US 5915001

I'll verify this patent against live sources before summarizing.

Let me check litigation and any 2026 docket activity specifically for this patent.

Let me get the full independent claim text to summarize accurately.

Verification notes first

I ran live searches against Google Patents, Espacenet, and RPX/Unified Patents mirrors. The ID 5915001 resolves to a single, real, expired US patent — no similar-number substitution was needed or made. Everything below is grounded in those sources.

Two things I could not confirm and will flag rather than invent:

1. No CAFC 2026 docket activity was found for this patent. My searches surfaced 2025–2026 PTAB/CourtListener material, but none of it named 5,915,001. It is also expired (see below), which makes fresh appellate activity on it unlikely.
2. Total claim count is unverified. The claim text I retrieved is partial.

Bibliographic data for US 5,915,001

Note on a discrepancy: the Unified Patents portal entry lists the priority date as 1996-11-13, whereas Google Patents and Espacenet both give 1996-11-14 (= the filing date). I report both rather than picking one.

Assignee chain: Vois Corporation (1996) → Nuance Communications (assigned 2002-09-20) → security interests to USB AG, Stamford Branch (2006) → patent releases recorded 2016-05-20 under prior Morgan Stanley facilities. Nuance was itself later acquired by Microsoft (completed March 2022), though that postdates any asserted ownership of this particular patent.

Abstract (as published)

A system and method provides universal access to voice-based documents containing information formatted using MIME and HTML standards using customized extensions for voice information access and navigation. These voice documents are linked using HTML hyper-links that are accessible to subscribers using voice commands, touch-tone inputs and other selection means. These voice documents and components in them are addressable using HTML anchors embedding HTML universal resource locators (URLs) rendering them universally accessible over the Internet. This collection of connected documents forms a voice web. The voice web includes subscriber-specific documents including speech training files for speaker dependent speech recognition, voice print files for authenticating the identity of a user and personal preference and attribute files for customizing other aspects of the system in accordance with a specific subscriber.

Plain-language overview of the independent claims

Note on scope of my knowledge: I have verbatim text for claim 1, and partial text for claims 4, 6, and 8, drawn from the RPX/Insight claim display. I do not have the complete claim set or the total number of claims, so the list of "independent" claims below is based on the structure of the retrieved text and should be treated as provisional.

Claim 1 — Method of delivering caller-customized voice-based information to a caller. The system keeps caller-specific data in a computer file sitting at a URL. When a call comes in, it figures out which URL belongs to that caller, fetches that caller's file, and interprets the caller's spoken/touch-tone commands over the phone to determine what service they want. It then retrieves information that depends both on the request and on the caller's stored personal data. Critically, the retrieval works by (a) generating a database query form for the request, (b) pre-filling/customizing that form using the caller's stored information, (c) performing the database search with the customized form — where the query form itself was built by storing a markup-language voice form at a URL, playing it back to the caller to produce prompts, collecting the caller's answers, and building the query from those answers. Finally, the retrieved results are played back to the caller over the telephone. In short: personalized-profile-driven voice querying of databases, built from URL-addressable markup voice forms.

Claim 4 — Method of processing voice-based information received from a telephone caller over a computer network. The core is storing a voice form at a URL in a markup language with voice extensions, then during a calling session collecting information from the caller, and storing that collected information into a first markup-language document that includes a hyperlink to a second markup-language document. In plain terms: the call's captured input is written into a web document that is itself hyperlinked into a web of documents — i.e., the session output becomes navigable, addressable content.

Claim 6 — Method of processing voice-based information received from a telephone caller over a computer network. Very close in structure to claim 4, but the hyperlink relationship is inverted: the collected information is stored in a first markup-language document, and a second markup-language document contains the hyperlink to the first. This reads as a linking-direction variant of the same concept.

Claim 8 — System for delivering information over a telephone. A hardware/data-structure claim. It requires a business white pages database (name, address, phone number); a database query form; a first processing agent programmed to collect user information via a voice telecommunications device, insert that information into the query form, search the database by applying the form, and generate a voice web page with a URL from the results; a yellow pages database of business advertising information; a second processing agent that the generated voice web page hyperlinks to, and which searches the yellow pages database and generates its own voice web page; and a voice web browser adapted to play voice web pages to a user. This claim maps directly to the FIG. 10 "white-yellow-order page system."

Cross-cutting themes in the claims: URL-addressable caller-specific profiles; server-side form generation from markup voice forms; hyperlinked "voice web" documents as first-class network resources; and telephone playback of retrieved data. Notably, the claims as retrieved center on profile-driven query customization and voice-form publishing, with the speaker-dependent speech training/voice-print authentication material appearing predominantly in the specification (FIGS. 4–7) rather than being reproduced in the independent claims I was able to retrieve.

Litigation and status (with caveats)

The Google Patents page for this patent lists five Delaware District Court matters:

• 1:22-cv-00105
• 1:20-cv-01593
• 1:20-cv-01435
• 1:20-cv-01433
• 1:20-cv-01010

All are sourced from Unified Patents' litigation dataset (CC BY 4.0), which Google surfaces as a flat list. I could not verify from these results what role 5,915,001 played in those cases — whether it was asserted, counter-asserted, or merely referenced. Because the patent's term expired 2016-11-14, any pre-expiration damages window in a 2020/2022 suit would be narrow, and no injunctive relief would be available. I'd treat the case list as a lead to verify against PACER/Unified Patents directly, not as a settled account. I found no CAFC 2026 docket entry tied to this patent.

The patent is also cited as prior art in a number of later voice-browser/voice-portal patents (e.g., EP 1,145,226 B1; KR 10-1027548 B1; US 2014/0040745), which is consistent with its recognized position as an early voice-web/voice-browser reference.

Where I'm uncertain

• Complete claim set and total claim count for US 5,915,001 — not authoritative from my sources. Independent-claim identification above is provisional.
• Nature of the five Delaware cases relative to this patent number specifically.
• Priority date — 1996-11-14 (Google/Espacenet) vs. 1996-11-13 (Unified Patents).
• I found no CAFC 2026 activity, and I want to state that as a negative finding rather than imply I exhaustively confirmed its absence.

I'll research litigation involving US 5915001 across patent litigation databases.

Let me search for the specific Delaware case numbers identified and the parties involved.

I hit the tool-call limit before I could confirm party names, so here is what I can state with grounding, plus an explicit flag on what I could not verify.

Bottom line

I found no reliable, party-level litigation record for US 5,915,001 in the sources I was able to reach. The only litigation data I can ground is a list of five case dockets in the U.S. District Court for the District of Delaware that are associated with the patent in Unified Patents' litigation database, as reproduced on the Google Patents page for US5915001A (the authoritative record supplied):

Critically: the Google Patents litigation section gives only jurisdiction, case number, and filing-year, and describes the data source as Unified Patents' "Litigation Data." It does not list plaintiff(s), defendant(s), precise filing dates, or outcomes for any of these five dockets.

What I could NOT confirm (and will not guess)

• Plaintiff(s) and defendant(s) for all five dockets.
• Exact filing dates (only the year is embedded in the case numbers: 2020 for four, 2022 for one).
• Outcomes / current status (e.g., dismissal, settlement, judgment, voluntary dismissal, transfer).
• Whether any of these cases actually asserted US 5,915,001 (a docket can appear in a patent's portfolio/litigation listing for reasons other than direct assertion, e.g., related-case linkage).

My searches (including on "5915001," the case numbers, and Unified Patents' litigation portal) did not surface the case captions. A search for case number 1:20-cv-01433 returned an unrelated Court of Federal Claims case (Begay v. United States), which is not the D. Del. matter referenced here — so I am deliberately not mapping that caption onto this patent. I have no verified plaintiff/defendant names for US 5,915,001 and will not supply them from memory.

Grounded context on the patent (from the authoritative record)

These bear on who would plausibly be the parties, though they are not litigation facts themselves:

• Original assignee: Vois Corp; inventor: Premkumar V. Uppaluru; filed/priority: 1996‑11‑14; granted: 1999‑06‑22.
• Assignment history (as listed): assigned to Nuance Communications (recorded 2002‑09‑20); subsequent security interests to USB AG, Stamford Branch (2006); later releases recorded in 2016 in favor of Nuance/ScanSoft/SpeechWorks entities.
• Current assignee (per page): Nuance Communications Inc.
• Family members that also carry their own litigation: US 6,400,806 B1 and US 6,885,736 B2 (both listed as priority-linked), plus PCT/US1997/020752 (WO1998021872A1) and AU52566/98 A.
• Legal status: "Expired - Lifetime." Note: the page lists an "anticipated expiration" of 1996‑11‑14, which is the same as the filing/priority date and is almost certainly a metadata artifact (a 20‑year term from 1996‑11‑14 would run to 2016‑11‑14). I am reporting it as literally displayed rather than correcting it. This matters because the 2020/2022 docket dates postdate a nominal 2016 expiry, so the term/status of the patent (e.g., any maintenance-fee lapse, revival, or a suit on a related continuation) should be independently checked before relying on these entries as active infringement suits.

How to get the missing party/status data

To fill the gaps without guessing, the definitive sources are:

1. CourtListener / RECAP dockets for each number, e.g. https://www.courtlistener.com/docket/?q=1:20-cv-01010 (and the analogous queries for 1:20‑cv‑01433, 1:20‑cv‑01435, 1:20‑cv‑01593, 1:22‑cv‑00105, filtered to the District of Delaware).
2. PACER docket for the District of Delaware (DED) for each case number.
3. Unified Patents Portal patent-detail/litigation views: https://portal.unifiedpatents.com/patents/patent/5915001 and https://portal.unifiedpatents.com/litigation/caselist?patents=5915001.
4. PTAB (for any IPR/PGR on the '001 patent) via the USPTO PTAB End‑2‑End system. In my searches, the only "5915001" hits in PTAB/IPR materials were prior-art citation strings inside another party's exhibits (e.g., an Apple IPR exhibit listing prior-art search terms and an EP search report citing US 5,915,001) — not a PTAB proceeding on the '001 patent itself.

Summary statement required by your instruction: Based on the records available to me, litigation involving US 5,915,001 is documented only as five District of Delaware dockets (1:20‑cv‑01010, 1:20‑cv‑01433, 1:20‑cv‑01435, 1:20‑cv‑01593, and 1:22‑cv‑00105). I could not verify the plaintiff, defendant, exact filing date, or outcome for any of them, and I am not aware of any additional confirmed litigation or PTAB proceeding. The party identity and disposition of those five Delaware cases should be confirmed via PACER/CourtListener before being relied upon.

I'll verify the structured data against live sources before writing anything up.

The first pass surfaced only same-numbered other patents (7,139,591; 8,650,591; 11,818,591). Let me run targeted checks on this patent's number, its continuations, and its known litigation track.

No hits yet on the '001 patent itself. Let me run a few final confirmation searches — looking for any petition caption using this patent number, any ex parte reexam, and the patent's litigation history.

Proceedings overview

Zero AIA trial proceedings are on file for US 5,915,001 — the structured "PTAB proceedings on file" block reports no AIA trials from the USPTO Open Data Portal ingest (breakdown: 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denied), and my independent web searches for petition captions, PTAB E2E and Docket Alarm hits, and the patent number itself surfaced no IPR, PGR, or CBM petition ever filed against this patent. Bottom line for a defendant: the patent is untested at the PTAB, so there is no cancellation ruling to hide behind — but it is also expired as of 2016-11-14, which is the far stronger defensive fact. No IPR-based defense exists and none is needed: the absence of PTAB activity here is almost certainly explained by the expiration, not by the patent's strength.

No proceedings to report — and an important number-collision warning

There is no proceeding number to list, so I am not going to manufacture one. Because "591" / "…591…" is an extremely common suffix, my searches returned heavy false-positive traffic from unrelated patents. Anyone doing a quick validity check on 5,915,001 should know these are not this patent:

I also note that Nuance — the current assignee of 5,915,001 — has been a frequent patent owner in PTAB proceedings on other patents in its portfolio (e.g., MMODAL LLC v. Nuance Communications, Inc., IPR2018-01431, on US 8,117,034, where a Final Written Decision of 2020-02-29/2020-04-03 found claims 8 and 13 unpatentable). That generalized litigation posture does not extend to the '001 patent, and I found no evidence it was ever challenged.

Verification note: I searched via public web sources and could not query PTAB E2E or Patent Center directly. Treat "no PTAB activity" as well-supported but independently confirmable at PTAB E2E and USPTO Patent Center before relying on it in a filing.

Strategic summary

Claim status: 100% untested. No claim of US 5,915,001 has ever been canceled, confirmed, or construed by the PTAB. There are no FWDs, no institution decisions, no settlement terminations, and therefore no Federal Circuit appeal to point to. For a defendant, that means you cannot say "claim 1 is dead" — but you also face no adverse PTAB precedent finding the claims valid over specific art. The claim set is whatever it was on 1999-06-22, unamended and un-reviewed.

The dispositive fact is the expiration date. The record shows an anticipated expiration of 2016-11-14 (20 years from the 1996-11-14 filing date) and a current legal status of "Expired - Lifetime." Practically:

• Prospective infringement is impossible. The only recoverable conduct is past infringement, capped by the 35 U.S.C. § 286 six-year lookback from the complaint date. The last possible infringing acts occurred no later than 2016-11-14, so a suit filed after roughly 2022-11-14 reaches nothing.
• The Delaware District Court cases linked from the patent's record (1:20-cv-01010, 1:20-cv-01433, 1:20-cv-01435, 1:20-cv-01593, 1:22-cv-00105) sit in exactly that window. I could not confirm from the sources available to me which patents in the family those complaints actually asserted, or what claims were at issue — do not assume from the Google Patents/Unified Patents linkage that the '001 patent itself was asserted. That is a fact to pull from the dockets, not infer.

Estoppel landscape: completely clean. Because no IPR was ever instituted against this patent, 35 U.S.C. § 315(e)(2) estoppel has never attached to anyone. A defendant today retains the entire field of prior art — patents, printed publications, and system art — with no "raised or reasonably could have raised" bar. The only limits are statutory, not estoppel-based: a PGR is unavailable (the patent is a pre-AIA 1999 grant, well outside the 9-month window), and CBM review is unavailable (its AIA transition period ended 2020-09-16, and the patent has expired). That leaves IPR and ex parte reexamination as the only Office-side invalidity vehicles, both of which remain available against an expired patent for purposes of past-damages exposure — though for an expired claim the Board applies Phillips claim construction rather than the broadest reasonable interpretation, which modestly favors the patent owner.

Pattern signals. No petitioner has filed anything against this patent, so there is no serial-petitioner or multi-petition pattern. The patent owner has not pursued PTAB appeals on this patent (there is nothing to appeal). Unified Patents appears in the record only as the source of the litigation dataset, not as a petitioner — I found no defensive-aggregator challenge. One further family note: the continuation US 6,885,736 (app. 10/057,508, granted 2005-04-26) carries a terminal disclaimer, and both it and US 6,400,806 (app. 09/286,194) claim priority to the same 1996-11-14 filing. If a demand letter is in play, confirm whether it actually cites the '001 patent or a live-family member with a different term — I did not find IPRs against those either, but I did not exhaustively verify them.

Recommended next steps

1. Do not build a defense around a nonexistent FWD. There is no final written decision to quote and no IPR number to cite. Any brief or demand-letter response asserting that "the PTAB canceled claims of 5,915,001" would be a sanctionable misstatement.
2. Attack on the expiration and the § 286 window instead. Pull the complaint date from the docket and compare it to the 2016-11-14 expiration. If the accused conduct postdates expiration, or the suit was filed more than six years after expiration, the damages case is exposed to a straightforward limitations challenge. Confirm the expiration date on the face of the patent and in Patent Center before relying on it.
3. Since no proceeding is pending, there are no trial-stage milestones to track — no institution deadline, no oral hearing date, no § 316(a)(11) one-year FWD deadline. If you need Office-side invalidity work, you must originate it yourself.
4. Re-confirm the null result directly against the Office's own systems, since I worked from the ODP block plus web search only: PTAB E2E, Patent Center, and the patent's Google Patents record. For any litigation-side questions, CourtListener and the D. Del. dockets are the right sources for which patents each 2020/2022 case actually asserted.
5. Read the absent-IPR signal correctly. Ordinarily, "no IPRs" on a well-asserted patent is suspicious — it can mean the art is weak, or that targets folded quickly. Here the simpler explanation is that the patent expired on 2016-11-14. Weigh the absence of PTAB art accordingly, and concentrate diligence on the family members that still had term when the Delaware suits were filed.

I'll verify the assignment chain and litigation records against live sources before reconstructing the timeline.

Key data point confirmed: reel 017435/0199 is a security agreement (Nuance → "USB AG, Stamford Branch," effective 2006-03-31). Now let me pin down the Delaware cases and the Vois→Nuance acquisition.

Ownership & Assignment Reconstruction — US 5,915,001 A

Patent: System and method for providing and using universally accessible voice and speech data files
App. 08/748,943 · filed 1996-11-14 · granted 1999-06-22 · statutory term ended 2016-11-14 (Expired – Lifetime)
Source of record used: Google Patents legal-events record (https://patents.google.com/patent/[US5915001A](/patent/US5915001A)/en#legalEvents), which mirrors the same underlying USPTO assignment data, corroborated against sibling filings in the same family.

Data-integrity caveat before you read the timeline. I was not able to complete a direct reel/frame-level query against USPTO Assignment Center (assignmentcenter.uspto.gov) in this session — the tool budget was exhausted. Everything below is grounded in the legal-events record actually fetched, plus corroborating family-member records that surfaced in search. Four of the six recorded events do not have a reel/frame visible in the fetched data. I have marked every field I could not verify as not surfaced rather than guessing. In particular, the correspondent-of-record field — the single most useful signal you asked for — is not exposed in the source I could retrieve. I say so plainly in Signal 3 rather than inferring a name.

Inventors

Notes and pattern checks:

• Sole inventor. No joint inventors are named on the issued patent or on the two continuation filings in the family (App. 09/286,194 → US 6,400,806; App. 10/057,508 → US 6,885,736), both of which list Uppaluru alone. A single-inventor, employer-assigned software patent is unremarkable; it does not by itself indicate a portfolio-sale setup.
• Name variant in the record. The EPO search report for EP 1 073 036 renders the inventor as "UPPALURU PREKKUMAR V." — a transliteration/OCR variant of the same person. Per your operating rule I am flagging it, not silently correcting it.
• Unusual-pattern test — inventor departure within 12 months of filing: NOT ESTABLISHED. The only evidence available is that the inventor's participation in the family extended at least to the 2002-01-25 continuation filing (App. 10/057,508), i.e. ~6 years past the 1996 filing. That is weak evidence against the "inventor bails immediately, portfolio gets flipped" pattern. I found no record of his resignation or of a post-Vois employer in the sources I could reach. (I have an unverified recollection that Uppaluru later led Movius Interactive, a voice-messaging company — I could not confirm that in this session and it should not be relied on.)

Original assignee

Vois Corporation (assignee per the 1996-11-14 assignment, executed on the filing date).

• Line of business. Speech/voice telephony infrastructure and "voice web" / voice-portal services. The specification itself is the best evidence of the operating business model: voice web gateways and sites, HVML (HyperVoice Markup Language) pages played over ordinary telephones, an 800-number calling-card account model, and the corporate URL http://www.voiscorp.com used throughout as the example domain.
• Did they ship a product embodying the claims? Moderate confidence, not independently verified. The spec describes a deployed service topology ("subscribers 107 access the voice web system 100 by dialing a single '800' service phone number"), which reads as a description of a real Vois service rather than a paper design — but I could not confirm a commercially shipped Vois product in this session.
• Current status. Vois Corporation no longer holds this patent; its rights were assigned to Nuance Communications by an assignment whose recorded date is 2002-09-20. Whether that transfer was a merger, a stock purchase, or an asset purchase is not determinable from the record I retrieved — I will not characterize it further. In practical terms Vois is defunct as an independent entity, and there is no Vois-side assignment activity after 2002.

Chain terminus / current owner: Nuance Communications, Inc. (Burlington, MA at the time of its 2017 litigation; Delaware corporation). Nuance was acquired by Microsoft under a merger agreement disclosed in Nuance's 2021 proxy statement ($56.00/share in cash; Nuance surviving as a wholly owned Microsoft subsidiary). Nuance had earlier spun off its automotive business as Cerence Inc. (Oct 2019), with a Separation and Distribution Agreement that includes IP licensing provisions — relevant background, but not an assignment of this patent.

Assignment timeline

Important framing: the six ownership-relevant events below are, in substance, only two ownership transfers (1996 inventor→Vois; 2002 Vois→Nuance). The 2006 and 2016 entries are liens and lien releases — recorded in the same database as "assignments" but they convey no ownership. Do not read the 2006 "Assigned to USB AG" lines as a sale of the patent.

• 1996-11-14 (executed) / recorded 1996-11-14 — Reel not surfaced

  • Conveyance: Assignment of assignors' interest
  • Assignor: Premkumar V. Uppaluru (individual)
  • Assignee: VOIS CORPORATION
  • Correspondent: not surfaced in the fetched record.
  • Context: Standard employee/inventor-to-employer assignment, executed on the filing date — not a fire-sale or reorg.

• 2002-09-20 (recorded) — Reel not surfaced (execution date not surfaced)

  • Conveyance: Assignment
  • Assignor: VOIS CORPORATION
  • Assignee: NUANCE COMMUNICATIONS (later Nuance Communications, Inc.)
  • Correspondent: not surfaced. Only one appearance in this chain, so even had it been captured it would not by itself meet your recurrence test.
  • Context: Acquisition / consolidation of the Vois speech portfolio into an operating acquirer — this is the only true ownership transfer after issuance, and it happened 17 years before the earliest litigation record on this patent page.

• 2006-03-31 (effective) / recorded 2006-04-07 — Reel 017435/0199

  • Conveyance: Security Agreement (grant of security interest — not a conveyance of title)
  • Assignor / grantor: NUANCE COMMUNICATIONS, INC. (Massachusetts)
  • Assignee / secured party: "USB AG, STAMFORD BRANCH," Connecticut — as literally recorded.
  • Correspondent: not surfaced.
  • Context: Securitization / collateral pledge securing Nuance debt. Corroboration is strong: the identical reel/frame 017435/0199 with effective date 20060331 appears on sibling Nuance-family patents (US 5,809,494; US 6,490,549; US 5,611,002), all reading "SECURITY AGREEMENT; ASSIGNOR: NUANCE COMMUNICATIONS, INC.; REEL/FRAME:017435/0199 … Owner name: USB AG, STAMFORD BRANCH, CONNECTICUT."
  • Identifier fidelity note (per your rules, no auto-correction): the record literally reads "USB AG." The well-known counterparty is UBS AG, Stamford Branch, and this is almost certainly a long-standing transcription error carried in the assignment data. I flag it; I do not correct it.

• 2006-08-24 (recorded) — Reel not surfaced on this patent (the corresponding released agreement is cited at 018160/0909; see next entry for the inference and its limits)

  • Conveyance: Security Agreement (second lien/refinancing pledge)
  • Assignor / grantor: NUANCE COMMUNICATIONS, INC.
  • Assignee / secured party: "USB AG. STAMFORD BRANCH" (punctuation as recorded)
  • Correspondent: not surfaced.
  • Context: Securitization / refinancing collateral pledge. The patent page shows a second 2006 security agreement, indicating a refinancing rather than a single lien.

• 2016-05-20 (recorded) — Reel cited: 017435/0199 — Release of the 2006-04-07 security interest

  • Conveyance: Patent Release (termination of security interest)
  • Assignor (releasing party): MORGAN STANLEY SENIOR FUNDING, INC., as Administrative Agent
  • Assignee / beneficiary of release: Nuance Communications, Inc. (as grantor — the collateral is freed back to the grantor)
  • Correspondent: not surfaced.
  • Context: Lien release / collateral cleanup roughly coinciding with the patent's expiration. Note the agent mismatch worth flagging: the 2006 liens were recorded in favor of USB (UBS) AG, Stamford Branch, but the 2016 release is executed by Morgan Stanley Senior Funding, Inc. as Administrative Agent. That is consistent with an agency/assignment-of-lien succession within a lender syndicate — I am labelling this an inference, not a recorded fact, because the intermediate agency-transfer document was not surfaced to me.

• 2016-05-20 (recorded) — Reel cited: 018160/0909 — Second patent release

  • Conveyance: Patent Release
  • Assignor (releasing party): MORGAN STANLEY SENIOR FUNDING, INC., as Administrative Agent
  • Assignee: Nuance Communications, Inc. (grantor)
  • Correspondent: not surfaced.
  • Context: Lien release of the second (2006-08-24) security agreement. The reel/frame 018160/0909 identifies the released agreement; the release's own recording reel/frame was not surfaced — do not conflate the two.

• 2016-11-14 — Anticipated expiration (legal status: Expired – Lifetime). Not an assignment; included because it is the terminal event that forecloses any later assertion of this patent.

Post-issuance assignee changes: none after 2002-09-20. No transfer to a licensing entity, no NPE, no defensive aggregator appears anywhere in this chain.

Litigation records attached to this patent page (D. Del.), reported separately from assignments:

Critical reading of that table: US 5,915,001 expired 2016-11-14, so it could not have been asserted in a 2020 or 2022 complaint. The five Delaware entries therefore cannot be assertions of this patent; they most plausibly attach to family members (US 6,400,806, expiring 2019; US 6,885,736, expiring 2022, terminal-disclaimer-encumbered) or are an artifact of Google's docket-linking. Either way, there is no recorded assignment of US 5,915,001 to Digi Portal LLC or to IP Edge. For completeness on the "Nuance as litigant" question: Nuance itself is a plaintiff-side operating company — Nuance Communications, Inc. v. MModal LLC, C.A. 17-1484-MN-SRF (D. Del.), complaint filed 2017-10-23, asserting US 9,564,126; 7,379,946; 6,308,158; 6,766,295; 6,999,933; and 8,117,034 against a direct competitor, announced by Nuance in an Oct 23, 2017 press release quoting VP of Litigation & IP David Greenbaum. Nuance also appears on the defendant side of NPE suits (e.g., AI Visualize, Inc. v. Nuance Communications Inc., D. Del. 21-1458).

Timeline diagram

timeline
    title Ownership of US 5915001
    1996 : Uppaluru assigns to Vois Corporation
         : Application filed 14 Nov 1996
    1999 : Patent issued 22 Jun 1999
    2002 : Vois assigns to Nuance Communications
    2006 : Security agreement recorded Reel 017435 0199
         : Second security agreement 24 Aug 2006
    2016 : Liens released by Morgan Stanley as agent
         : Patent expires 14 Nov 2016
    2022 : Nuance becomes Microsoft subsidiary

(Reel 018160/0909 is the second released security agreement and is omitted from the diagram only for width; it belongs on the 2016 release line.)

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. There is no transfer to an entity with an "IP / Patents / Licensing / Holdings / Ventures" suffix, no single-purpose LLC, and no registered-agent-service address anywhere in the chain. The only post-issuance ownership assignment (recorded 2002-09-20) runs from an operating speech company to a larger operating speech company, Nuance. The 2006 and 2016 entries are liens and lien releases at Reel 017435/0199 and Reel 018160/0909 and convey no title.

2. Known asserter in the chain — NOT PRESENT for the chain. No assignee in this patent's history matches Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or any Spangenberg entity. Current owner Nuance Communications (now a Microsoft subsidiary) is an operating company, not an NPE-list member. Separate, non-chain observation: one of the five Delaware docket entries on this patent's page — 1:20-cv-01010, Digi Portal LLC v. Thryv Inc. — is an IP Edge LLC–affiliated NPE case per Unified Patents. Because US 5,915,001 expired in 2016 and no assignment to Digi Portal or IP Edge exists, that entry is not evidence of NPE activity in this chain.

3. Repeat correspondent across the chain — INSUFFICIENT DATA (cannot be scored). The correspondent-of-record field does not appear in the Google Patents legal-events data, and I could not reach USPTO Assignment Center to retrieve it. I am recording this as an unresolved gap rather than as a negative finding — it is exactly the field your methodology weights most heavily, and it is the one field this session could not produce. To close it you need the per-document Assignment Center records for the 1996-11-14, 2002-09-20, 2006-04-07, 2006-08-24 and 2016-05-20 filings. My prior expectation, given the pattern (an operating-company chain with lender liens), is that these will show corporate-side counsel and lender counsel — i.e. different correspondents on the 2002 transfer versus the 2006/2016 lien work — which would fail the recurrence test; but that is an expectation, not a finding.

4. Cascading transfers — NOT PRESENT. One ownership transfer after issuance (2002-09-20), then liens 16 and 20 years ago. Nothing resembling consecutive LLC-to-LLC hops in a sub-24-month window.

5. Pre-litigation transfer — NOT PRESENT. The last ownership transfer (2002-09-20) precedes the earliest listed Delaware case (2020-07-28) by roughly 18 years. There is no assignment within six months of any listed suit naming this patent.

6. Bankruptcy fire-sale — NOT PRESENT. No recorded Chapter 7/11 sale, no trustee conveyance, no stalking-horse transfer involving this patent or the Vois/Nuance assignors. (The closest bankruptcy in the surrounding record is the defendant-side In re Legend Parent, Inc., Bankr. S.D.N.Y. No. 14-10701 — a 2014 MModal affiliate reorganization cited in Nuance's 17-1484 complaint. It has no connection to this patent's chain.)

7. Privateering — NOT PRESENT. No operating-company-to-NPE transfer and no evidence of Nuance (or Vois) funding assertion through a third-party vehicle on this patent. Nuance's own enforcement (17-1484 against MModal) is first-party, named-plaintiff litigation by the patent owner.

8. Defensive aggregator (anti-NPE) — NOT PRESENT. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN. It terminates at Nuance/Microsoft — an operating company — and then the patent expired. So the patent is not "neutralized by aggregation"; it is neutralized by the passage of its 20-year term on 2016-11-14.

Verdict

Operating-company assertion.

Justification: the entire chain is two operating companies and two lender liens — inventor Uppaluru → Vois Corporation (assignment executed 1996-11-14, the filing date), then Vois Corporation → Nuance Communications (recorded 2002-09-20) — followed only by security interests at Reel 017435/0199 (effective 2006-03-31) and a second 2006 pledge, both released 2016-05-20 by Morgan Stanley Senior Funding as Administrative Agent against the released agreements at Reels 017435/0199 and 018160/0909. The current owner, Nuance Communications (a Microsoft subsidiary since 2022), ships speech products and sues actual competitors — Nuance v. MModal, C.A. 17-1484, complaint filed 2017-10-23. Zero of the eight NPE signals is present. Two caveats you should carry forward: (a) the correspondent-of-record was not retrievable in this session, so Signal 3 is a genuine gap, not a clean negative; and (b) US 5,915,001 expired 2016-11-14, so it was never assertable in the 2020/2022 Delaware dockets listed on its page — the Digi Portal LLC / IP Edge entry (1:20-cv-01010) is therefore not attributable to this patent's chain.

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ — and Google Patents legal events — https://patents.google.com/patent/US5915001A/en#legalEvents. To complete this file, pull the five per-document records above to capture reel/frame and correspondent for each; those are the two fields still missing from every link in the chain.

I'll verify this against live sources — specifically the prior-art "Patent Citations" section of US 5,915,001 on Google Patents/USPTO mirrors.

The Unified Patents "Patent Art" list is for the continuation (6,400,806), not the original's own front-page citations. Let me target the cited references on US 5,915,001 itself.

The results are returning forward citations (later documents citing '001). I need the front-page references cited in US 5,915,001. Let me target the FPO/Justia citation tables.

I hit the tool-call ceiling on follow-ups, so I'll answer now with what the live searches actually returned — and I'll be explicit about the one thing I could not retrieve verbatim.

1. ID verification (literal, no substitution)

5915001 resolved to US 5,915,001 A — System and method for providing and using universally accessible voice and speech data files, Uppaluru, Vois Corporation, filed 1996-11-14, published 1999-06-22, expired-lifetime 2016-11-14. No similar-number substitution was required or made.

One artifact to flag rather than auto-correct: the Google Patents page for US 680,753 A ("Gate," 1901) displayed a "Family Applications" table listing US5915001A as an application claiming US 680,753. That is plainly a scraped-page data error — the 1901 gate patent has no family relationship to a 1996 voice-web application. Reported as-is, per the literal-interpretation rule.

Unchanged discrepancies from the prior section (I found no new evidence resolving them):

• Priority: 1996-11-14 (Google/Espacenet) vs. 1996-11-13 (Unified Patents).
• Publication: 1999-06-22 (Google/Espacenet) vs. 1999-06-21 (Unified Patents family list).

2. Provenance caveat on the prior-art list — read this before the tables

What I could not do: my searches repeatedly returned forward citations (documents citing '001) rather than the front-page "References Cited" list of '001 itself. I could not retrieve US 5,915,001's own examiner-citation table verbatim. I will not fabricate it.

What I did retrieve: the "Patent Art (83)" citation set published on Unified Patents' page for the continuation US-640,0806-B1 (portal.unifiedpatents.com/patents/patent/US-6400806-B1). That set is family-derived — it even includes US-5915001-A itself — so it is a proxy, not certification, and the result was truncated at roughly 59 of 83 entries. Treat every row below as "cited somewhere in the '001/6,400,806/6,885,736 family," not "certainly on the '001 front page."

Why that's still analytically load-bearing: the pre-1996-11-14 items dominate this set, which is exactly the citation profile you'd expect of the earliest-filed member ('001). Post-1996 items in the set (e.g., US 5,930,804, priority 1997-06-09) most plausibly attach to the later continuations.

Legal frame (pre-AIA, since '001 filed 1996): §102(b) for publications more than one year before filing; §102(e) for US patents granted on applications filed before the applicant's date. Since '001's operative date is 1996-11-14, a reference needs an effective filing/publication date before that to be §102 prior art at all. A reference that postdates '001 cannot anticipate it — that disqualifies every forward citation, including US 6,400,806 and US 6,885,736 (the patent's own continuations, which are family, not art).

3. Prior-art inventory with §102 claim mapping

I have only partial claim text of '001 (claim 1 verbatim; claims 4, 6, 8 partial). So the "claim(s)" column is a screening judgment, not a legal conclusion. Recall the four claim themes: (A) URL-addressable caller profile driving a customized database query form, played over the phone (claim 1); (B) storing caller-collected input into a markup-language document hyperlinked to a second such document (claims 4, 6); (C) white-pages/yellow-pages/order-entry system with two agents and a voice browser (claim 8).

3.1 The dangerous short list — closest single-reference §102 candidates

3.2 IVR / telephone-service platform art (§102(b))

3.3 Speech recognition / speaker-dependent training (§102(b))

3.4 Network security / distributed authentication (§102(b)–(e))

3.5 Information retrieval / profile & document infrastructure (§102(b))

4. Honest limits on the §102 conclusions

• No retrieved reference alone appears to anticipate the full combination of claim 1. The novelty of '001 sits in the conjunction: URL-addressable caller profile → customized query form → markup-language voice form played to the caller → telephone playback. The nearest single-reference threats are US 5,884,262 and US 6,240,448 (network audio access), US 5,537,586 (profile-constrained retrieval), US 5,799,063 (pre-recorded audio over the Internet), and US 5,526,520 (hypermedia linking for claims 4/6). Expect the real §102 fight — had one occurred — to have been narrower claims, with §103 combinations doing the heavy lifting.
• Claim 8 is the most exposed. US 5,448,625 (telephone advertising) + US 5,479,510 (card payment) + US 4,253,157/US 5,255,305 (telephone database access) + US 5,517,605/US 5,530,852 (document generation) collectively cover much of its white-pages/yellow-pages/order-entry architecture. I could not identify a single reference containing all of claim 8's agents.
• The voice-imprint/speech-training subject matter (FIGS. 4–7) faces the densest §102(b) wall — the five Nuance/Verizon 1990-05-14 references are all filed six years before '001 and cover simultaneous telephone-network speaker verification and recognition. That the same entity later acquired '001 is a noteworthy commercial fact, not a legal one.
• Date certainty: for most rows I have priority dates only, as published by Unified Patents — not verified issue dates. For §102(e) the filing date governs, so the priority dates above are the operative ones, but I have not independently confirmed each filing date against the USPTO Patent Center. Where a §102(e) reference is close to 1996-11-14 (notably US-6233318-B1, 1996-11-04), the as-filed disclosure must be checked directly.

5. Forward-citation context (cannot anticipate, included for completeness)

Two EPO search reports treat '001 as background art (category "A"), which is the standard fate of an early pioneer reference: EP 1,073,036 A3 cites US 5,915,001 against claims 1–55 of that EP application, pinpointing col. 2 lines 21–41, col. 8 lines 25–29, and col. 22 line 30 – col. 23 line 51 (this incidentally confirms '001's specification runs past column 22). EP 1,156,649 A1 ("Voice portal hosting system and method") cites US 5,915,001 as [A] alongside EP 0,782,318 A2 as [Y]. KR 10-1027548 B1 (Motorola Mobility, "voice browser dialog enabler") lists US5915001 A under its searched prior art. None of these can anticipate '001 — they all postdate it.

6. Open items to resolve with primary sources

1. US 5,915,001's actual front-page "References Cited" table (USPTO Patent Center / FPO / Justia) — my proxy list is family-derived and truncated at ~59/83.
2. Full claim set and total claim count of '001 (carried over unresolved from the prior section).
3. Issue dates for each cited reference, to confirm §102(e) qualification beyond the priority dates reported.
4. Whether the post-1996 entries in the "Patent Art (83)" set attach to '001 or only to US 6,400,806 / US 6,885,736 — I inferred the latter but did not verify.

I'll research the prior art on the '001 patent's face and the voice-web/IVR art cited against and around it.

Good haul. Now let me pin down the '001's own cited references and the full claim 1 text so I can map elements precisely.

Obviousness Analysis — US 5,915,001 under 35 U.S.C. § 103

0. Sourcing caveat you should read first

I could not retrieve the '001 patent's own front-page "References Cited" / IDS list in this session. The authoritative full text supplied for US5915001A contains a "Prior art keywords" block (voice, information, page, web, caller — that is Google's concept extraction, not evidence of record) but no U.S. Patent Documents table, no Foreign Patent Documents table, and no Non-Patent Citations table. My attempts to pull the front page (uspto.report/patent/grant/5915001, the patentimages PDF, Espacenet biblio) returned the specification/abstract but not the citation tables before I hit the tool-call limit.

Accordingly, everything below is built from records I did retrieve, and I label each reference's provenance and date status. Where I could not verify a reference's critical date, I say so rather than assume it. Nothing here should be filed as an invalidity contention without pulling the '001's actual IDS and the full text of each reference.

Two further literal-reading notes: the supplied record contains OCR artifacts ("HIML" for HTML, "keyword" for keyboard, "informaton", "Useenables ide", "prefered", "requeste"), and it displays an "anticipated expiration" of 1996-11-14 — the filing date — which I flagged in the prior section as a metadata artifact (a 20-year term runs to 2016-11-14). I am not silently correcting these. Claim-and-spec wording should be verified against the USPTO copy before any construction depends on a spelling artifact.

1. Legal frame and the critical date

• The application was filed 1996-11-14 (US 08/748,943). It is therefore pre-AIA, and § 103 is governed by pre-AIA § 102's categories of prior art (a), (b), (e), (g), with Graham v. John Deere factors and KSR v. Teleflex permitting combinations of known elements with predictable results and rejecting a rigid TSM requirement.
• Critical date for § 102(b): 1995-11-14.
• § 102(e) is unusually important here. Several of the closest references issued after 1999-06-22 but were filed in 1996 before 1996-11-14. Under pre-AIA § 102(e) they are prior art as of their U.S. filing dates. Their publication dates are therefore a red herring — do not discard them.
• The priority-date discrepancy flagged earlier (1996-11-13 per Unified Patents vs. 1996-11-14 per Google/Espacenet) matters only if a reference falls in that one-day gap. I found none; I flag it for completeness.

2. The prior-art set I can ground (with dates as displayed)

Two references I must exclude or downgrade — do not auto-correct these into the record

• GB 2,307,619 A (Pollitt, "Internet information access system") — priority 1995-11-21, published 1997-05-28. A foreign (UK) publication is not § 102(e) art, and its publication postdates the 1996-11-14 filing. Unless it maps to an English-language PCT publication designating the US (which I did not verify), it is not § 102 prior art to the '001 at all. I flag this because it is frequently cited in the neighboring art and is easy to mis-use.
• US 5,819,220 A (Sarukkai et al., issued 1998-10-06) — appears in the WO 99/48088 ISR as category A,P. I did not verify its subject matter or filing date, so I will not map it to any element. (WO 99/48088 ISR)

Prior art the family attracted (useful as field evidence, not as anticipatory art)

• EP 1,073,036 A3 European search report lists US 5 915 001 A as category "A" against claims 1–55 — i.e., a later examiner treated the '001 as background art, not as a blocking reference. (EP1073036A3)
• KR 10-1027548 B1 (Motorola Mobility) lists US5915001 A in its "선행기술조사문헌" (prior-art search documents) for a voice-browser dialog enabler with markup-based voice dialog forms and downloaded speech grammars. (KR101027548B1)
• Numerous later patents cite it (e.g., US 6,606,611, US 7,127,403, US 7,266,181, US 7,457,876, US 9,369,545). This is consistent with it being a foundational background reference, which weakens any non-obviousness story rather than supporting one.

3. Element mapping — where the art already sits

The strongest structural point in an obviousness attack is the '001's own Background admissions. The specification concedes that conventional IVR systems already: allow a user to "place a telephone call into a system, navigate (generally using touch tone input) through a hierarchy of options in response to voice prompts and retrieve information stored in a computer database"; that "caller-specific information in the form of user preferences" could "minimize the size of a command tree"; and that "caller-specific speech training profiles could be used to implement speaker dependent speech recognition." Those are admitted, and the delta the '001 claims is essentially the URL-addressability / markup layer on top of them.

4. The combinations, and why a POSITA would make them

Ground A — Wise (US 5,884,262) + Herz (US 5,754,939): the core of claim 1

Rationale. Wise supplies everything except personalization: telephone-side rendering, hyperlink navigation, DTMF/voice command interpretation, and audio playback of a network document located by a link. Herz supplies the missing personalization half: generation and use of per-user profiles to customize which objects are presented.

Motivation / predictable result. The stated problem in the '001's own Background — "caller-specific information in the form of user preferences can aid in minimizing the size of a command tree which the user must navigate" — is precisely the problem Herz's profile-driven selection solves. Both are software/data-structure techniques operating on the same conventional network stack; combining them is "the mere arrangement of old elements... each performing the same function it had been known to perform." Under KSR, the combination yields the predictable result of fewer navigation steps, and the '001 itself frames this as a desirable goal, not a discovered result.

Ground B — Wise + Herz + Nat'l Semiconductor '736 (HTML telephone apparatus): the "markup voice form" element

Rationale. Claim 1's hardest element is that the query form is built by storing a markup-language form at a URL, playing it as prompts to the caller, and constructing the query from the collected answers. Wise teaches parsing and playing markup documents and invoking links, but is weaker on form capture. US 5,923,736 — an HTML-based telephone apparatus with priority 1996-04-02, squarely § 102(e) art — is directly on point for markup forms rendered on a telephone-form device.

Motivation / predictable result. Adding telephone form-fill to a telephone HTML browser is exactly the "familiar element... according to known methods" combination KSR endorses. The result — prompts in, answers out, server-side query — is the ordinary request/response semantics of HTML forms relocated to a voice channel, with no change in principle of operation.

Ground C — Wise + IBM '367 (and/or Bowater '270): the "voice publishing" cluster (relevant to claims 4 and 6)

Rationale. Claims 4 and 6 are the direction-of-linking variants of "collected caller input is written into a markup-language document that is hyperlinked into a web of documents." IBM '367 discloses dynamically generating HTML from voice-response-system data and hyperlinks invoking voice-response-system functions, plus password protection and over-the-WWW development/modification of voice applications — which is substantially the FIG. 9 "voice publishing form" concept. Bowater '270 adds document-per-action hyperlinking and passing parameters (including a password) through hyperlink navigation.

Motivation. Both references and Wise share the same architecture (telephony server ↔ HTTP server ↔ HTML). A POSITA building a voice web would plainly reuse the existing HTML hyperlink paradigm to make session output navigable, because that is the only linking machinery the Web stack provides. Direction of linking (parent→child vs. child→parent pointers) is a design choice with no new technical effect.

⚠️ Date risk: I did not verify the filing dates of US 6,052,367 or US 6,282,270. Their use in a § 103 ground depends entirely on their pre-1996-11-14 U.S. filing/§ 102(e) dates. Treat Ground C as a lead to check, not a finished ground.

Ground D — Claims 4/6 with a second-rung reference, and the speech-training/voiceprint subject matter

Because the previously generated summary notes that the independent claims I could retrieve center on profile-driven query customization and voice-form publishing, with the speaker-dependent training and voice-print material living in the specification (FIGS. 4–7), I will not manufacture claim-mapping for elements that may not be claimed. To the extent dependent claims recite URL-addressable, MIME-typed speech-training and voiceprint resources: Herz '939 (profile data organized for retrieval) + Wise '262 (documents and components addressed by URL and fetched on demand) + the MIME/HTTP standards the '001 expressly invokes (including the '001's cited HTML 2.0 / RFC 1866) make storing a per-user resource as a URL-addressable MIME object the default implementation choice, not an invention. Biometric voice-print matching and speaker-dependent recognition were themselves acknowledged as conventional in the '001 ("Test 611 uses conventional voice authentication methods"; "Personal voice web 300 uses conventional speaker dependent recognition").

Ground E — Claim 8 (white/yellow/order-page system)

Claim 8 reads on FIG. 10. Its elements are: a white-pages database, a query form, a first agent that searches it and generates a URL-addressable voice web page, a yellow-pages database, a second agent hyperlinked from that page, and a voice web browser. Murex '214 (directory lookup/intelligent call processing) supplies the directory-database lookup behavior used in telephony; Wise '262 / '736 supply voice-page generation and playback; the "home page as gateway to hyperlinked resources" pattern was already known in the art cited across the family. The gap a POSITA would need to bridge — linking a generated directory page to an advertising page to an order form — is the ordinary hypertext model applied to a known directory service, with a predictable commercial benefit (advertising upsell), which is itself a recognized KSR motivation.

5. Where an obviousness case is genuinely weak (state these, don't bury them)

1. Claim 1's "query form built from a markup-language voice form stored at a URL" is the narrow point. If the Office/PTAB reads this as requiring the form itself (not merely its rendered prompts) to be the URL-addressable, markup-encoded object that produces the query, then Grounds A/B need the reference to disclose form-at-URL, not merely page-at-URL. Wise '262's claim 1 as retrieved emphasizes link-retrieved documents parsed into segments, and does not, in the text I have, expressly recite form submission producing a database query. That is the factual pivot of the whole attack.
2. GB 2,307,619 is not prior art on its face (foreign publication after the filing date; no verified § 102(e) route). Any chart built on it is defective.
3. No verified § 102(e)/§ 102(b) date for US 5,953,392, US 6,052,367, or US 6,282,270 in this session.
4. No verified evidence on secondary considerations — I found no record evidence either way on long-felt need, failure of others, or commercial success attributable to the '001. Absence of such evidence is not proof of obviousness; it just means the Graham factor 4 record is empty.
5. The '001's category-A citations in EP 1,073,036 and its presence in Motorola's and others' prior-art lists cut both ways: they show the field regarded it as background, but they are also consistent with a genuinely early disclosure.

6. Practical posture

The patent is Expired – Lifetime, anticipated expiration 2016-11-14. That shapes the remedy, not the analysis:

• A § 103 challenge now would most plausibly arise defensively in the five D. Del. dockets listed earlier (1:20-cv-01010, 1:20-cv-01433, 1:20-cv-01435, 1:20-cv-01593, 1:22-cv-00105) or on the priority-linked continuations (US 6,400,806 B1; US 6,885,736 B2), not as a fresh infringement claim on the '001 — and, as flagged in the earlier sections, I could not verify the parties or whether the '001 was actually asserted in those cases.
• Ex parte reexamination remains available against an expired patent, but the practical value is limited with no enforceable term left; the real targets are the unexpired family members, which would require their own claim charts.
• No PTAB proceeding on the '001 was found; my only "5915001" hits in PTAB materials were prior-art citation strings in other parties' exhibits.

7. Bottom line

On the record I can ground, claim 1 is the most vulnerable and the least vulnerable at the same time: its preamble and most of its body are admitted or squarely disclosed by Wise US 5,884,262 (filed 1996-03-28), Herz US 5,754,939 (priority 1994-11-29), and US 5,923,736 (priority 1996-04-02) — all pre-filing § 102(e) art — but the specific "URL-stored markup voice form as the query generator" limitation is the one element for which I have not seen express disclosure. Claims 4 and 6 (session output written into a hyperlinked markup document) are the weakest claims, because the linking-direction distinction is a design choice and IBM/Bowater-family art teaches dynamically generating linked HTML from voice-system data — subject to date verification I could not complete. Claim 8's white/yellow/order chain is a straightforward application of known directory lookup plus known hypertext linking.

That is my analysis; I have deliberately not filled in the '001's missing IDS entries, the unverified reference dates, or the D. Del. case captions.