US 4635253

I'll verify against live sources before summarizing.

US 4635253 — Verification & Summary

Search scope note: I ran targeted queries against Google Patents, EPO Global Patent Index, Unified Patents, and uspto.report. I found no CAFC 2026 docket activity, no district-court litigation, and no PTAB proceeding involving US 4635253. That is consistent with its status: "Expired – Lifetime," with anticipated expiration in 2004 (2004‑08‑06 per Google Patents). I did not query PACER/CAFC directly, so treat "no litigation" as a search-based inference rather than a docket-certified conclusion.

Bibliographic data

Date discrepancy flag: Google Patents and the patent's own front page give filing 1984‑08‑06, priority 1983‑08‑12, issue 1987‑01‑06; Unified Patents shows 1984‑08‑05 / 1983‑08‑11 / 1987‑01‑05 — a uniform one-day offset, likely a time-zone/parsing artifact. I am not auto-correcting either source; the Google Patents/original-document values are the more authoritative.

Prior art cited: US 3,535,457 (Siemens); US 3,632,882 (General Datacomm); US 3,806,654 (North Electric); US 4,530,093 (International Standard Electric, "PCM Telecommunications system for voice and data," filed 1983‑07‑05). The last is the closest in subject matter and was filed just five weeks before this application's priority date.

Abstract (verbatim)

An exchange system includes a plurality of terminal devices, a plurality of interface devices for transmitting signal data and control data among the terminal devices and the interface devices on the basis of a two way bidirectional burst transmission system, and a control circuit which assigns a time slot to a PCM highway according to control data transmitted through a control highway from one of the interface devices, and transmits within the assigned time slot the signal data through the PCM highway to a selected one of the interface devices, through the PCM highway. Each of the terminal circuits generates 2-bit mode data representing the type of the terminal device. The control circuit assigns one or two time slots to the PCM highway according to the mode data.

Claims — plain-language overview

The patent has 3 claims; claim 1 is the sole independent claim. Claims 2 and 3 depend from claim 1.

Claim 1 (independent) — an exchange system, comprising:

• Several interface means (the digital/analog line circuits, e.g. items 10/11) and several terminal means coupled to them, sending both signal data (payload — voice/synchronous data D1 and asynchronous data D2) and control data in a fixed 13-bit burst format (F + D1(8) + D2(2) + C + P).
• A control highway carrying the control data, and a separate PCM highway carrying the signal data.
• Control means that: (a) assigns a time slot on the PCM highway whose length is determined by mode data supplied by an interface and forwarded over the control highway; (b) transmits, inside that assigned time slot, only the data part specified by the mode data (i.e. D1 and/or D2, whichever is populated) to whichever interface the control data selects.
• Key structural hook: each interface must contain a mode data generator that can selectively generate one of several mode-data values representing the type of signal data that interface will carry. In the embodiment this is 2-bit mode data: 00 = asynchronous data only, 01 = synchronous data only, 10 = voice only, 11 = voice + asynchronous data.

Claim 2 (dependent on 1) — the mode data generator is 2 bits wide, representing the type of the interface.
Literal-reading note: claim 2 recites "said mode data generator (MCG)," whereas the specification consistently uses MDG (mode data generator). This appears to be a typographical error in the printed claim; per the instruction to interpret identifiers literally, I am reporting it as "MCG" in the claim text rather than silently correcting it to MDG.

Claim 3 (dependent on 2) — the control means assigns one or two time slots on the PCM highway depending on the mode data (one slot for mode 00, 01, or 10; two slots for mode 11).

What the invention is actually solving

The background section frames the problem concretely: because an exchange conventionally had to be pre-configured (via a maintenance program) with information about which type of terminal sat behind each interface circuit, any terminal swap forced a reconfiguration. The workaround of permanently assigning two time slots per interface — one for voice, one for data — meant that single-service terminals wasted one slot, degrading PCM highway utilization.

Claim 1's contribution is to make the slot allocation self-describing and dynamic: the interface itself declares its service mix via the mode data generator, the control highway carries that declaration to the central controller, and the controller then assigns a corresponding number of slots and routes only the populated data fields. The specification's FIG. 6 discussion notes that when a voice/asynchronous device replaces an asynchronous-only device (mode 00 → 11), the remaining devices need no time-slot reallocation — the unchanged interfaces keep TTS/RTS assignments and the new device simply takes the first and Nth slots.

Uncertainty / confidence

• Bibliographic data and abstract: high confidence (corroborated across Google Patents, EPO GPI, Unified Patents, uspto.report).
• Claim scope description: high confidence — based on the full verbatim claim text.
• "No 2026 CAFC/USPTO litigation": medium confidence — based on absence of results in open web sources, not a formal court-docket (PACER/CM‑ECF) or USPTO Patent Center query, which I cannot execute here.
• If you need the litigation/assignment chain certified, the authoritative sources are USPTO Patent Center for application 06/637,832 and the USPTO Assignment database (reel/frame 004602/0792, assignment recorded 1986‑09‑12).

I'll search for litigation involving US patent 4635253 across patent litigation sources.

Let me check additional litigation-specific sources.

Let me try a couple more targeted litigation sources.

Litigation Search: U.S. Patent No. 4,635,253

Result: No known litigation found.

I searched the litigation-focused sources you identified — including the Unified Patents litigation portal (portal.unifiedpatents.com), Google Patents' litigation/legal-events data for US4635253, and general federal court / PACER-style docket results — and found no district court action, ITC investigation, PTAB proceeding, or CAFC appeal that names U.S. Patent No. 4,635,253 as a patent-in-suit or challenged patent.

What I checked and what it returned

No false positives reported

Per your instruction, I excluded near-miss numbers. For transparency, the litigation results that did appear involved different patents — e.g., Encryptawave Technologies LLC v. Toshiba Corporation (E.D. Tex., 4:25-cv-00171, asserting U.S. 7,233,664), St. Clair IP Consultants v. Toshiba (D. Del. 09-354), and Toshiba Corp. v. Imation Corp. (W.D. Wis. 09-cv-305, asserting U.S. 5,831,966 and 5,892,751). None of these involve 4,635,253.

Context on why no litigation is expected

The patent record is consistent with a patent that was never asserted:

• Grant date: 1987-01-06 · Priority: 1983-08-12 · Filing: 1984-08-06
• Status: "Expired – Lifetime"; anticipated expiration 2004-08-06 (i.e., full 20-year term from the 1984 filing date run)
• Assignee: Kabushiki Kaisha Toshiba
• Maintenance fee payments through the 12-year window were recorded; the patent simply lapsed at end of term.

Because the patent expired in 2004 (before the modern NPE/PAE assertion wave), and no litigation is recorded anywhere in its file history or in the litigation databases, I have no basis to report any case.

Important caveats

• I could not fully verify PACER directly — PACER docket querying requires a paid/login-gated account, so my check relied on aggregated docket mirrors and the litigation portals above. If you need absolute certainty, a direct Docket Navigator or Lex Machina query (both gate behind subscriptions) would confirm.
• Absence of records in the sources I reached does not mathematically prove no case ever existed; it means none is discoverable in the searched sources as of this date.
• No ID was auto-corrected. "4635253" was interpreted literally as U.S. Patent No. 4,635,253 (application 06/637,832).

Bottom line: Based on available sources, there is no known litigation involving U.S. Patent No. 4,635,253.

I'll verify the absence of PTAB activity against live sources before drafting.

Let me run two more confirmatory searches against PTAB-specific docket sources.

Search complete. Both the structured ODP block and my independent web checks agree: zero AIA trial proceedings. Drafting accordingly — no proceeding numbers invented.

Proceedings overview

Total AIA trial proceedings on US 4,635,253: 0. The USPTO Open Data Portal record carries no IPR, PGR, or CBM for this patent, and my independent checks of the Unified Patents patent page and general PTAB docket mirrors returned no proceedings, no institution decisions, no Final Written Decisions, and no Federal Circuit appeals. Breakdown by status is therefore vacuously: 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution-denied.

Bottom-line defensive posture: this is not a "hardened by surviving IPRs" situation and not a "claims cancelled" situation — it is a situation where the PTAB lever was never available. US 4,635,253 issued 1987-01-06 and expired 2004-08-06, roughly eight years before the first IPR was ever filed (AIA trials began 2012-09-16). There is no inter partes review to point a defendant at, because the patent's entire enforceable life predates the AIA. The correct defensive move for anyone receiving a demand letter citing this number is not an IPR — it is a standing/enforceability kill-shot on expiration grounds.

No proceedings to report

The requested per-proceeding template (### {PROCEEDING_NUMBER} — {Petitioner} v. {Patent Owner}) has no entries. I am not populating it with placeholders.

Verified negative — what I checked

Excluded false positives (transparency): the searches surfaced Toshiba-related PTAB matters involving different patents — e.g., Toshiba America Electronic Components, Inc. v. Monument Peak Ventures, LLC, IPR2021-00330 (POP decision on fee-payment filing date), and LG Electronics, Inc., Toshiba Corp., et al. v. Straight Path IP Group, Inc., IPR2015-00209 (U.S. 6,108,704). Neither involves 4,635,253. I also saw reexamination papers for unrelated patents (control nos. 90/019,244 on U.S. 7,627,781; 90/014,871 on U.S. 10,844,697) — not this patent.

Nearest-miss check: I also looked for older post-grant activity outside the AIA framework — ex parte reexamination and the pre-AIA inter partes reexamination (which went to the Board of Patent Appeals and Interferences, not the PTAB). The Google Patents legal-events record for 4,635,253 shows only fee payments (1990-06-25 year 4; 1994-06-20 year 8; 1998-06-22 year 12) and no reexamination certificate or control number. I found no reexamination of any kind.

Strategic summary

Claim status on the PTAB ledger: all of claims 1–3 are UNTESTED at the Board — not because they survived, but because nothing was ever filed. This is a critical distinction for any defendant. There is no Final Written Decision to cite, no claim that was cancelled, no claim that was confirmed. Contrast this with the patent's own cited-by trail, which shows the '253 disclosure being absorbed into later work on ISDN-style voice/data multiplexing (e.g., U.S. 5,579,313, Fujitsu, "Method of multiplexing speech signal and control signal in ISDN B-channels") — evidence the technical space matured around it, but not evidence of a validity challenge.

Estoppel landscape: § 315(e)(2) is a dead letter here. The estoppel provision bars a petitioner "from requesting or maintaining a proceeding before the Office" or from asserting invalidity in a civil action on any ground raised or reasonably could have been raised "during that inter partes review." Estoppel attaches only when an IPR was instituted and reached a final written decision. Since no petition was ever filed, no institution ever issued, and no FWD ever entered, nothing is estopped. Every prior-art ground — including the four references cited on the face of the patent (US 3,535,457 Siemens; US 3,632,882 General Datacomm; US 3,806,654 North Electric; and most interestingly US 4,530,093, International Standard Electric, "PCM Telecommunications system for voice and data," filed 1983-07-05, just five weeks before this application's 1983-08-12 priority date) — remains equally available to every party, with no estoppel asymmetry between them. Conversely, no party has "used up" the closest art, so there is no residual-art advantage to being the first challenger.

Pattern signals: none. No repeat petitioners, no serial filing behavior, no defensive aggregator involvement. There is no Unified Patents (or RPX, or AST) IPR in the chain — notable, because Unified Patents' portal does index this patent, and it shows no challenge. There has never been a Federal Circuit appeal from a PTAB FWD on this patent, so no Rule 36 affirmances, no claim-construction precedent, nothing.

The real reason there is no PTAB history — and it isn't a compliment to the patent. Three independent bars:

1. Timing. IPRs (and CBMs) were only available for petitions filed on or after 2012-09-16; PGRs apply only where the patent's effective filing date is on or after 2013-03-16. This patent's effective filing date chain runs to 1983-08-12. It expired 2004-08-06. The IPR window opened ~8 years after the patent died.
2. Subject-matter gate. CBM review reached only patents with at least one claim directed to a financial product or service — the transitional program sunset 2020-09-16. A PCM-highway time-slot exchange system (CPC H04M 11/06, H04L 5/14/5/16, H04Q 11/04) was never CBM-eligible. PGR is categorically unavailable (pre-AIA).
3. Term. Per the Google Patents record, all three pre-URAA maintenance fees were paid (years 4, 8, and 12, the last on 1998-06-22), and the patent reached its natural end of term. The 20-year-from-filing measure (1984-08-06 → 2004-08-06) governs because the UPRA term is the greater of 17 years from issue or 20 years from filing. So this was not a fee-lapse situation — it simply ran out.

(Note: an expired patent is not categorically immune from IPR — the Board has reviewed expired patents. The point here is purely practical: with no live infringement exposure, the filing fee buys nothing.)

Recommended next steps

If you are a defendant or a recipient of a demand letter citing U.S. 4,635,253:

• Do not budget for an IPR. There is nothing for the PTAB to do. Cite the absence of proceedings and, more decisively, the expiration date.
• Lead with the expiration defense and link the primary records. The patent expired 2004-08-06 — see Google Patents US4635253A (legal status: "Expired – Lifetime," anticipated expiration 2004-08-06) and the Unified Patents record (expiration date 2004-08-05; the uniform one-day offset flagged in the earlier bibliographic section). No pre-suit demand can reach post-expiration conduct, and any damages theory is confined to a pre-2004 window now far outside § 286's six-year lookback. That is a dispositive motion, not an invalidity trial.
• If a plaintiff nevertheless pleads live infringement, the correct motion is a Rule 12(b)(6) / summary-judgment attack on the expiration and damages window — not a § 112 or § 103 fight, and certainly not a trip to Alexandria.
• If you are doing freedom-to-operate or clearance work, the patent is dedicated to the public; treat the 13-bit burst format (F + D1(8) + D2(2) + C + P) and the 2-bit mode-data time-slot assignment scheme (mode 00/01/10 → one slot; mode 11 → two slots) as public-domain prior art of independent evidentiary value, not as a live threat.

If no PTAB activity had been your working assumption and you were waiting for a challenge to appear: do not wait. For this patent, the absence of PTAB activity is not the usual "well-asserted patents eventually attract IPRs" signal. It is a structural artifact of a pre-AIA, pre-2012 expiry, and it will never change.

Caveats — read these before relying on the negative:

• My PTAB search relied on the USPTO ODP structured record plus aggregated docket mirrors and the Unified Patents portal; I could not execute a direct, login-gated PTAB E2E (ptacts.uspto.gov) or Docket Navigator / Lex Machina query. For a certified negative, pull the PTAB E2E proceeding-search by patent number and the full file wrapper via USPTO Patent Center for application 06/637,832.
• The ODP ingest is strongest for AIA-era filings. It is theoretically possible — though I found no trace of one — that an older inter partes reexamination or BPAI appeal exists in a paper-era file not surfaced by ODP. The Google Patents legal-events list shows none.
• No identifier was auto-corrected. "4635253" was read literally as U.S. Patent No. 4,635,253, application 06/637,832. Note the earlier sections' inconsistent near-miss string "4835253" appears only as a search query term in the litigation pass and is not a real patent number referenced here.

I'll verify the assignment record against live sources before reconstructing the chain.

Let me look for the recorded correspondent and any post-issuance transfer.

Ownership-chain reconstruction — US 4,635,253

Source check performed. I queried Google Patents legal events, Unified Patents' patent page (which mirrors USPTO assignment/legal-event data), USPTO legacy-assignment document retrievals, uspto.report, and general web searches for any post-issuance transfer. The USPTO Assignment Center itself (https://assignmentcenter.uspto.gov/) is a JavaScript application I cannot drive directly here, so I relied on its indexed mirrors. I found exactly one recorded assignment for this patent. I did not find any Change of Name, Security Agreement, Merger, License, or Release record. I did not find the recorded "correspondent of record" field in any source I could reach — see the caveat under the timeline.

---

Inventors

Employer determination. All three assigned to Kabushiki Kaisha Toshiba by the standard employee-inventor assignment (reel 004602/0792), and the patent issued to Toshiba — consistent with all three being Toshiba R&D employees at the 1983‑08‑12 Japanese priority filing. The application relates to Toshiba's digital switching/PABX development, matching the assignee's line of business.

Unusual-pattern check — NOT present. There is no evidence that any inventor left Toshiba. To the contrary, PatentLeaderboard's inventor profile lists Michiaki Okano with 8 granted US patents at Kabushiki Kaisha Toshiba, the first granted 1983 and the most recent December 2012 — a ~29-year continuous association. That is the opposite of the "all inventors depart within 12 months of filing" fire-sale precursor. No departure signal to report. (Confidence: medium — inferred from patent-assignment continuity, not employment records.)

---

Original assignee

Kabushiki Kaisha Toshiba (a/k/a Toshiba Corporation; historically Tokyo Shibaura Denki Kabushiki Kaisha), 72‑34 Horikawa‑cho, Saiwai‑ku, Kawasaki / 1‑1 Shibaura 1‑chome, Minato‑ku, Tokyo, Japan. A Japanese corporation; the assignee address recorded in the assignment event is 72 Horikawa‑cho, Saiwai‑ku (Kawasaki).

• Shipped a product embodying the claims? The claimed subject matter is a digital exchange / PCM PABX with mixed voice+data terminal interfaces — squarely Toshiba's digital PBX and key-telephone product lines of that era. Toshiba was an active manufacturer of digital PABX/key systems. I regard product embodiment as likely, but I am flagging this as inference from the assignee's known product line, not from a specific product record I verified in this session. Confidence: medium.
• Primary line of business: diversified electrical/electronics manufacturer — heavy electrical, semiconductors, consumer electronics, computers, and telecommunications switching equipment. The last is the relevant division here.
• Current status: Operating. Note the 2026 Bermuda Trade Mark Journal entry recording that Kabushiki Kaisha Toshiba changed its registered address from 1‑1 Shibaura 1‑chome, Minato‑ku, Tokyo to 72‑34 Horikawa‑cho, Saiwai‑ku, Kawasaki on 2026‑04‑20 — which is the same Horikawa‑cho district already shown in this patent's 1986 assignment record. Toshiba remains an active, publicly traded operating company (post-2017 restructuring of its memory business into Kioxia, in which Toshiba retained a stake).

---

Assignment timeline

Important scoping finding: The Assignment Center / legal-events record for US 4,635,253 contains only the original inventor→assignee conveyance. There are no post-issuance assignments of any kind. I am reporting that plainly rather than filling the section with invented links.

• 1984‑07‑24 (executed; Google Patents lists this as the "effective date" of the assignment) / recorded 1986‑09‑12 — Reel 004602 / Frame 0792

  • Conveyance: Assignment (Assignment of Assignors' Interest)
  • Assignors: Kiyoshi Urui; Kiyoshi Ishikawa; Michiaki Okano (jointly, all three)
  • Assignee: KABUSHIKI KAISHA TOSHIBA, a corporation of Japan, 72 Horikawa‑cho, Saiwai‑ku (Kawasaki)
  • Correspondent: Not available from the sources I could reach. Google Patents' legal-events table and Unified Patents' mirror expose the reel/frame, conveyors, and assignee, but not the recording correspondent (the attorney/agent) field. I will not guess a name. To capture this field you need the Assignment Center detail view for reel 004602/0792 or the recorded document image at https://legacy-assignments.uspto.gov/. This is the one open item in the chain.
  • Context: Standard pre-issuance employee-inventor assignment — inventors' rights conveyed to their employer before the US patent issued (recorded 1986‑09‑12, after the 1984‑08‑06 US filing and before the 1987‑01‑06 grant). Not an acquisition, fire-sale, securitization, or reorg.

• No further recorded conveyances. Post-issuance legal events are maintenance-fee payments only:

  • 1988‑12‑07 — fee-payment procedure event (payor number assigned)
  • 1990‑06‑25 — FPAY, year 4
  • 1994‑06‑20 — FPAY, year 8
  • 1998‑06‑22 — FPAY, year 12
  • 2004‑08‑06 — anticipated expiration (Google Patents: "Expired – Lifetime"). No 16th-year-cascade event beyond the 12-year payment; the term simply ended.

Date-discrepancy flag (carried forward): Unified Patents shows a uniform one-day offset (filing 1984‑08‑05, priority 1983‑08‑11, grant 1987‑01‑05) versus Google Patents/original document (1984‑08‑06 / 1983‑08‑12 / 1987‑01‑06). I am not auto-correcting. The assignment record's own dates are 1984‑07‑24 executed / 1986‑09‑12 recorded, which are independent of that offset.

---

Timeline diagram

timeline
    title Ownership of US 4635253
    1983 : Japan priority filing
    1984 : Inventors assign rights to Toshiba
         : US application filed
    1986 : Assignment recorded reel 004602 frame 0792
    1987 : US patent issued to Toshiba
    1990 : Maintenance fee paid year 4
    1994 : Maintenance fee paid year 8
    1998 : Maintenance fee paid year 12
    2004 : Patent expires end of term

---

NPE / troll-pattern signals

Net signal count: zero. No shell entity, no known asserter, no cascading transfers, no pre-litigation transfer, no fire-sale, no privateering, and no defensive-aggregator endpoint. Nothing in the record supports an NPE characterization.

---

Verdict

Insufficient data — under the taxonomy's own definition ("no records, or only the original assignment"). The entire recorded chain is a single link: three Toshiba employee-inventors → Kabushiki Kaisha Toshiba, executed 1984‑07‑24 and recorded 1986‑09‑12 at reel 004602/0792. There are no post-issuance assignments whatsoever — no NPE, no shell LLC, no defensive aggregator, no litigation funder. The 1988/1990/1994/1998 legal events are maintenance-fee payments only, and the patent lapsed at term-end 2004‑08‑06.

Substantively, this is a clean single-owner chain in which the original operating manufacturer retained the patent for its entire life — i.e., the strongest negative NPE finding the record can support. I do not elevate it to "Defensive / non-asserting" because that verdict is reserved for chains ending at RPX/AST/LOT/Unified/OIN, and no such entity appears. The only unresolved datum is the correspondent of record, which no accessible source exposed.

Verify at: https://assignmentcenter.uspto.gov/ (search patent number 4635253, or reel 004602 frame 0792) · mirror: https://assignment.uspto.gov/patent/index.html · source page for this patent: https://patents.google.com/patent/US4635253/en

---

Caveats / confidence

• Single-assignment finding: high confidence — corroborated by the patent's own front-page legal events, Google Patents, and Unified Patents, all of which show the same lone reel/frame plus fee events.
• "No Change of Name / Merger recorded": medium-high confidence — these would appear as separate conveyance types in the same legal-events feed and none do. Note that the October 1986 assignment spells the assignee "KABUSHIKI KAISHA TOSHIBA, A CORP OF JAPAN," and no separate Change-of-Name record for the historical Tokyo Shibaura Denki → Toshiba rename appears for this patent; the rename predates this chain and is not a link in it.
• Correspondent of record: not determinable from the sources reached. I have explicitly not fabricated one.

I'll verify the cited prior art against live sources before analyzing each reference.

Let me verify the family-cited references and the EP search report, which often contains additional prior art.

Most Relevant Prior Art for US 4,635,253 — Reference-by-Reference § 102 Analysis

Scope note. I verified the citation record for US 4,635,253 against the USPTO-derived full text on Google Patents (patents.google.com/patent/US4635253) and the original USPTO full-page images (patentimages.storage.googleapis.com), plus uspto.report for cross-checks. Google Patents aggregates this patent's file into four "Patent Citations" (examiner-cited) and three "Family Cites Families" (cited during prosecution of the JP/EP/CA counterparts). I treat both groups as "citations for 4,635,253" as the task directs, but I keep them clearly separated because they enter the record through different routes.

Governing-law frame. US 4,635,253 has a JP priority of 1983-08-12 and a US filing of 1984-08-06. Pre-AIA § 102 therefore applies. Two consequences matter for the analysis below:

• Anything published/granted before 1983-08-06 is available as § 102(b) art.
• Any US patent granted on an application filed before the applicant's invention date is available as § 102(e) art — this is how the examiner-cited US 4,530,093 (filed 1983-07-05, granted 1985-07-16) enters, i.e., its filing date precedes the priority date even though its grant date follows.

Headline finding. Applying the strict single-reference "arrange every element" test, none of the cited references anticipates any of claims 1–3. The claim-1 element that defeats all of them is the mode data generator located in each interface that selectively generates one of a plurality of mode data representing the type of signal data — none of the references discloses a per-interface type-declaration code fed to the switch's central controller. Several references do read on sub-elements (variable/multiple time-slot allocation), so they remain strong § 103 candidates, which is presumably why the examiner cited them.

---

A. Examiner-cited references (Google Patents "Patent Citations," 4)

1. US 3,535,457 A — "Time division multiplex system provided with bandwidth multiplication feature"

Brief description. A two-stage time-division exchange in which both voice and data subscriber stations hang off the same multiplex highways (M1, M2, "MS"). The distinguishing feature is "bandwidth multiplication": voice stations connect to a highway through a single switch (one time slot), while each data station connects through a pair of switches — i.e., it is allocated two time slots to obtain greater bandwidth. Intermediate registers (capacitors C1…Ck) temporarily store samples so that a station's samples can be shifted to an available slot, enabling time-slot interchange between path segments. Assignment is made by circulating-address delay-line storage devices (U1, U2, UK, UC).

Claim mapping / anticipation.

• Claim 3 ("one or two time slots … according to the mode data") — the number-of-slots-per-station-type idea is genuinely present: Siemens expressly gives data stations two slots and voice stations one. But selection is by hard-wired switch topology per station type, not by "mode data" of the kind claimed. No anticipation; relevant art.
• Claim 1 — No. There is no interface that generates a mode-data code, no control highway carrying such a code to a central control, and no PCM-highway time-slot assignment driven by it. Anticipation: No.
• Claim 2 — No (no 2-bit mode data generator).

2. US 3,632,882 A — "Synchronous programable mixed format time division multiplexer"

Brief description. A synchronous TDM multiplexer for many asynchronous sources having mixed rates and mixed character lengths. Its stated advance is to avoid the prior-art inefficiency of uniform slots: programmable transmitter/receiver timing generators build a frame of variable-duration time slots, "each having a predetermined duration of one to N bits," so that each source gets exactly the slot width its data needs. The timing is set by a "programable time-slot-per-channel matrix" (FIGS. 4, 4A), and frames are delimited by a Barker-type frame sync word detected at the receiver.

Claim mapping / anticipation.

• Claim 1 — This is the closest cited art on the "time slot with a length determined by …" concept: GDI discloses variable-length slots sized to the served source's data format, and a frame sync/format structure. But it is a point-to-point data multiplexer, not an exchange; there is no interface/terminal array, no control highway, no PCM switching matrix, and critically no mode data generator producing a type code that the switch reads to size the slot. The "programable matrix" is pre-set configuration, not a self-declaring terminal type. Anticipation: No; strong § 103 art.
• Claim 3 — GDI varies slot length, whereas claim 3 varies slot count (one vs. two). Not the same; no anticipation.
• Claim 2 — No.

3. US 3,806,654 A — "Arrangement for transmitting digital pulses through an analog TDM switching system"

Brief description. Digital pulse streams are carried across an existing analog PAM/TDM matrix. Incoming pulses are buffered until an assigned time slot; the number of slots assigned per frame scales with the digital data rate — ≤12.5 kb/s gets one slot per frame, 12.5–25 kb/s gets two, 37.5–50 kb/s gets four — with "stuff" pulses filling idle intervals. (The related U.S. 3,652,800, cited within it, covers the same buffer/stuff principle.)

Claim mapping / anticipation.

• Claim 3 — This is arguably the most literal match to the "one or two time slots" mechanic: North Electric dynamically assigns one, two, or four time slots per frame according to the character of the data being switched. Still, the criterion is measured data rate/elastic-store fill, not a mode-data code declared by an interface representing the type of signal data. No full anticipation of claim 3 (which requires the count be "according to the mode data"); powerful § 103 art.
• Claim 1 — No. No mode-data generator, no control highway/PCM-highway split as claimed, no terminal-array exchange.
• Claim 2 — No.

4. US 4,530,093 A — "PCM Telecommunications system for voice and data" ← closest overall

Brief description. A digital PBX handling both voice stations and data terminals, with a group controller per port group and a system controller governing a time-slotted system PCM bus. Claim 1 recites that the system controller is responsive to a signal from a group controller that a data terminal is seeking service to assign a time slot for data within a recurrence period, and is responsive to further data terminals seeking service to assign other recurrence sub-slots within the same already-assigned time slot (channel submultiplexing / sub-rate allocation). Bus/slot structure is 24-slot shelf buses mapped onto 192-slot system buses on the 125 µs frame.

Claim mapping / anticipation.

• Claim 1 — This is the nearest miss: it is an exchange, it does carry voice and data, it does have a central controller that dynamically assigns time slots in response to a signal arriving from the port/interface side, and it separates control from the PCM path. If the "signal that a data terminal is seeking service" were read as the claimed "mode data … supplied from one of said interface means," this would come close. But on the face of the reference it is a service request / call-origination signal tied to data rate, not a mode data generator that selectively generates one of several values representing the type of signal data to be carried, and the slot sizing is a sub-rate subdivision (multiple data terminals sharing one slot), not the claimed one-or-two-slot allocation per terminal type. Anticipation of claim 1: No — but this is the single most § 103-relevant reference, and the only one that could also be framed as § 102(e).
• Claim 3 — North Electric-style multi-slot scaling is not the mechanism here (this reference subdivides a slot rather than awarding an additional one). No anticipation.
• Claim 2 — No 2-bit mode-data field disclosed.

---

B. Family-cited references (Google Patents "Family Cites Families," 3)

These were cited during prosecution of the JP/EP/CA counterparts, not by the US examiner of record. I verified metadata for US 4,432,089 directly; for GB 2,085,264 A and JP 57-79793 A I obtained the Google Patents family entry only (the detail queries were cut off — see caveats).

5. US 4,432,089 A — "Digital loop transceiver for interfacing a digital PABX to a digital subscriber set via a subscriber line"

Brief description. A PABX line card and matching subscriber set each use a digital loop transceiver (DLT) operating in master/slave mode over the existing 2-wire loop, carrying two independent time-multiplexed channels (voice and data) in one burst frame (8-bit word per channel + per-channel signalling bits + parity). A time-slot-assignment-and-control (TSAC) circuit enables the codec/DLT during predetermined receive and transmit time slots assigned by the PABX for each channel. The two-wire (half-duplex) variant carries only a single data bit per frame in the second signalling position.

Claim mapping / anticipation.

• Claim 1 — Notably parallel in architecture: 2-wire bidirectional burst ("ping-pong"-type) transmission between an exchange line card and a remote terminal, a fixed multi-field frame (D1/D2/signalling/parity), and time-slot enable signals for voice and data channels. What is missing is the claimed mode data generator in the interface declaring a type, and any teaching that the slot count/length is chosen from that code. In '089 the two channels' slots are fixed by TSAC configuration. Anticipation: No; meaningful § 103 art on claims 1 and 3.
• Claim 3 — '089's dual-channel frame superficially resembles "two time slots," but the second slot is dedicated to signalling/data by design, not allocated according to mode data. No anticipation.
• Claim 2 — No.

6. GB 2,085,264 A — "Digital concentrator"

Brief description (metadata-level only). A digital concentrator for concentrating subscriber digital lines onto a TDM/PCM link. Concentrators by nature perform time-slot assignment to map many subscribers onto fewer highway slots. I was not able to retrieve its claims or full text within this session, so I cannot responsibly map it element-by-element. Directionally it is relevant to the time-slot allocation in a PCM highway sub-element of claim 1, and it is § 102(b) art — but I cannot state that it discloses the mode-data-generator element, and I have no basis to assert anticipation.

7. JP 57-79793 A — "Multi-dimension traffic time division exchanger"

Brief description (metadata-level only). A multi-dimensional-traffic time-division exchange — i.e., a TDM switch handling traffic of more than one "dimension"/class. I could not retrieve the Japanese text or an English abstract in this session, so I again decline to assert specific claim mapping. It is § 102(b) art and, by title, directed to handling mixed traffic classes in a TDM switch, which touches the claim-1 environment.

---

C. Summary table

---

D. Bottom line on anticipation

• Claims 1, 2, 3: no single cited reference anticipates. Claim 1's load-bearing limitation — a mode data generator physically in each interface that selectively generates one of a plurality of mode-data values representing the type of signal data, with the central control assigning a slot whose length/count is determined by that mode data — is not disclosed, as a whole, by any reference. Claims 2 and 3 depend from that limitation and therefore stand or fall with it.
• The most dangerous art is US 4,530,093 (exchange + dynamic data-slot assignment triggered from the interface side, and the only reference available as § 102(e) on a near-contemporaneous filing date). A § 103 combination of US 4,530,093 (dynamic slot assignment), US 3,806,654 or US 3,535,457 (number of slots varying with traffic type/rate), and US 3,632,882 (slots of programmed, source-specific format) would be the examiner's most natural attack on claim 1 and, especially, claim 3.

E. Confidence and explicit uncertainties

• High confidence: citations list and bibliographic data for the four examiner-cited US patents and for US 4,432,089 (corroborated by the US full-text/PDF images and uspto.report).
• Medium confidence on § 102(e) characterization: it rests on comparing each reference's filing date to the 1983-08-12 priority/invention proxy; the true "invention date" could differ, and a terminal disclaimer / intervening-rights analysis is outside what I can do here.
• Low confidence / unverified: GB 2,085,264 A and JP 57-79793 A — I have only Google Patents family metadata (title, dates, assignee). I did not obtain their claims or descriptions, so I make no anticipation assertion for either. If you need them fully assessed, the EP 0 135 342 search report (Family ID 15428266) and the JP H0667019 B2 file would carry the examiner's own reasons for citation.
• No identifier was auto-corrected. All numbers are read literally; note that the printed claim 2 recites "mode data generator (MCG)" while the specification uses MDG — reported as-is, not silently fixed.

I'll ground this analysis in the actual prior-art references listed on the patent page before building the § 103 combinations.

Obviousness Analysis — U.S. Patent No. 4,635,253

Under 35 U.S.C. § 103 (pre-AIA), using the references of record

Scope note: This is a technical/patent-analyst opinion on the strength of a § 103 case, not a legal conclusion or validity opinion. Per the operating rules, I interpret all identifiers literally (e.g., I do not silently correct the "MCG" typo in claim 2, or the "PG,4" scan artifact in the claim-1 preamble). Where a reference's full text was not retrievable, I say so rather than guess.

Governing law / framework. Pre-AIA § 103 controls (application filed 1984-08-06; priority 1983-08-12), so the Graham v. John Deere Co., 383 U.S. 1 (1966) factors apply, as refined by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007). The critical date for prior-art status is the 1983-08-12 Japanese priority date (JP 58-147353), assuming perfected § 119 priority.

---

1. Level of ordinary skill in the art (PHOSITA)

A person having ordinary skill would have: a B.S. in electrical engineering (or equivalent), plus roughly 2–4 years of experience in digital telephony / digital PBX line-card and PCM switching design. By 1983 the relevant knowledge base included: TDM/PCM framing (125 µs frames; 64 kbit/s voice channels), channel-associated and common-channel signaling, time-slot interchange (TSI) memories, and the two-wire time-compression-multiplexed "ping-pong" digital subscriber loop (well commercialized by Motorola's UDLT/DLT parts, see below).

---

2. A threshold claim-construction issue that drives the whole analysis

Claim 1 recites assignment of "a time slot with a length determined by mode data," while claim 3 says the control means "selectively assigns one or two time slots." Two readings are possible:

• Reading A (number of slots): "length" = the number of slots occupied (1 or 2). This is the specification's actual embodiment (the table at col. re mode 00/01/10 → 1 slot; 11 → 2 slots).
• Reading B (duration of a single slot): "length" = variable bit-duration of a slot.

This matters enormously, because each reading is literally met by a different one of the four examiner-cited references:

I flag this because the usual prosecution posture ("my slot length is determined by mode data") is weaker than it appears: whichever way the term is construed, an examiner has a reference that discloses the slot-sizing mechanism.

---

3. The prior art of record — what each actually teaches

I verified the four examiner-cited references and the three family-cited references against live sources.

3.1 US 3,535,457 — Siemens AG (1970-10-20) — "Time division multiplex system provided with bandwidth multiplication feature"

(Google Patents)

This is the most structurally damaging reference and, tellingly, the examiner cited it. Its disclosed system has:

• Voice and data subscriber stations on the same multiplex highways (M1, M2).
• Voice stations use one switch → one time slot; data stations use a pair of switches → two time slots ("data subscriber stations are connected to the multiplex highway M1 by a pair of switches, for each data station"). This is bandwidth multiplication based on the nature of the station, i.e., a de facto one-vs-two-slot assignment keyed to station type.
• Intermediate registers (capacitors C1…Ck on highway MS) that "temporarily register the samples only until they can be conveyed to the following path segment during time slots which are available." This is functionally a time-slot-conversion/temporary store, i.e., the ancestor of the '253 PCM switching circuit 14 "time slot conversion memory."
• Assigning/seizing time slots independently "per path segment" with "considerably greater… adaptability… in the respect of the seizure of time slots."

3.2 US 3,632,882 — General Datacomm (1972-01-04) — "Synchronous programmable mixed format time division multiplexer"

(Google Patents) · (uspto.report)

This reference supplies both the mechanism and, more importantly, the explicit motivation. It states the problem in the same terms as the '253 background:

"When data sources having different character lengths are multiplexed, the operation is inefficient due to the fact that there have been more bits assigned to the composite data frame than are necessary to convey the data."

And it solves it with programmable timing generators that "transmit and receive any desired sequence of variable duration time slots in a frame, each having a predetermined duration of one to N bits that can readily and efficiently accommodate a plurality of different data sources having both mixed rates and mixed character lengths." FIG. 4/4A show a programmable time-slot-per-channel matrix — i.e., the slot count/duration per channel is a programmable parameter. It also teaches frame-sync detection on a defined sync word (Barker sequence) — used for burst alignment.

Note: General Datacomm's own later EP application (EP 0251587, "Variable control and data rates in highly efficient multiplexer," (EPO)) shows this was a sustained design philosophy at that company — evidence the field viewed demand-based slot/bandwidth allocation as routine.

3.3 US 3,806,654 — North Electric (1974-04-23) — "Arrangement for transmitting digital pulses through an analog TDM switching system"

(cited on Google Patents, above)

Least directly relevant of the four; supports the general proposition that digital data and PCM voice could be carried through a common TDM switching fabric. It is a weak secondary reference.

3.4 US 4,530,093 — International Standard Electric (filed 1983-07-05; granted 1985-07-16) — "PCM Telecommunications system for voice and data"

(Google Patents) · cited in (US 4,892,804)

This is the closest art and was filed only ~5 weeks before the '253 priority date, making it § 102(e) prior art. Its claim 1 recites, verbatim from the fetched text:

"a data interface coupled to one or more data terminals over a pair of data conductors and a pair of control data conductors … said system controller responsive to a signal from a group control that a data terminal is seeking service through said system for assigning a time slot for data for the terminal seeking service within a recurrence period for that time slot, said system controller responsive to further data terminals seeking service for assigning other recurrence sub-slots…"

That is: (i) separated data bus vs. control-signal conductors, (ii) a controller that assigns a time slot responsive to a signal sent up from the interface/group control, (iii) different slot allotments depending on terminal demand/type (subrate channel multiplexing, "Channel Submultiplexing," with three disclosed table-mapping algorithms for allocating sub-slots). Its disclosure also distinguishes T1/E1 shelf buses, mapping, and 24/48-slot-group structures.

3.5 Family-cited references (also pre-1983-08-12)

• US 4,432,089 — Motorola (filed 1981-12-24; granted 1984-02-14) — "Digital loop transceiver for interfacing a digital PABX to a digital subscriber set via a subscriber line." (Google Patents) This reference is remarkably on-point for '253's two-wire bidirectional burst framing. It discloses a ping-pong / half-duplex time-compression-multiplexed loop at 256 kbit/s DPSK over a single twisted pair; a frame carrying two channels (a 64 kbit/s voice channel + data channels); a sync detector that validates a frame "of the proper frame length"; a "time slot assignment and control circuit (TSAC)" responsive to a control bus; and — critically — "neither, either or both of these channels may carry PCM voice data words, depending upon the requirements of a given subscriber." That is a channel-count that is a function of the subscriber's terminal configuration: the conceptual core of the '253 mode data.

• GB 2,085,264 — Western Electric (1982-04-21) — "Digital concentrator" (family citation); and JP S57-79793 — NEC (1982-05-19) — "Multi-dimension traffic time division exchanger." Both pre-date the priority. I could not retrieve full texts; they appear to be cumulative background on digital concentration and multi-stage TDM switching. Flagging as not fully verified.

• Also referenced in the '089 family: Wurzburg, Ser. No. 416,441, "Simultaneous Voice and Data Transmission Circuit Having a Digital Loop Transceiver," filed Sept. 10, 1982 (cross-referenced in (US 4,484,028)). If it issued, it is potential § 102(e) art on simultaneous voice + data over a digital loop. Not fully verified.

References I did not treat as prior art: everything in the "Cited By" lists (US 4,787,081; 4,796,256; 5,291,302; etc.) and family members (EP 0135342, DE 3484762, CA 1219939) — these post-date the priority date and/or are the same invention.

---

4. Element-by-element mapping of claim 1

Conclusion of mapping: every element of claim 1 except the mode data generator per se is disclosed or plainly suggested. The case therefore rises or falls on that last limitation.

---

5. Proposed § 103 combinations

Combination A (primary): ISE '093 + Siemens '457 + General Datacomm '882

• '093 supplies the system architecture: voice + data PCM PBX, separate data vs. control conductors, a controller that assigns a time slot in response to a signal from the interface that a terminal needs service.
• '457 supplies the type-dependent slot count: voice stations get one slot, data stations get two.
• '882 supplies the programmable, per-channel slot matrix and the express motivation (eliminate "more bits assigned… than are necessary").

Rationale for combining: All three are in the same field (TDM/PCM telephony and data multiplexing), address the same problem (inefficient fixed slot allocation), and their combination yields nothing more than the predictable sum of their parts. Under KSR, "if a technique has been used to improve one device, and a person of ordinary skill in the art would recognize that it would improve similar devices in the same way, using the technique is obvious."

Combination B: General Datacomm '882 + Motorola '089 + Siemens '457

• '089 supplies the two-wire bidirectional burst transfer (ping-pong, 256 kbit/s, 125 µs bursts) between exchange and subscriber set, the frame sync detector, the two-channel frame (voice + data), and the key suggestion that the number of active channels depends on "the requirements of a given subscriber."
• '882 supplies the programmable mixed-format slot assignment and the anti-waste motivation.
• '457 supplies the one-slot/two-slot type-keyed allocation.

This is the combination most likely to be run against the dependent claims, because '089's "voice, data, or both depending on subscriber requirements" is essentially the '253 mode-data concept (00/01/10/11) denuded of the explicit 2-bit encoding.

Combination C: ISE '093 + Motorola '089 (minimal two-reference case)

If the examiner were allowed only two references: '093 teaches dynamic slot assignment responsive to interface-originated signaling and separate control/PCM paths; '089 teaches the burst transfer format and subscriber-dependent voice/data channel count. The only gap is the discrete multi-value ID generator — which, on the reasoning in § 7, is arguably an obvious design choice.

Dependent claim 3 ("one or two time slots") is met literally by Siemens '457 alone.
Dependent claim 2 ("2-bit mode data") is a design choice (see § 7.3).

---

6. Motivation to combine — the KSR rationales

1. The references themselves identify the problem '253 purports to solve. '253's background complains that permanently assigning two slots per interface wastes one slot "resulting in deterioration of the transfer efficiency." '882 says almost the identical thing about wasted bits/slots; '457's whole title is "bandwidth multiplication" — i.e., not wasting bandwidth on stations that don't need it. A PHOSITA who had the '253 background's problem statement was being told by the prior art where the solution lay. This is not hindsight reconstruction; it is the actual motivation articulated in the art.

2. "Known technique to improve a similar device in the same way" (KSR). Programmable/variable slot allocation ('882) and station-type-keyed slot counts ('457) are established, off-the-shelf tools for TDM efficiency. Applying them to the '093/'089 exchange architecture is a garden-variety combination.

3. Market/design pressure. The patent's own background concedes the driving force: exchanges had to "concurrently handle both voice and data," and reconfiguring a maintenance program "every time the terminal device is changed from one to another" is administratively expensive. Design incentives of exactly this sort "can prompt variations in the prior art" (KSR). Eliminating truck-rolls/re-provisioning is a recognized market pressure, and the prior art already pointed to self-describing channels ('089's subscriber-dependent channel count).

4. Finite, predictable solutions → obvious to try. Given an exchange that must assign one or two slots by terminal type, there is a small, closed set of ways to convey the type to the controller: a strapping/ID field, a dedicated lead, or an in-band control field. '093 already carries a control-signaling path from interface to controller; '089 already carries signaling bits within the frame. Adding a 2-bit ID field is a "finite number of identified, predictable solutions."

5. No teaching away. Nothing in '457, '882, '093, or '089 warns against type-dependent or self-reported slot assignment.

---

7. The strongest non-obviousness arguments — and the rebuttals

A competent validity defense would press on the mode-data generator. Here is how each argument fares.

7.1 "The interfaces must self-describe via a mode data generator — that's more than the art."

Rebuttal: '089 discloses channels that are active "depending upon the requirements of a given subscriber," and '093 discloses a controller that acts on "a signal from a group control that a data terminal is seeking service." Both are interface-originated type/demand signaling transmitted to a central controller over a control path. The remaining delta — encoding the type in a small field rather than inferring it — is a matter of implementation detail, and the '253 specification itself treats mode data as merely 2 bits of an already-existing 12-bit control word (the C bit accumulated across 12 bursts). Adding two ID bits to an existing control message is not an inventive leap.

7.2 "The art required manual pre-configuration (maintenance program); '253 makes it automatic/dynamic."

Rebuttal: '093's claim 1 already makes assignment responsive to signaling from the interface at service-request time, not from a maintenance program. And '882's timing generators are programmable rather than hard-wired. The patent's own characterization of the prior art (that it was always manually provisioned) is accordingly overbroad and, to the extent it is contradicted by '093, subject to challenge.

7.3 Claim 2 — "2-bit mode data."

This is a classic arbitrary design choice. Given four terminal configurations (async-only, sync-only, voice-only, voice+async), 2 bits is the minimum-width encoding — the selection is dictated by the number of categories, not by invention. KSR: "if a person of ordinary skill can implement a predictable variation, § 103 likely bars its patentability." The specific mapping (00/01/10/11) is likewise a design choice with no unexpected result.

7.4 Claim 3 — "one or two time slots."

Literally disclosed by Siemens '457 (voice = 1 switch/slot; data = 2 switches/slots). Even putting '457 aside, the '253 specification's table merely follows from mode data; there is no unexpected technical result from selecting exactly "one or two."

7.5 A genuine (if narrow) foothold for patentability.

The strongest argument for '253 is not the outcome (one-or-two slots) but the architectural placement: a resident, multi-value mode data generator inside the interface means, whose output travels through the control highway and is consumed by the PCM-highway controller to size the slot. If the PTAB/court read claim 1 to require that the mode data travels on the control highway and that the same mode data both sizes the slot and selects which data field is transmitted (a dual-function limitation), then '093 (which teaches slot assignment but not a mode field routed over a discrete control highway to size the slot) and '457 (which has no such control highway at all) may not, individually or on the record as cited, disclose that specific conjunctive routing. That is the limitation most likely to have carried the claim to allowance. But a two-reference combination — '093 (control-highway-driven slot assignment) + '089 (subscriber-dependent voice/data channel count decided at the interface) — comes close, and the difference (2-bit field in the control word) is small.

---

8. Secondary considerations (Graham factor 4)

Nothing in the record supplies objective evidence of non-obviousness:

• No commercial-success, licensing, copying, or long-felt-need evidence appears in the patent or the file materials I retrieved. Toshiba's mere ownership is not evidence of nexus.
• There is a plausible "long-felt need" story (the background's complaint about maintenance-program re-provisioning), but long-felt need alone, without a showing that others failed, is weak — especially where '093 and '089 show the field was already moving toward demand-responsive allocation.
• A "teaching away" argument is unavailable: no reference disparages mode-data or variable slot assignment.

Net: No secondary considerations rescue the claims on this record; they would have to be developed de novo.

---

9. Bottom line

• Claims 1 and 3 are, in my assessment, vulnerable to a strong § 103 challenge on ISE '093 + Siemens '457 + General Datacomm '882, and separately on '882 + Motorola '089 + '457. The best combination uses:
  • '093 for the voice/data PCM exchange with separate data/control paths and controller-assigned, interface-signaled time slots;
  • '457 for type-keyed one-vs-two-slot allocation (directly meeting claim 3, and the "length determined by mode data" element under Reading A);
  • '882 for programmable variable-duration slots per channel (meeting the element under Reading B) and for the express anti-waste motivation that parallels the '253 background; and
  • '089 (family-cited, pre-priority) for the two-wire bidirectional burst framing, frame sync, and subscriber-dependent voice/data channel count.
• Claim 2 is highly vulnerable as an arbitrary 2-bit encoding choice.
• The one limitation not squarely met by the cited art is the in-interface, multi-value mode data generator routed over the control highway. Whether the claims survive turns on how tightly that conjunctive limitation is construed — it is the pivot point of the entire invalidity/validity question.

---

10. Confidence and residual uncertainty

Explicit non-findings: I have not verified that Wurzburg Ser. No. 416,441 ever issued (if it did, it is likely § 102(e) art and would strengthen Combinations B/C materially). I have not reviewed the '253 prosecution history wrapper, which may contain applicant arguments distinguishing these very references — those arguments, if any, are the best predictor of how an examiner or the PTAB would treat this case. A direct review of the JP 58-147353 priority document and the EP 0135342 examination file would also be warranted before relying on this analysis for any adversarial purpose.

Reminder: no patent number or identifier in this analysis has been auto-corrected. "MCG" in claim 2 is reproduced as it appears; "PG,4" in the claim-1 preamble is a scanned-page artifact I have not treated as claim text.