the initiating node, and if it is not already in both calls, add it to whichever call it is missing from first.
• 12 — Recites the identical "select and add the node" limitation as claim 11 (apparent duplication/error in the printed claim set; noted literally, not corrected).
• 13 — The first call's name is included in the messages to the first set, and the second call's name in the messages to the second set.
• 14 — Combination of 10 + 11 (name plus pre-merge common-node selection).
• 15 — Combination of 14 with the "select and add" limitation restated a second time (again apparently duplicative).
• 16 — Combination of 13 + 11.
• 17 — Applies to claims 9, 10, 13 or 16, adding the no-remote-compulsion / responds-to-request limitation.

6. Litigation and post-grant activity — and stated uncertainty

• CAFC 2026 dockets: I found no Federal Circuit case for US 5,491,798. My searches for "5491798 CAFC 2026," "patent 5491798 appeal Federal Circuit 2026," and "5,491,798" … litigations/IPR returned only unrelated documents. Note that the numeric string "5491798" also appears as an unrelated Brazilian municipal contract number and a U.S. DOE/OSTI report identifier; those are different artifacts and were not used here.
• PTAB / USPTO: No IPR, PGR, or reexamination record for this patent appeared in the retrieved material. Google Patents lists USPTO PatentCenter, USPTO Assignment, and Global Dossier links, but I did not retrieve the individual PatentCenter record itself, so I cannot independently verify current fee/status fields beyond what the sources above show.
• Confidence: The bibliographic data (title, inventors, assignee, dates, claims, status) is high confidence — corroborated by two independent full-text sources and the supplied record. The absence of 2026 CAFC/PTAB activity is moderate-to-high confidence, based on the searches run and reinforced by the fact that the patent lapsed in 2004 and its statutory term would have expired 2013-11-10, making a live 2026 appeal highly unlikely. I have no authoritative information indicating any pending 2026 proceeding.
• OCR/printing artifacts in the record — "laving" (for "having"), "CC1" (for "CCL"), "particants" (for "participants"), "tile Call Manager" (for "the"), and the claim 1 phrase "for but in" — are reproduced as they appear; I flag them rather than silently correcting them, per the instruction to interpret identifiers literally.

I'll search multiple litigation sources for US patent 5491798. Let me run several queries.

Initial searches returned nothing relevant. Let me try more targeted litigation databases.

Let me check the patent litigation-specific databases directly.

Litigation Search Results — US Patent 5,491,798

Bottom line: No litigation involving US Patent 5,491,798 is known. I found no record of any lawsuit, ITC action, PTAB proceeding, or CAFC appeal in which this patent was asserted.

What I checked

I searched for the specific number in the forms 5491798, 5,491,798, and US5491798, combined with terms like "litigation," "lawsuit," "infringement," "case number," "plaintiff," "defendant," "CourtListener," and "Docket Navigator," and also queried sources touching the Unified Patents litigation portal. Results were dominated by false positives from unrelated numeric matches — none of them the patent:

• A Florida LLC's FEI number 82-5491798 (RMC Property Investments)
• A Lithuanian business-registry notice number 5491798
• A Brazilian contract alteration ID 5491798
• A Chilean municipal permit number 5491798
• A New York LLC entity ID 5491798 (Justin S. Kerr Consulting)

None of these relate to the patent. I found no docket, complaint, or opinion citing US 5,491,798.

Supporting evidence that no litigation exists

Several structural facts are consistent with the absence of litigation:

1. The patent lapsed early for non-payment of maintenance fees. Per the Google Patents legal-events record, the 4-year fee was paid in 1999, but a maintenance-fee reminder was mailed 2003-09-03, lapse for failure to pay was recorded 2004-02-13, and the patent lapsed effective 2004-02-13 (noted again 2018-01-26 as "PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362"). A patent that went abandoned in 2004 was never a meaningful assertion target during the modern NPE era.
2. The patent is IBM-owned and was never transferred to a monetization entity. The current/original assignee is International Business Machines Corp.
3. The Google Patents record lists no litigation. Its forward-citation list ("Cited By") shows only later patents referencing it (e.g., Apple's US 5,973,724 "Merging multiple teleconferences," Nokia's US 6,304,652 "Merging of calls," Cisco, Verizon, Xerox/Capital One), not infringement suits.

Important caveats about this answer

• Free-web-source limitations. The Unified Patents litigation portal and PACER/Docket Navigator are not fully exposed to general web search; a paywalled/commercial litigation database (e.g., Docket Navigator, Lex Machina, RPX Insight) or a direct PACER docket query is the definitive check. My search did not return a queryable caselist entry for this patent number.
• No litigation ≠ never licensed or never in a portfolio. IBM historically licensed large portfolios; such activity would not appear as "litigation."
• I could not confirm a negative from a purpose-built litigation database within this search, so my statement is: no known/recorded litigation was found, rather than an absolute certainty that none has ever existed.

If you need a definitive confirmation

I'd recommend querying, in order:

1. Unified Patents Portal — https://portal.unifiedpatents.com/litigation/caselist (filter by patent number)
2. CourtListener / RECAP — https://www.courtlistener.com (search "5,491,798" in the RECAP docket archive)
3. Docket Navigator / Lex Machina — paywalled, but the authoritative source for district court, ITC, and PTAB assertions
4. PTAB — https://ptacts.uspto.gov for any IPR/PGR/CBM (none appeared)

Filing-date and status note: If any case did exist, its filing would fall in the window before the patent's 2004 lapse, or in the 1996–2004 enforcement window. I found none.

Summary table

If you can share a specific docket or case caption you've seen referenced, I can search for that exact case to confirm or refute it.

Proceedings overview

Total AIA trial proceedings on US 5,491,798: 0 (0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denials). The structured USPTO ODP block supplied with this task — the canonical proceedings list — returns no AIA trial proceedings for this patent, and independent web checks for an IPR/PGR/CBM bearing this number returned nothing on point.

Bottom line for a defendant: there is no PTAB record to work from, but that absence is not a signal of a "hardened" patent — it is a consequence of chronology. US 5,491,798 issued 1996-02-13 and lapsed for non-payment of maintenance fees effective 2004-02-13, with its statutory term ending 2013-11-10. Both windows in which a challenge would have been economically rational (and, for PGR/CBM, legally available) closed long before the AIA trial regime mattered. A demand letter citing this patent today faces a far more basic problem than validity: essentially no live damages window (see Strategic summary).

No proceeding sections are rendered below because the canonical list is empty. I decline to invent proceeding numbers. What follows is the verification trail and the strategic read.

Verification performed (no proceedings found)

Confidence: high that no IPR/PGR/CBM was ever filed against this patent, corroborated by (a) the empty ODP list, (b) the patent's status history showing lapse by 2004, and (c) the complete absence of any docketed PTAB trial in any search. Two caveats stated plainly: CBM review was only available for patents issued from the pre-AIA regime and only to parties charged with infringement, and my web searches are not exhaustive across pre-2012 paper records; but for a patent that lapsed in 2004, any such record would be a curiosity rather than a defensive asset. If you need a belt-and-braces confirmation, the primary sources are PTAB E2E / PTAB Center (https://ptacts.uspto.gov/ptacts/) and the USPTO Patent Center record for US 08/150,675.

Availability of each forum today (for completeness — none is practically useful here):

• IPR — no statutory expiry bar; in principle still filable against an expired patent, but there is nothing to win (see below).
• PGR — unavailable. PGR applies only to patents with an effective filing date on or after 2013-03-16; this patent's priority is 1993-09-25.
• CBM — unavailable. The transitional program was limited to pre-AIA patents meeting the "covered business method" test and to petitioners charged with infringement; the program sunset 2020-09-16 (see the U.S. Solicitor General's brief discussing § 18(a)(3)(A), https://www.justice.gov/osg/media/[967881](/patent/967881)/dl?inline).
• Ex parte reexamination — also absent from the record (the only reexamination-style activity in this family is the withdrawn EP 0650278 application; no EPO opposition could arise from a withdrawn application).

Strategic summary

Claim status: every claim 1–17 is UNTESTED at the PTAB, and none has ever been canceled or confirmed in an AIA proceeding. There is no FWD to link to, no claim-level disposition to quote, and therefore no IPR-narrowed claim set. The only claim-level guidance available is the intrinsic record summarized in the earlier sections — independent claim 1 (splitting) and independent claim 9 (merging), with dependents 2–8 and 10–17 — plus the drafting quirks already flagged (the "for but in" artifact in claim 1; the duplicated "select and add the common node" limitations at claims 11/12 and 14/15), which are more useful to a § 112 or claim-construction attack in district court than to an IPR.

Estoppel landscape: § 315(e)(2) estoppel is a non-issue, and so is the benefit it usually buys. Because no IPR or PGR was instituted, no petitioner (or privy) is estopped from anything, and there is no patent-owner-favorable FWD giving a defendant a roadmap. Conversely, a defendant today gets no free "the Board already killed claim 1" argument. Practically, any IPR you might consider would be limited to § 102/103 on patents and printed publications — and the four examiner-cited references plus the two NPL items already in the file would be the obvious starting art:

• US 4,544,804 (AT&T Bell, "Conference control transfer arrangement")
• US 4,555,781 (Reuters, "Conversational video system having local network control")
• US 4,754,478 (Siemens, "Method of adding additional subscribers and signals to existing connection")
• US 4,811,334 (U.S. Philips, nodal telephone/data network)
• Aldred et al., An Application Programming Interface for Collaborative Working, 4th IEE Conf. on Telecommunications (1993-04-18/21), pp. 146–151
• Conti et al., Requirements for a Multiparty Data Communication Protocol, 1992 Int'l Zurich Seminar on Digital Communications, pp. E6/1–8

Note the § 315(b) one-year bar and § 325(d) discretion would be live concerns if this patent were asserted today, and the Aldred reference is the authors' own prior paper — a potentially useful § 102(b)/§ 103 anchor given the 1993-09-25 GB priority and 1993-11-10 filing dates.

Pattern signals: none — no repeat petitioner, no serial patent-owner appeal, no defensive aggregator in the chain. There is no Unified Patents (or any aggregator) filing against this patent, no joinder, and no CAFC appeal from a PTAB decision. The citing-document trail (e.g., Apple's US 5,973,724 "Merging multiple teleconferences"; Nokia's US 6,304,652 "Merging of calls") shows the technology attracted later patenting and, in Apple's case, litigation — but this patent was never the vehicle.

Current-law context worth knowing (not specific to this patent). If a demand letter did cite this patent, current USPTO practice makes institution materially harder for old patents. As of 2025-10-16, Director Squires centralized institution decisions in the Director rather than the merits panel (https://www.willkie.com/publications/2025/10/ptab-announces-new-institution-policy), and proposed rules published 2025-10-15 would bar IPR against a patent that had already survived a validity challenge and require petitioners to stipulate away § 102/103 arguments in other venues. A related "settled expectations"-style discretionary-denial practice has been reported in cert-stage briefing as having been invoked to deny many IPR petitions against older patents (see the Unified amicus filing, https://www.supremecourt.gov/DocketPDF/25/25-1230/[412129](/patent/412129)/20260529145347123_Unified%20Amicus%20Brief%2025-1230.pdf) — I flag this as party advocacy in a pending cert petition, not as a decision I have verified, and the case name associated with No. 25-1230 is not something I can confirm from the retrieved material. None of this changes the analysis here, because the patent is expired and lapsed; it matters only if you are weighing an IPR against some other aging patent.

Recommended next steps

• No FWD exists to link or quote. The disposition statement requested for a defendant cannot be made: there is no Final Written Decision, no institution decision, and no cancellation of any of claims 1–17. Do not build a defense narrative around PTAB outcomes.
• If a demand letter arrives citing US 5,491,798, lead with the expiration and lapse record, not validity. The maintenance fee lapsed effective 2004-02-13, the term ended 2013-11-10, and royalty recovery is limited by 35 U.S.C. § 286 to six years before suit — which, from 2026, does not reach back to the 2013 expiration. The practical posture is that any assertable damages window is closed; ask the asserting party to identify the accused acts and their dates.
• If you nonetheless want a validity record, the realistic forum is district court or the ITC § 337 track, not the PTAB — and there, the strongest levers against claim 1's "for but in" verbiage, claim 9's preambular "call list" structure, and the duplicative claim 11/12 and 14/15 limitations are §§ 101/112 and claim construction, none of which IPR can reach (IPR is confined to § 102/103 on patents and printed publications; PGR, the only forum that reaches § 101/112, is unavailable for a 1993-priority patent).
• For a definitive "no proceedings" certification, pull the PTAB E2E record for the patent (https://ptacts.uspto.gov/ptacts/) and the Patent Center file for application US 08/150,675, and cite those primary URLs rather than this analysis.
• Do not represent to any tribunal or counterparty that this patent was IPR-tested. The correct statement is: no AIA trial proceeding has ever been filed or instituted against US 5,491,798; the patent lapsed in 2004 and expired in 2013.

I'll research the assignment record for US 5,491,798 across the assignment databases and related sources.

Let me search more specifically for any assignment records and the reel/frame data.

Let me try to pull the actual assignment record and check for any post-issuance transfers.

I was unable to load the Assignment Center's dynamic search results directly (it is a JavaScript application that my retrieval could not render), so the assignment record below is reconstructed from the USPTO legal-events data mirrored in the authoritative patent text supplied in the task (which carries the recorded reel/frame for the one assignment) plus the Google Patents legal-events entry, cross-checked against Justia and the EPO family record. Where a field (notably the correspondent) was not present in any retrieved source, I say so rather than guess.

---

Inventors

Four named inventors, all appearing to be IBM UK (Hursley Park, Winchester) employees at the time of filing — the specification's own working environment (a "Communications Control Layer" plus a per-node "Call Manager") and one of the cited non-patent references (Aldred et al., "An Application Programming Interface for Collaborative Working," Fourth IEE Conference on Telecommunications, London, Apr. 1993) are IBM UK/Hursley work product.

Assignment execution dates: 1993-10-19 to 1993-10-21 (per the recorded assignment "SIGNING DATES FROM 19931019 TO 19931021").

Unusual patterns: None detectable from the available record. There is no evidence in any retrieved source of any inventor departing IBM within 12 months of filing, and — more importantly — none of the usual consequences of such a departure (no subsequent inventor-originated assignments, no startup chain, no portfolio spin-out). I cannot affirmatively confirm inventor employment continuity beyond the filing, so I flag this as not determinable rather than "clean." No finding either way.

---

Original assignee

International Business Machines Corporation (Armonk, New York) — named as assignee on the issued patent (Google Patents lists IBM as both original and current assignee).

• Primary line of business: Mainframe/enterprise computing, software, and services. In 1993 IBM UK's Hursley laboratory was the center of its collaborative-working and network software research, which matches the subject matter of this patent.
• Product embodying the claims: Unclear. The specification describes a working software architecture (CCL + Call Manager) rather than a discrete product. A 1994 Aldred et al. paper titled "Call Management in a Lakes environment" appears in the Google Patents "Similar Documents" list, suggesting the invention was reduced to practice inside an IBM internal/product environment, but I found no evidence of a commercially shipped product specifically practicing the split/merge claims. I will not assert one.
• Current status: Operating. IBM remains a going concern. It did not sell or spin out this patent.

---

Assignment timeline

There is exactly one recorded assignment in the chain, plus the terminal legal event (lapse).

• 1993-10-19 → 1993-10-21 (executed) / recorded 1993-11-10 — Reel 006785 / 0103

  • Conveyance: Assignment — "ASSIGNMENT OF INTEREST (SEE DOCUMENT FOR DETAILS)"; recorded in USPTO legal events as ASSIGNMENT … ASSIGNORS: BONSALL, G. W.; LAMBERT, H. S.; MITCHELL, H. D.; AND OTHERS; REEL/FRAME:006785/0103; SIGNING DATES FROM 19931019 TO 19931021
  • Assignor: Gordon W. Bonsall; Howard S. Lambert; H. David Mitchell; Barry K. Aldred
  • Assignee: IBM Corporation, New York
  • Correspondent: Not determinable. The recorded-assignment abstract surfaced in the sources does not include the filing attorney/agent. (The EP family names IBM UK's Simon Robert Davies, IBM UK Intellectual Property Dept., Hursley Park, as the EP prosecution representative — that is a European patent agent of record, not the US assignment correspondent, and I am deliberately not conflating the two.)
  • Context: Original inventor-to-employer assignment (standard employment/confirmation-of-assignment transfer executed at filing).

• 2003-09-03 — Maintenance-fee reminder mailed (4-year fee had been paid 1999-06-25).

• 2004-02-13 — Lapsed for failure to pay maintenance fees (effective date 2004-02-13); formal discontinuation recorded 2018-01-26 (37 CFR 1.362).

No post-issuance assignments exist. The patent was never transferred away from IBM — no license recording, no security agreement, no merger, no change of name, and no sale. It simply expired for nonpayment in 2004, roughly nine years before its anticipated statutory expiration (2013-11-10).

---

Timeline diagram

timeline
    title Ownership of US 5491798
    1993 : Inventors execute assignment
         : Recorded Reel 006785 Frame 0103
         : Assigned to IBM Corporation
    1996 : Patent issues Feb 13
    1999 : IBM pays 4 year maintenance fee
    2003 : Maintenance fee reminder mailed
    2004 : Patent lapses for nonpayment
    2018 : Formal discontinuation recorded

---

NPE / troll-pattern signals

Net: All eight signals are absent or unassessable. This is an unusually "quiet" record — the opposite of a troll pattern.

---

Verdict

Insufficient data — only the original assignment.

Justification: The complete recorded chain is a single inventor-to-employer assignment to International Business Machines Corporation, executed 1993-10-19/21 and recorded 1993-11-10 at Reel 006785 / Frame 0103; no subsequent assignment, license, security interest, merger, or change of name ever appears in the record, and Google Patents legal events show the patent simply lapsing for nonpayment of maintenance fees effective 2004-02-13 — nine years before its 2013-11-10 anticipated expiration. There is therefore no LLC chain, no repeat correspondent, no pre-litigation transfer, and no asserter of any kind to evaluate; under the rubric this is the "only the original assignment" case, not an NPE finding. (Because I could not render the Assignment Center's dynamic results page directly, the residual possibility of an unrecorded/abandoned transfer is negligible but not literally zero — however, an unrecorded transfer would in any event be ineffectual against third parties under 35 U.S.C. § 261.)

Verification links:

• USPTO Assignment Center: https://assignmentcenter.uspto.gov/
• Legacy Assignment Search (same records): https://assignment.uspto.gov/patent/index.html
• Google Patents legal events: https://patents.google.com/patent/US5491798/en
• Justia: https://patents.justia.com/patent/[5491798](/patent/5491798)

Caveat on method: The assignment-center host is a client-rendered application and did not return parseable results to my retrieval; the reel/frame citation above is taken from the recorded-assignment data embedded in the Google Patents legal-events entry, which reproduces the USPTO record. I did not fabricate any reel/frame, correspondent, or assignee. No EP0650278-family or GB priority assignment should be read into the US chain.

I'll research this patent on USPTO and dig into each cited prior-art reference.

Prior-Art Analysis — US 5,491,798 ("Method for network call management")

Scope and source caveat (read first)

I was asked to search the USPTO database for patent number 5491798 and to analyze each patent citation of record. Two honest limitations on what follows:

1. I did not retrieve the USPTO PatentCenter page itself. The searches I was able to run returned Google Patents, Espacenet/EPO, and uspto.report records for the patent and for each cited reference. The bibliographic record was already verified in the earlier section and I do not repeat it here. Everything below is grounded in the full patent text supplied in the task (authoritative) plus the retrieved records for each cited reference.
2. No similar numbers were substituted. Where the numeric string "5491798" collides with unrelated artifacts, those were ignored, as in the earlier litigation section.

Critically, I distinguish two different lists that are easy to conflate:

• References Cited on the face of US 5,491,798 (examiner-cited prior art) — 4 U.S. patents. This is the prior art.
• "Cited By" / forward citations (Apple US 5,973,724; Nokia US 6,304,652; Cisco; Verizon; Capital One; etc.) — these post-date the patent and are not prior art to it. The task's wording ("each patent citation for 5491798") could be read to include these, so I flag them explicitly and exclude them from the §102 analysis.
• A third list, "Family Cites Families" (US 5,195,086 and CA 2,094,410), are references cited in the prosecution of related family members, not on the face of the US patent. They are analyzed separately below because they are potentially the closest art.

§102 legal frame. The earliest priority is 1993-09-25 (GB 9319851). All four examiner-cited U.S. patents issued in 1985–1989 — more than one year before that date — so they are pre‑AIA §102(b) art (patented / printed publication more than one year before the applicant's date). US 5,195,086 issued 1993-03-16 on an application filed 1990-04-12 and is best characterized as §102(e) art (granted U.S. patent with an earlier effective filing date).

Threshold point: the independent claims issued over all four cited references. My conclusion is that none of them is a true §102 anticipation of claim 1 or claim 9; the value of the list is that it shows what the examiner treated as the closest art, and it delimits the inventive contribution (distributed per-node call lists + a single "leave-and-join" instruction used to split and merge multi-party calls).

---

1. The four examiner-cited references (patent citations of record)

(a) US 4,544,804 — "Conference control transfer arrangement"

Description (from abstract): "A dial-up audio/data conference arrangement … wherein the originator of the conference has certain speaker priorities and control capabilities e.g., adding and deleting conferees … The originator, under the guidance of voice prompts, can transfer this control to another party on the conference."

§102 relevance: This is a centralized conference-bridge arrangement. Its "adding and deleting conferees" is the functional ancestor of "join/leave a call," and its "transfer of control" is a form of call management. It is the most apt reference for the preamble-type concept of joining and leaving a multi-party call and for control transfer.

Why it does not anticipate:

• No computer-workstation nodes with each node maintaining a list of calls and the identities of the other nodes (claim 1/claim 9 preamble).
• No "leave the initial call and join the first call" single instruction (claim 1 step 3); deletion and addition are distinct bridge operations.
• No splitting of one call into two disjoint subsets; no merging of two multi-party calls into a third call.

Potential claim mapping (relevance, not anticipation): the broad "join or leave a call" language in claim 1's preamble / claim 9's preamble. Does not anticipate any claim.

---

(b) US 4,555,781 — "Conversational video system having local network control"

Description (from abstract): A "video conversational data communication network … subscribers … may conduct conversational video textual data communications with one or more keystations … Each keystation is associated with a keystation terminal controller interface … connected to a switching concentrator computer and a message switching node … The controller locally stores video conversational textual data for its associated keystations and enables two different designated keystations to conduct two different video conversations with a common keystation in a split screen display … Prior to completion of a call, the controller provides an incoming calls queue."

§102 relevance: This is the closest of the four on the architecture side. It discloses local (per-node) control, a terminal controller maintaining conversational state, and the ability of one station to be engaged with more than one conversation at once — which is conceptually adjacent to "each node maintaining a list of calls in which that node is participating."

Why it does not anticipate:

• Its "split screen" is a display split, not a call split into two disjoint subsets (claim 1). Do not confuse the two.
• "Two different video conversations" is a station holding parallel one-to-one conversations, not a mechanism for one multi-party call to be divided or for two multi-party calls to be merged.
• No instruction of the form "leave call A and join call B" (claim 1/claim 9 steps).

Potential claim mapping: peripheral support for the claim 1/claim 9 preamble's "list of calls in which that node is participating" and for the multi-conversation capability. Does not anticipate any claim.

---

(c) US 4,754,478 — "Method of adding additional subscribers and signals to existing connection"

Description (from abstract): On initiation of the "offering" service feature, "the existing telephone connection is reswitched onto two subscriber inputs of a conference or offering device … arranged in a peripheral unit … The two subscribers are connected via a preset conference switching matrix array. The additional telephone paths or audible signals are conducted to a further subscriber input of the conference or offering device." Realizes the subscriber features "offering" and "rapping" (i.e., adding a third path or audible signal onto an existing two-party connection).

§102 relevance: This is the classic "convert a two-party call into a three-party conference by adding a participant" reference — a PSTN/Siemens version of "extend a call," which the present patent itself discusses in its "Extending a call from a two-way call to a multi-party call" passage. It therefore bears directly on the general concept of a call changing its participant set.

Why it does not anticipate:

• It adds a party to an existing connection; it does not merge two existing multi-party calls into a new single call across a superset (claim 9), nor split one call into two disjoint calls (claim 1).
• Control is centralized in switching-matrix/conference hardware, not distributed to networked computer workstations each maintaining its own call list with peer identities.
• No "leave the old call and join the new call" dual-action message (claim 1/9).

Potential claim mapping: general relevance to "add a party / join a call"; possible §103 motivation with the distributed-computing art, but no §102 anticipation.

---

(d) US 4,811,334 — "Method of operation of a nodal telephone and data communication network to provide specific facilities in changeable domains of the network"

Description (from abstract/description): A nodal PABX communication network in which "local data is provided in each node for a specific facility-bound program which dictates the operation of the control module, the local data containing an adjustable address file of the addresses of those nodes via which a call for a specific facility … can be accepted." Nodes provided with the same facility are extended and their addresses incorporated in the file of each such node so as to define a domain in which the facility can be offered; domains can be overlapping or non-overlapping and can be set up or changed during operation.

§102 relevance: This is the strongest of the four on the distributed-node + per-node stored address/identity data axis. It shows that, before the 1993 priority date, it was known to give each node its own local data file of node addresses governing which other nodes a facility may involve — an architectural cousin of "each node maintaining a list of calls … including the identities of all the other nodes in that call."

Why it does not anticipate:

• The stored data is a facility-domain address file, not a per-call participant list keyed to call identity.
• It concerns routing/offering of facilities (operator assistance, common answering night service, "call me" paging), not splitting or merging multi-party calls.
• No leave-and-join instruction; no disjoint-subset splitting logic (claim 1); no superset merge (claim 9).

Potential claim mapping: preamble "each node … maintaining a list" (as a stored list of identities, though not call-specific). Does not anticipate any claim.

---

2. Family-cited references ("Family Cites Families") — potentially the closest art

These were cited in the prosecution of a related family member, not on the face of the U.S. patent, so I flag them as a separate category. I was able to confirm their bibliographic facts from the supplied record but not to retrieve full-text spec content within the available steps, so the mapping below is provisional and I say so explicitly.

(e) US 5,195,086 — "Multiple call control method in a multimedia conferencing system"

Why this is the most concerning reference: by its title it is a "multiple call control method" in a "multimedia conferencing system." That is squarely on the call-control axis of the ’798 patent. If it discloses establishing, controlling, and changing multiple calls among endpoints of a multimedia conference, it is the reference that most directly tests the "each node maintaining a list of calls in which that node is participating" and "multiple-call management" concepts.

Provisional assessment: I cannot state it anticipates claim 1 or claim 9. From the title alone one cannot establish the required elements — in particular the disjoint-subset split of claim 1 and the two-calls-into-a-third-superset merge of claim 9, nor the "leave A and join B" instruction. Because I could not pull its specification, I mark this unresolved, and it is the single reference I would retrieve first if a definitive §102 answer is required.

(f) CA 2,094,410 C — "Distributed management communications network"

Relevance: A "distributed management" communications network with a 1992 filing date falls within the year before the ’798 priority — its status as §102 art depends on publication timing. Conceptually it is on the distributed control axis. As with US 5,195,086, I could not retrieve full text in the available steps; treat as unresolved / background.

---

3. Non-patent literature cited (also prior art of record)

---

4. References that are NOT prior art (forward citations / "Cited By")

For completeness, and to prevent mis-citation, these post-date US 5,491,798 and are not §102 art against it (they are later documents citing it):

• US 5,973,724 (Apple, "Merging multiple teleconferences," priority 1995-02-24)
• US 6,304,652 B1 (Nokia, "Merging of calls," priority 1996-11-14)
• US 5,805,846 (IBM, dynamic application sharing among conference devices, 1994)
• US 5,572,582 / US 5,854,898 / US 5,999,977 (Apple) and the Apple reissue family (US RE40,704; RE42,442; RE44,306; RE44,341; RE44,395)
• Cisco US 7,602,704 / US 8,971,511; Verizon US 9,392,120; Capital One US 2023/0262197

These are useful only for showing what the ’798 disclosure subsequently influenced (notably Apple's and Nokia's own "merge calls/conferences" art). They carry no §102 weight against US 5,491,798.

---

5. Bottom line

Conclusion: The four patents cited on the face of US 5,491,798 are all pre‑1993, mostly centralized PSTN/PABX conference-control art. They establish the state of the art in adding, dropping, and controlling parties on a conference (US 4,544,804; US 4,754,478), local/terminal-controller control of multiple video conversations (US 4,555,781), and per-node stored identity/address data in a nodal network (US 4,811,334). None discloses the two limitations that carry the ’798 independent claims: (i) splitting one call into first and second disjoint node subsets by a single "leave the initial call and join the first call" message (claim 1), and (ii) merging two calls into a new single call across the superset, again via leave-and-join instructions (claim 9) — the latter being expressly motivated in the specification by the asymmetry problem in token-passing resource allocation.

The one reference I would not characterize as conclusively cleared is US 5,195,086 (family-cited AT&T "Multiple call control method in a multimedia conferencing system"): its title and 1990 filing date place it closest to the ’798 call-control subject matter, but I could not extract its specification within the available steps, so I deliberately refrain from declaring it non-anticipatory. Obtaining its full text (and CA 2,094,410 C's) is the necessary next step to close out the §102 question.

Caveats to carry forward: (1) I retrieved the record via Google Patents / Espacenet / uspto.report, not a direct USPTO PatentCenter/Patent Public Search record; (2) the US 4,555,781 dates differ slightly between the patent front page (filed Mar. 30, 1983; issued Nov. 26, 1985) and a secondary database (priority Mar. 29, 1983; grant Nov. 25, 1985) — reproduced as found, not auto-corrected; (3) the claim-text quirks and OCR artifacts noted in the earlier section (e.g., "for but in," "CC1," "laving") are carried forward unchanged per the no-auto-correction rule.

I'll ground my analysis in the actual prior-art references listed on the patent page. Let me pull details on each.

Let me pull the two cited non-patent references and the family-cited US 5,195,086, which bear directly on the merge/split reasoning.

Obviousness Analysis — US 5,491,798 under 35 U.S.C. § 103

Building on the bibliographic and claim material already established above; I do not repeat that content here.

---

1. Framework and legal standard

Because US 5,491,798 has an earliest priority date of 1993-09-25 (GB 9319851) and was filed 1993-11-10, it is governed by pre-AIA § 102/103. Analysis proceeds on the Graham v. John Deere, 383 U.S. 1 (1966), factors — scope/content of the prior art, differences from the claims, level of ordinary skill, and secondary considerations — as refined by KSR Int'l v. Teleflex, 550 U.S. 398 (2007) (analogous art; predictable combinations; "obvious to try"; design incentives; market demand).

Level of ordinary skill (my estimate): a bachelor's degree in EE/CS (or equivalent) plus ~2–3 years in computer networking and multi-party/conferencing systems, familiar with LAN/ISDN transport, CCITT/IEEE standards, and the packet/link-level "sharing-set" primitives that the patent itself attributes to its own earlier collaborative-working API. The patent's own background cites the MERMAID multiparty desktop conferencing system (CSCW '90) and the CCITT T.12x Multi-media Control Unit draft — corroborating an active, mature field by the 1993 priority date.

Important evidentiary caveat: I have retrieved only the abstracts and selected text of US 4,544,804; US 4,555,781; US 4,754,478; and US 4,811,334, and only a partial text of the Conti paper. A full-disclosure review of each reference would be required to convert any of the below into a defensible rejection. My conclusions are therefore framed as grounds, not as proven rejections.

---

2. What the prior art of record actually teaches

---

3. The distance the claims must bridge

From the claim overview already generated: the independent claims 1 and 9 are essentially method steps for re-grouping a set of networked workstations among calls — pick a controlling node, enumerate a target group, message each member to leave the old call and join the target call, and have each member act. The specification concedes that the two design choices at the heart of the claims are known alternatives with known trade-offs:

• For merging, the spec expressly states the alternative is to "invite all those nodes in … the second call to join the first call," rejected only because of an asserted "inequality between old and new members … (this may surface, for example, in resource allocation in a token passing environment)." A choice between two known merge strategies, justified by a routine resource-allocation concern, is a classic design choice / predictable variation.
• For splitting, the spec concedes that if some node fails to leave the old call, "the practical effects … are likely to be minimal"; the same point is made for merge ("the only disadvantage … is … relatively small … continued overheads"). That is a low-cost, expected-result characterization — the hallmark of obviousness.

The only arguably distinctive limitations are (a) nodes are peer workstations each holding their own call list of the other participants' identities, and (b) the leave+join instruction is delivered in a single message (in the spec, carried in the opaque "user information" field because remote unshare is impossible).

---

4. Claim 1 (splitting) — proposed combination

Primary: US 4,811,334 (arbitrary, run-time-reconfigurable node "domains") in view of US 4,544,804 (add/delete conferees; transfer control) and Conti et al. (add/quit primitives; notify all parties; unique IDs).
Secondary/platform: US 4,555,781 (distributed terminal-controller nodes that locally store call state and exchange join/end messages).

Motivation / rational underpinning:

1. Same field, same problem. All five touch multi-party calls among addressable network endpoints. A node that can be added to a call (4,754,478) or deleted from a conference (4,544,804) can likewise be moved between two calls — splitting is the mirror image of adding, applied to the social structure of the call rather than the media path.
2. Conti supplies the exact protocol primitives. "Any communicating party must be able to add a new party … without affecting the running conference" and "each CP must be able to quit … at any time" together teach the join and leave operations that claim 1 combines into one instruction. Conti's requirement that "every CP … be notified when a CP leaves" supplies the membership-tracking that claim 1's "list of calls … identities of all the other nodes" presupposes.
3. 4,811,334 supplies the partitioning. Teaching domains that may be "overlapping or not" and reconfigurable "during operation" gives the claimed "two disjoint subsets."
4. 4,555,781 supplies the platform. Nodes that are terminal controllers locally storing conversation state and messaging one another for call setup/teardown map onto the claim's "computer workstation … responsive to messages … to join or leave a call."

Reasonable expectation of success: High. Every constituent step (identify members; send a message; add/remove a member; keep a membership list) is individually disclosed or conventional, and the combination produces no new mechanism — it merely severs an existing call into two via the same add/quit machinery.

---

5. Claim 9 (merging) — proposed combination

Primary: US 4,754,478 (add a party to an existing connection) and US 4,544,804 (a controlling node that adds/deletes conferees) — in view of Conti (any party may add; unique ID per party; notify all on membership change) and US 4,555,781 (distributed nodes).

Motivation: Merging is the repeated application of "add a party to a call." The specification's own admitted alternative (invite all of call 2's nodes into call 1) is a known, simpler variant of the same concept. The only refinement — creating a third call and migrating both groups — is a predictable choice justified by the token-passing resource-allocation rationale, i.e., the kind of "known design incentive" KSR holds insufficient to confer patentability absent unexpected results.

---

6. Dependent claims

• Claims 3 & 10 (unique call name per call): Conti expressly requires unique addresses and an ID negotiation guaranteeing uniqueness "even if two new parties are added simultaneously." The patent's own expedient (node-name + incrementing counter) is a routine naming convention. Strongly obvious.
• Claims 4 & 13 (include the old call's name in the message): A natural necessity — a node cannot leave call X unless told which call X is; plainly an obvious implementation detail in view of Conti's addressing scheme.
• Claim 7 (each node confirms; the initiator forwards the confirmation to others so far joined): This is precisely Conti's "every CP should be notified when a CP leaves, and the notification … forwarded to the applications," and mirrors the share_confirm fan-out the patent itself describes. Very strong obviousness ground.
• Claims 2, 5, 6, 8, 17 ("a node cannot be remotely compelled to leave … but … responsive to requests"): A functional/design limitation. Conti's model — the party itself quits and others are notified — is functionally the claimed "request, not compulsion" asymmetρy. No new structure.
• Claims 11–16 (select a node common to both calls; add it if absent): This is the enabling precondition for 4,544,804-style "control by a participant" applied to two calls; selecting/nominating a participating node and ensuring its presence in both is an obvious administrative step. (Note: as flagged in the earlier claim overview, claim 12 is textually identical to 11, and 15 re-recites the same "select and add" limitation — a drafting defect that, if anything, narrows the value of these claims rather than saving them.)

---

7. Counter-arguments a patent owner could raise, and their strength

1. "The prior art is centralized; '798 is serverless/peer-to-peer." This is the patent's best argument (it criticizes the CCITT T.12x MCU's "relatively fixed, hierarchical structure"). Counter: Conti expressly starts from "all parties having equivalent privileges" with no centralized control, and US 4,555,781 distributes call control to per-node terminal controllers. The distributed architecture is itself in the art.
2. "Merge must create a new call because call IDs are immutable and unshare cannot be done remotely." Counter: the specification frames the alternative as viable; the immutability/remote-unshare constraints are properties of the applicant's own prior API (the Aldred paper), and tailoring a known method to a known API constraint is routine engineering.
3. "The 'user information' string is the only channel to instruct a peer to leave." Counter: implementation detail/signal carrier; obvious in view of any message-based add/quit protocol (Conti).

No secondary considerations appear in the record: the patent was IBM-owned, lapsed for non-payment effective 2004-02-13 (per the legal-events data already reported), and I found no evidence of commercial success, licensing-driven nexus, long-felt need, or failure of others attributable to these claims.

---

8. Conclusion

Bottom line: The independent claims are directed to a predictable re-grouping of multi-party calls using join/leave primitives, membership notification, and unique identifiers — each taught or suggested by the cited art in the same field. Conti et al. in particular recites the add/quit/notify/unique-ID requirements almost verbatim. A § 103 rejection combining the centralized conference-control references (4,544,804; 4,754,478) with the distributed-conferencing/protocol references (4,555,781; 4,811,334; Conti) is more likely to succeed than any single-reference § 102 attack, which the record does not support.

Two caveats I will not paper over: (i) my reference characterizations rest on abstracts/partial texts, not the full disclosures — a definitive opinion requires the complete documents and their filing/issue dates verified against the 1993 priority date; and (ii) Aldred et al. is the inventors' own paper, so it is probably not citable § 102/103 art and should not be relied upon as an invalidating reference — only as evidence of the skill level and the state of the art.