US 6823370

I'll verify details against live sources before summarizing.

Let me check the litigation history specifically and any 2026 appellate activity.

Identification & verification

I searched for the literal number 6823370. Two important identifier-collision caveats up front, per your strict rule:

• JP 6823370 B2 is an unrelated Japanese patent (Asahi Kasei / AKM, lens drive and calibration, issued 2021-02-03) — different document, different number system. Source: https://patentimages.storage.googleapis.com/29/ae/c6/d9d22f4d6dc7e6/JP6823370B2.pdf
• U.S. Trademark Registration Nos. 6823370 and 6823371 were inadvertently issued Aug. 23, 2022 and referred to the Commissioner for Trademarks (TTAB Opp. No. 91266109). These are trademarks, not patents. Source: https://ttabvue.uspto.gov/ttabvue/ttabvue-91266109-OPP-30.pdf

Neither is the patent you asked about. The U.S. patent is below.

US 6,823,370 B1 — bibliographic data

Primary source: https://patents.google.com/patent/US6823370/en

Abstract (verbatim): "An operator assisted system helps a caller browse the World Wide Web without requiring the caller use a computer. The invention enables a caller on a POTS or cellular/wireless telephone connection to connect with an operator to search and select Web-based content. The invention collects desired search results and passes them to an IVR (Interactive Voice Response) system for presentation to the POTS/Cellular caller. The invention facilitates high-quality search request interpretation and highly-efficient Web searches by a trained operator, using a visual PC based browser."

Independent claims in plain language

Claim 1 — System. Four cooperating elements: a workstation, an interactive voice response (IVR) module, an audio web server, and a switch. The workstation talks to the IVR, the audio web server, and the World Wide Web, and it contains means for selecting portions of Web pages from locations on the Web and storing those selections as files on the audio web server. The audio web server stores those files. The IVR generates audio signals corresponding to the stored files. The switch connects the telephone caller to the workstation, and later connects the IVR to the caller.

Claim 6 — Method. Four steps: (1) select one or more Web pages or portions of pages; (2) remove non-textual indicia from the selection to form selected files and store them on an audio server in accordance with caller identification data; (3) conduct an IVR session with the caller; (4) play one or more of the stored files to the caller according to the results of that IVR session.

Claim 8 — Computer workstation. A workstation with: means for receiving/transmitting telephone voice signals over a PSTN or a packet-switched data network; means for browsing the Web and selecting pages/portions (which may include both text and non-text indicia); means for removing non-text indicia to form a selected text file; means for storing that text file; and means for converting the stored text file to audio representative of the selected text and transmitting it to a telephone caller.

Dependent claims: 2–5 depend from claim 1 (2: switch-generated caller-ID data governs storage; 3: audio web server is a computer with memory storing selections by caller ID; 4: workstation software selects Web page portions and strips non-textual indicia; 5: workstation administered by a live operator). Claim 7 depends from 6 (selection performed by a live operator). Claims 9–12 depend from 8 (9: voice inquiry + output voice signal; 10: receive caller-ID signal and store accordingly; 11: software for selecting and de-texting; 12: browsing/selecting performed by a live operator).

Two literal-text issues worth flagging (not corrected):

• Claim 12 opens "The system of claim 8" although claim 8 is drawn to a computer workstation — an antecedent-basis/typographical inconsistency on the face of the claim set.
• Claims 8–11 are drafted in "means for" form without corresponding disclosed structure; the specification expressly says the selection/de-texting program's "construction and operation … are not … disclosed because one skilled in the art may implement such programs in a variety of ways." That is a classic §112(f) indefiniteness / written-description exposure.

The commercial core of the invention: a live operator uses a PC browser to find what the caller wants, highlights the relevant text, software strips graphics, the text is saved (as HDML/WML) on an audio Web server keyed to the caller's CLID, and the call is handed off to an IVR that reads it back via text-to-speech or WAV, with DTMF menus (back, forward, next, previous, exit, fax the source URL). Claim 5 and claim 7 and claim 12 each make the live-operator feature an element — which is the key distinction over the automated prior art (Wise, US 5,884,262; General Magic's Web-On-Call).

Litigation / docket check

One district court case is associated with this patent, and it is not from 2026:

• Vonage Holdings Corp. v. Nortel Networks Inc. and Nortel Networks Ltd., C.A. No. 1:07-cv-00507-GMS, D. Del. (Wilmington), filed August 17, 2007. Vonage sought a declaratory judgment on three Nortel patents (6,091,808; 6,445,695; 7,050,861). Nortel counterclaimed on those three plus nine others — the counterclaim expressly asserts infringement of "the '370 Patent" (U.S. Pat. No. 6,823,370) and alleges willfulness.
• The case was dismissed without prejudice on March 12, 2008 by stipulation, pursuant to a patent settlement agreement, each side bearing its own costs.

Sources: https://dockets.justia.com/docket/delaware/dedce/1:2007cv00507/38754; counterclaim doc. 28 (Dec. 14, 2007) and stipulation doc. 34 (Mar. 12, 2008) via https://www.findforms.com/single_form.php/form/97710/ and https://www.findforms.com/single_form.php/form/97703/; litigation link from https://portal.unifiedpatents.com/litigation/Delaware%20District%20Court/case/1%3A07-cv-00507.

CAFC 2026 dockets: I found no Federal Circuit 2026 docket entry, opinion, or appeal involving U.S. Patent 6,823,370. My searches returned unrelated 2026 Federal Circuit matters (e.g., Spero v. Volkswagen Group of America, Nos. 2025-1064 et al., decided May 14, 2026, concerning U.S. Pat. No. 10,894,503). I do not have authoritative confirmation of any pending 2026 appellate or PTAB proceeding on this patent, and given the status history below, live 2026 litigation would be surprising.

Status caveat — a genuine inconsistency in the record: The Google Patents status timeline shows "Expired – Fee Related" and an "Adjusted expiration" of 2020-05-02, while the Legal Events show a maintenance-fee lapse with "PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362" effective 2008-11-23. The two figures refer to different things: 2020-05-02 is the nominal adjusted statutory term (20 years from the 1999-10-18 parent filing, plus adjustment), whereas the patent actually ceased in 2008 for failure to pay maintenance fees. I report both literally rather than harmonizing them.

Uncertainty / method limits

• I did not query PACER, the CAFC CM/ECF docket, or USPTO PatentCenter/PTAB APIs directly; findings come from open web search over Google Patents, Justia Dockets, court-document mirror sites, and TTAB/PTAB PDFs. Absence of a 2026 CAFC hit is a negative search result and may reflect indexing limits rather than certainty that no such docket exists.
• Google Patents labels the current assignee "Nortel Networks Ltd"; other citation listings render it "Nortel Networks Limited." I could not authoritatively reconcile the entity-name variants, though Nortel's assignment/terminal disclaimers are not in the material available to me.
• Everything above the "Litigation / docket check" heading is taken from the authoritative full text you supplied and the Google Patents page; the prosecution-history, ownership-chain (e.g., any Nortel bankruptcy-era transfer of this asset), and any reexamination/reexam certificate details are not confirmed here.

I'll research litigation involving US patent 6823370. Let me start with the lead already present in the patent record (a Delaware District Court case) and search litigation databases.

The first result is revealing but needs verification — the Delaware case appears tied to different patents. Let me dig deeper.

Findings: Litigation involving US Patent 6,823,370

1. What the patent's own record shows (primary lead)

The Google Patents record for US 6,823,370 (fetched from https://patents.google.com/patent/US6823370/en) lists exactly one litigation entry:

Following that case number to the underlying docket (Docket Alarm / Justia) resolves it to:

2. Important discrepancy you should not overlook (attorney's caution)

I was not able to confirm from the docket that US 6,823,370 was actually asserted or challenged in 1:07-cv-00507.

• The AO 120 "Report on the Filing of an Action Regarding a Patent" filed in that case (Document 4, filed Aug. 20, 2007) lists only three patents:
  • 6,091,808 (Nortel Networks Corporation)
  • 6,445,695 B1 (Nortel Networks Limited)
  • 7,050,861 B1 (Nortel Networks Limited)
• The Vonage–Nortel settlement agreement's "Exhibit A – Nortel Patents" likewise lists only those same three patents.
• None of those three is 6,823,370, and 6,823,370 does not appear on either the AO 120 report or the settlement exhibit I could retrieve.

So there are two possibilities: (a) 6,823,370 was among the broader set of Nortel patents at issue in the related Nortel/Vonage dispute (press coverage referenced a Nortel suit alleging Vonage infringed "12 of Nortel's patents") but was not carried on the formal AO 120/settlement lists; or (b) the Google Patents/Unified Patents database has mis-associated the '370 patent with case 1:07-cv-00507. On the evidence I retrieved, I cannot resolve which — and given the instruction to treat search results as ground truth and not to fabricate, I flag this rather than assert it.

3. Other proceedings

• No PTAB proceedings (IPR/PGR/CBM) involving US 6,823,370 were found in the search results.
• No CAFC appeal involving US 6,823,370 was found.
• No other district-court case specifically identifying US 6,823,370 was found.

4. Literal-identifier note (avoiding false positives)

Per the rule to interpret identifiers literally, the following are different rights, not US Patent 6,823,370, and were excluded:

• JP 6823370 B2 (Japanese patent, Asahi Kasei Microdevices — lens-actuator drive device, published 2021).
• US Trademark Registration No. 6823370 (inadvertently issued, referenced in TTAB Opposition No. 91266109) — a trademark, not a patent.

5. Practical/status context

US 6,823,370 (filed July 10, 2002, as a continuation of Ser. No. 09/420,154 filed Oct. 18, 1999; granted Nov. 23, 2004; assignee Nortel Networks Ltd.) lapsed for failure to pay maintenance fees, effective November 23, 2008 (Patent Expired Due to Nonpayment of Maintenance Fees, 37 CFR 1.362; legal status "Expired – Fee Related," adjusted expiration 2020‑05‑02). This is consistent with the short, early-terminated life of any enforcement activity.

Bottom line

• One litigation is associated with US 6,823,370 in the patent record: Vonage Holdings Corp. v. Nortel Networks Inc. et al, No. 1:07-cv-00507 (D. Del.), filed August 17, 2007, a declaratory-judgment action that settled in 2008. However, the case's own AO 120 report and settlement exhibit list three different Nortel patents, so whether the '370 patent was truly at issue in that case is unconfirmed and should be verified against PACER (D. Del. 1:07-cv-00507) and the full complaint/answer before relying on it.

Caveat on completeness: I was unable to complete a full sweep of the Unified Patents litigation caselist, PACER, and the CAFC docket before reaching my search limit. A definitive "no other litigation exists" statement therefore cannot be made with high confidence; a direct PACER/CAFC and Unified Patents portal query on "6823370" is recommended to close that gap.

I'll verify the structured data with live searches before writing anything.

Let me run a couple more targeted checks to make the negative finding defensible.

PTAB / AIA Trial Analysis — U.S. Patent No. 6,823,370 B1

"System and method for retrieving select web content" — Kredo et al., assignee Nortel Networks Ltd.

Proceedings overview

There are zero AIA trial proceedings on file against US 6,823,370 B1: 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denials. The structured PTAB-proceedings block from the USPTO Open Data Portal (the canonical list provided in this prompt) is empty for this patent, and my independent web searches surfaced no IPR, PGR, or CBM petition, institution decision, Final Written Decision, or appeal. The defensive posture this gives a defendant is not "the patent survived IPRs and is hardened," and it is certainly not "the claims are canceled" — it is that the patent was never tested at the PTAB because it lapsed for non-payment of maintenance fees on 2008-11-23, roughly fourteen months after the only litigation that ever touched it, and its full statutory term (adjusted expiration 2020-05-02) has since run. The absence of PTAB activity here is not a sign of a robust patent; it is a sign of an abandoned one.

Proceedings

None. There is no proceeding to report at § 1, claim-level granularity, because no petition was ever filed.

I will not manufacture a docket. For the record, the searches I ran (querying the patent number alone and in combination with "IPR," "PTAB," "inter partes review," and "Nortel") returned no PTAB document referencing US 6,823,370. The only hits on the numeral string "6823370" were the two unrelated identifier collisions already flagged in the identification section:

• JP 6823370 B2 — Japanese patent, lens-actuator drive/calibration (Asahi Kasei Microdevices), published 2021-02-03. Not a PTAB matter. https://patentimages.storage.googleapis.com/29/ae/c6/d9d22f4d6dc7e6/JP6823370B2.pdf
• U.S. Trademark Registration No. 6823370 — inadvertently issued 2022-08-23, referred to the Commissioner for Trademarks in TTAB Opposition No. 91266109 (Therapy Brands Holdings LLC v. TherapyNotes, LLC). That is a TTAB matter about a trademark, not an AIA trial about this patent. https://ttabvue.uspto.gov/ttabvue/ttabvue-91266109-OPP-30.pdf

Neither is an AIA trial proceeding on US 6,823,370, and neither supports an inference of one.

Strategic summary

Claim status: all twelve claims are UNTESTED and, as a practical matter, unreachable. Claims 1–12 of US 6,823,370 have never been canceled, narrowed, or adjudicated by the PTAB. No independent claim (1, 6, 8) and no dependent claim (2–5, 7, 9–12) has a PTAB outcome. That sounds like good news for a patent owner until you pair it with the status history: per the patent's own legal-events record, the patent expired due to nonpayment of maintenance fees under 37 C.F.R. § 1.362, effective 2008-11-23, and the nominal adjusted statutory term of 2020-05-02 has also passed. A defendant facing a demand letter today therefore does not need a PTAB outcome — there is no live patent to infringe for any post-2008 conduct (and, on a § 286 damages theory, no more than six years back from any hypothetical suit, which likewise reaches only pre-lapse activity). The absence of PTAB activity is explained by lapse, not by invulnerability.

Estoppel landscape: empty, and in the defendant's favor. Because no petitioner ever filed, no one is subject to the § 315(e)(2) estoppel bar on grounds raised or reasonably raisable, and there is no IPR-driven prior-art record to work around. If a defendant were nonetheless forced to litigate invalidity (for example, over a pre-2009 damages period in some revived dispute), the entire prior-art field remains open: the thirteen references of record (Wise/US 5,884,262, Kanoh/US 5,873,077, Dekelbaum/US 5,838,682, Noonen/US 5,761,280, Kikinis/US 5,727,159 and US 5,875,436, plus General Magic's Web-On-Call™ product), and any additional art, can be asserted in district court without any § 315 estoppel constraint, and without the PTAB's narrower Phillips-adjacent claim-construction posture. The patent's own specification concedes a § 112 soft spot that is worth developing as a fallback: it states the selection/de-texting program's "construction and operation … are not … disclosed because one skilled in the art may implement such programs in a variety of ways," while claims 8–11 are drafted in "means for" form — a written-description / § 112(f) indefiniteness argument that never got an airing anywhere because no AIA trial was ever filed.

Pattern signals: none of the usual markers. There is no repeat petitioner, no serial IPR campaign, no patent-owner appeal to the Federal Circuit, and no defensive aggregator in the petitioner column. Note carefully that Unified Patents appears in this patent's record only as the source of litigation data for the Delaware case entry — not as a petitioner. It filed nothing. The absence of an aggregator IPR is itself consistent with the patent's early lapse: defensive aggregators target asserted, live patents, and this one was neither after 2008.

Recommended next steps

• If you are a defendant receiving a demand letter citing US 6,823,370: the decisive fact is not a PTAB FWD — there isn't one to link. It is the maintenance-fee lapse. The patent's legal-events record states "PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362," effective 2008-11-23 (Google Patents, https://patents.google.com/patent/US6823370/en). Ask the asserting party to identify (a) the allegedly infringing acts and (b) their dates. Any demand framed on post-2008-11-23 conduct, or on sales within six years of today (2026-09-28), is facially baseless.
• Do not assume a PTAB shortcut exists. There is no FWD to cite and no claims previously canceled by the Board. If you need an invalidity record, it must be built in district court (or in a new AIA petition, if a live controversy over pre-lapse conduct were ever to justify one — though the Board has no reason to institute on an expired patent over long-past damages).
• If instead you are evaluating the patent historically (portfolio, licensing diligence, or the Nortel estate): treat it as expired, never-PTAB-challenged, and adjudicated only tangentially. The one associated court case, Vonage Holdings Corp. v. Nortel Networks Inc. et al., No. 1:07-cv-00507 (D. Del., filed 2007-08-17), settled in March 2008 and ended without a merits ruling; and as flagged in the prior section, the AO 120 report and settlement exhibit in that case list three different Nortel patents, so even that single litigation link to the '370 is unconfirmed. Verify via PACER before relying on it.
• Searches to run to close the gap (I was not able to complete these before reaching my search limit, and the negative finding should be treated as a negative search result, not proof of nonexistence): a direct query of PTAB E2E (https://ptab.uspto.gov/) and the ODP AIA-trial API on "6823370"; the Unified Patents litigation portal for any proceeding not surfaced in open results; and a CourtListener / CAFC docket sweep for any appeal. Given the 2008 lapse, I would be surprised by a hit — but the standard for a "no proceedings" statement is a completed portal query, which I have not executed directly.

Constraint compliance note: No proceeding number is invented; the count is zero and is reported as zero. No FWD is quoted because none exists. Because there is no opinion to cite, no link to a specific PTAB or CourtListener decision is offered — the appropriate destinations are the searchable portals above. Dates are in YYYY-MM-DD throughout. Where the record is genuinely contradictory (the 2008-11-23 lapse versus the 2020-05-02 adjusted expiration in the prior section), both are reported literally rather than harmonized.

I'll research the assignment chain for US 6,823,370 before writing anything.

Let me try the USPTO assignment databases directly and search for any Nortel-era transfers.

Let me check the assignment records and the original assignee/employment context.

I'll verify each cited reference against live sources before analyzing § 102 exposure. Let me run the first batch of lookups.

I've hit my tool-call limit, so I'll synthesize now. The patent's own full text (supplied) is authoritative for the citation list and dates; the searches confirmed the substance of the four most anticipation-relevant references (Wise '262, Imielinski '448, Bellcore '789, General Magic '650) and Kikinis '159.

Prior-Art Analysis — U.S. Patent No. 6,823,370 B1

"System and method for retrieving select web content" — Kredo, Kohl & Knight; Nortel Networks Ltd.

0. Scope, method, and the "6823370" search itself

Identifier discipline. A USPTO/Google Patents query on the literal string 6823370 returns exactly one U.S. patent: US 6,823,370 B1 (https://patents.google.com/patent/US6823370/en). Two non-patent collisions with the same numeral string recur in any such search and are excluded here, consistent with the earlier sections of this analysis: JP 6823370 B2 (Asahi Kasei Microdevices lens-actuator patent) and U.S. Trademark Registration No. 6823370 (TTAB Opp. No. 91266109). Neighboring patent numbers (e.g., 6,823,046, 6,823,046-family) were not substituted.

Method. Every reference analysed below is drawn from the front-page "Patent Citations (13)" list of US 6,823,370 itself, as reproduced in the authoritative full text. I verified bibliographic data and technical substance for the highest-relevance references against Google Patents, USPTO publication mirrors, and the printed specifications; references I could not independently re-verify in this session are marked [record-only] and are described on the basis of the patent's own citation table plus the specification's own prior-art discussion.

Note on the citation-list markers. Google's legend is "* Cited by examiner, † Cited by third party." Seven of the thirteen carry the asterisk; six do not. Tellingly, the six unmarked references (Wise '262, Kanoh '3077, Dekelbaum '8682, Noonen '1280, Kikinis '159 and Kikinis/'5436) are exactly the six the applicant discusses by name in the specification's "Discussion of Prior Art." That is consistent with those six being applicant-submitted IDS art, and the seven asterisked items being examiner-cited. I flag this rather than assert it, because the Google marker convention is not self-documenting.

1. Legal framework (which § 102 doors are even open)

Consequence up front: two of the thirteen cited references — US 6,643,621 (Cisco) and US 6,621,502 (Awa) — were filed after 1999-10-18 (2000-09-14 and 2001-05-02 respectively). Against claims entitled to the 1999-10-18 priority, they cannot be § 102(a), (b) or (e) art at all. They are discussed in § 4 below as "date-barred."

2. Inventory — the thirteen cited references at a glance

Non-patent art of record (from the specification itself): General Magic, Inc.'s Web-On-Call™ product — a software product installed in a Web server letting a telephone user browse the Web automatically. Under pre-AIA § 102(a)/(b) a prior public product/printed description can anticipate, but the specification only characterises it functionally and identifies no enabling printed publication, so its evidentiary value is thin.

3. Reference-by-reference § 102 analysis

3.1 US 5,884,262 — Wise et al. (Bell Atlantic) — the strongest § 102 candidate

• Full citation: U.S. Pat. No. 5,884,262, "Computer network audio access and conversion system," filed 1996-03-28 (App. 623,103), issued 1999-03-16, assignee Bell Atlantic Network Services, Inc. https://patents.google.com/patent/US5884262A/en
• Disclosure (verified): A caller dials a number and requests information by DTMF or voice; the system analyses the request, connects to the target network, and retrieves the requested document in a standard format such as HTML; a Parser analyses document structure, selects what portions of the document are converted to speech, and passes the text through a text-to-speech engine into audio files played by an audio file player ("Call Manager" architecture). Hyper-text lists are read as audio menus; the caller navigates with DTMF ("press 1 to select a hyper-text link") or voice. AIN caller-ID features can “allow access to individual user profiles using caller identification (ID) information … [and] to designate a home page for an individual user.”
• Potential § 102 mapping (as a single reference):

• Assessment: Wise is the only cited reference that maps onto all the automated limitations of the independent claims. It does not, however, disclose a separate "audio web server" as a distinct networked node storing the operator-selected files keyed to CLID, and it cannot touch the live-operator claims. Expect it to be run as a § 103 primary reference, not a clean § 102 reference, against claims 1/6/8.

3.2 US 6,240,448 B1 — Imielinski & Virmani (Rutgers) — the strongest structural "audio web server" reference

• Full citation: U.S. Pat. No. 6,240,448 B1, "Method and system for audio access to information in a wide area computer network," inventors Tomasz Imielinski & Aashu Virmani, assignee Rutgers, The State University of New Jersey; PCT filed 1996-12-20 (PCT/US96/20409), WO 97/23973 published 1997-07-03, U.S. § 371(c) date 1999-05-07, provisional 60/009,153 filed 1995-12-22, issued 2001-05-29. https://patents.google.com/patent/US6240448
• Disclosure (verified): Describes an "Audio Web" of "audio enabled pages" authored in an "audio text manipulation language" (ATML) — pages that link particular text data which can be from conventional Web pages. "Audio enabled pages are stored at an audio web server." An audio web server provides text-to-speech conversion and call-processing, with DTMF keypad and speech input, retrieves pages by URL over http, and "translates the audio enabled pages into audio." The specification expressly frames this as an improvement on "existing approaches [that] involve an automatic translation of html documents into audio" — i.e., it positions itself squarely in the '370's problem space.
• The tell: this reference uses the phrase "audio web server" verbatim — the exact term in claims 1 and 6 of the '370.
• Potential § 102 mapping:

• Assessment: the most dangerous single reference for claim 1, because of the literal "audio web server" + "select Web text into audio pages stored on that server" combination. Its weaknesses are (i) the absence of a live operator and (ii) the absence of CLID-keyed, per-call storage.

3.3 US 6,366,650 B1 — Rhie et al. (General Magic) — the Web-On-Call™ patent

• Full citation: U.S. Pat. No. 6,366,650 B1, "Method and apparatus for telephonically accessing and navigating the internet," inventors Kyung H. Rhie, Richard J. Kwan, Lee E. Olsen, John S. Hahn; assignee General Magic, Inc., Sunnyvale CA (later Intellectual Ventures I LLC); App. 09/205,269 filed 1998-12-04, a continuation of App. 08/609,699 filed 1996-03-01 (issued as US 5,953,392); granted 2002-04-02.

Literal-text flag: one USPTO-derived listing renders the parent as "now 5,953,369," while Google Patents and FreePatentsOnline render it "5,953,392." I do not auto-correct; the parent's own family listing shows US 5,953,392. This is a citation-transcription discrepancy in a secondary source, not in the '370 itself.

• Disclosure (verified): A "Voice Browser" with an "HTree Generator" that accesses the Web and converts web-page text to speech; the "Call Menu Script Interpreter" drives voice prompts, plays Web-page voice segments, and accepts DTMF from the caller's keypad; hyperlinks are signalled audibly; documents can also be delivered by fax, e-mail or mail. https://patents.google.com/patent/US6366650
• Potential § 102 mapping:

• Assessment: this is the patent behind the Web-On-Call™ product the '370's specification attacks. It is the natural § 103 partner for Wise and Imielinski, and a serious § 102 candidate against claims 8 and 6.

3.4 US 5,926,789 — Barbará et al. (Bellcore) — the "Audio Web" reference

• Full citation: U.S. Pat. No. 5,926,789, "Audio-based wide area information system," filed 1996-12-19, issued 1999-07-20, assignee Bell Communications Research, Inc. (Bellcore); inventor of record Daniel Barbará (George Mason University faculty page lists it: "Audio-based wide area information system, US5926789, granted July 20, 1999"). https://patents.google.com/patent/US5926789A
• Disclosure (verified): A "virtual audio client" used to navigate the "Audio Web"; commands input by telephone keypad (mapped to STOP/PLAY/FORWARD/BACK/PRESET/RETURN/GO) or by speech; hierarchical "directory pages" of menu choices delivered audibly; "hyperaudio links" analogous to hypertext links; bookmarking of audio pages; a "Digital HyperRadio system."
• Potential § 102 mapping: potentially relevant to claim 6(d) ("playing … the selected files to the caller in accordance with the results of the interactive voice response session") and to the general IVR-navigation concept of claims 1 and 8. However, the '789 system operates on a purpose-built corpus of pre-authored audio pages; it does not disclose "selecting portions of pages from locations on the World Wide Web" nor "removing non-textual indicia from the selected pages." Those are express limitations of claims 1 and 8 and step (2) of claim 6. It therefore cannot anticipate any independent claim; it is background/§ 103 material only. It cannot touch claims 5, 7 or 12.

3.5 US 5,727,159 — Kikinis (the "proxy server for low-power portables" patent)

• Full citation: U.S. Pat. No. 5,727,159, "System in which a Proxy-Server translates information received from the Internet into a form/format readily usable by low power portable computers," filed 1996-04-10 (App. 08/629,475; PCT/US97/05545, WO 97/38389 published 1997-10-16), issued 1998-03-10. https://patents.google.com/patent/US5727159A/en
• Disclosure (verified): A Proxy-Server that browses the WWW on behalf of a low-end/battery-powered field computer, reduces information density, and transposes downloaded Web pages to match the client display via a TCP/IP link; caching; display-size/format tailoring.
• Potential § 102 mapping: None of claims 1–12. There is no telephone interface, no IVR, no audio conversion, and no caller. The '370's specification says as much: "It does not address any telephone interfaces." It is § 102 art in the abstract (it qualifies under § 102(b)) but discloses no claimed limitation beyond generic Web-page retrieval and format conversion. Useful, if at all, as § 103 support showing that selecting/reformatting Web content for a constrained client was known.

3.6 US 5,761,280 — Noonen et al. (8×8, Inc.)

• Full citation: U.S. Pat. No. 5,761,280, "Telephone web browser arrangement and method," filed 1996-09-04, issued 1998-06-02, assignee 8×8, Inc. [record-only — description from the '370's own citation table and specification]
• Disclosure (per the '370's specification): Web browsing using telephone DTMF inputs, a menu system, and a display system attached to the telephone. The specification expressly notes: "It does not address the use of audio speech-to-text or text-to-speech."
• Potential § 102 mapping: fails the audio-translation heart of every independent claim — claim 1 ("generating audio signals corresponding to the files"), claim 6 ("playing … the selected files to the caller"), claim 8 ("converting the stored text file to audio signals … transmitting"). It could supply the "switch connects the caller to the workstation" and "browsing/selecting" elements of claims 1 and 8, but cannot anticipate any claim as a whole. § 103 material at best.

3.7 US 5,838,682 — Dekelbaum et al. (Bell Atlantic) — dual-link secure transaction system

• Full citation: U.S. Pat. No. 5,838,682, "Method and apparatus for establishing communications with a remote node on a switched network based on hypertext dialing information received from a packet network," filed 1995-11-28, issued 1998-11-17. [record-only]
• Potential § 102 mapping: None. The specification concedes the point: it "does not address customer-driven browsing and searching the Web, independently of a merchant or sales entity." Its dual-link (PSTN + data network) architecture is at most § 103 support for claim 8's "public switched telephone network or a packet switched data network" alternative wording — but that wording is a disjunctive of alternatives, and Dekelbaum shows the two-network combination, not the claimed single-workstation telephony interface. No anticipation of any claim; no bearing on claims 5, 7, 12.

3.8 US 5,873,077 — Kanoh et al. (Ricoh) — fax-based Web retrieval

• Full citation: U.S. Pat. No. 5,873,077, "Method and apparatus for searching for and retrieving documents using a facsimile machine," filed 1995-01-13, issued 1999-02-16. [record-only]
• Potential § 102 mapping: the specification says it discloses "a fax-based Web-access method … the Website scans the user's faxes to select a course of action or a collection of data to return via a fax" and that it "does not describe any voice access or any non-fax telephone usage." It arguably touches claim 6's "selecting one or more Web pages or portions" (a human's fax choices drive the selection) and claim 8's "selecting … portions of pages," but it discloses no audio translation and no IVR session — fatal to every independent claim. Cannot anticipate. No bearing on 5, 7, 12.

3.9 US 5,875,436 — Kikinis / Data Link Systems — "Virtual transcription system"

• Full citation: U.S. Pat. No. 5,875,436, "Virtual transcription system," filed 1996-08-27, issued 1999-02-23, Data Link Systems, Inc. [record-only] The '370's specification describes it as a "work-order transcription and communication system using the Internet" and expressly states it "does not address Web browsing at all."
• Flag (entity discrepancy, not auto-corrected): the specification attributes this reference to Kikinis, while the front-page citation table lists Data Link Systems, Inc. as the entity. These are not necessarily inconsistent (inventor vs. assignee), and Kikinis's patent portfolio does include Data Link Systems assignments, but I could not verify the inventor of record for '436 in this session. I report both renderings literally.
• Potential § 102 mapping: None. Not directed to Web browsing, audio Web access, or operator assistance. Qualifies as § 102(a)/(e) art formally; discloses no claimed limitation.

3.10 US 5,799,063 — Talk Web Inc.

• Full citation: U.S. Pat. No. 5,799,063, "Communication system and method of providing access to pre-recorded audio messages via the Internet," filed 1996-08-15, issued 1998-08-25, Talk Web Inc. [record-only] — the only cited reference to qualify under § 102(b).
• Potential § 102 mapping: potentially relevant to the "playing … the selected files to the caller" limitation of claims 1, 6(d) and 8 — pre-recorded audio messages delivered to a caller from Internet-side storage is the same delivery primitive. But it is a system of pre-recorded messages, not operator-selected Web page text converted to audio; it therefore fails "selecting portions of pages from locations on the World Wide Web" and "removing non-textual indicia," and cannot anticipate any independent claim. § 103 background only. No bearing on claims 5, 7, 12.

3.11 US 6,298,129 B1 — MCI Communications

• Full citation: U.S. Pat. No. 6,298,129 B1, "Teleconference recording and playback system and associated method," filed 1998-03-11, issued 2001-10-02. [record-only] § 102(e) art as of 1998-03-11.
• Potential § 102 mapping: None on point. Audio recording/playback to callers implicates only the generic "playing audio to a caller" idea already disclosed better by Wise or Imielinski. No Web-page selection, no text stripping, no operator. Cannot anticipate any claim.

3.12 US 6,643,621 B1 — Cisco Technology — date-barred

• Full citation: U.S. Pat. No. 6,643,621 B1, "Methods and apparatus for referencing and processing audio information," filed 2000-09-14, issued 2003-11-04. [record-only]
• § 102 status: Not prior art against claims entitled to the 1999-10-18 priority — it was filed after that date, so it is unavailable under § 102(a), (b) and (e). It would become § 102(e) art only for a claim of the '370 that is not supported by the 1999-10-18 parent disclosure and therefore takes the 2002-07-10 filing date as its effective date (a contingency that the § 112 concerns noted in the earlier sections make worth keeping in reserve, and that is also § 103-relevant on that alternative date). It cannot anticipate any claim on the correct priority date.

3.13 US 6,621,502 B1 — Awa, Inc. — date-barred

• Full citation: U.S. Pat. No. 6,621,502 B1, "Method and system for decoupled audio and video presentation," filed 2001-05-02, issued 2003-09-16. [record-only]
• § 102 status: Not prior art for the same reason — filed about 19 months after the 1999-10-18 priority date. Same contingency caveat as § 3.12. Cannot anticipate any claim.

4. Consolidated § 102 verdict table

Bottom line on anticipation: No single cited reference anticipates the claims as issued. The only references with genuine single-reference anticipation exposure are Wise '262, Imielinski '448 and General Magic '650, and even they cannot reach claims 5, 7 and 12, each of which recites that the workstation/browsing/selection is administered by a live operator — the load-bearing limitation of the whole patent. The three references survive as a formidable § 103 combination (Wise alone, or Imielinski '448 as primary with Wise '262 / General Magic '650 as secondary, teaching every automated element: telephone access, Web-page selection, stripping of non-text, text-to-speech, caller-ID profiles, DTMF navigation, even fax delivery of the source URL).

5. Flags, anomalies, and confidence limits

1. Two "cited" references post-date the patent's own priority date. US 6,643,621 (filed 2000-09-14) and US 6,621,502 (filed 2001-05-02) cannot be § 102 art against claims entitled to 1999-10-18. Their presence in the front-page citation table (and, per Google's marker convention, in the "cited by examiner" set) is anomalous. This is worth reconciling against the file history/pair of the ‘370, because it bears directly on whether the examiner treated the claims as having the 1999-10-18 effective date.
2. The specification concedes the key software is undisclosed. It states the selection/de-texting program's "construction and operation … are not … disclosed because one skilled in the art may implement such programs in a variety of ways," while claims 8–11 are in "means for" form. That is a § 112(f) written-description exposure that never received an airing (no PTAB proceeding, single 2007–2008 district case that settled). It also means the § 102/§ 103 art has to be judged against broadly functional claim language.
3. Claim 12 recites "The system of claim 8" although claim 8 is drawn to a computer workstation — an antecedent-basis inconsistency on the face of the claims, reported literally and not corrected.
4. Secondary-source transcription discrepancy: the General Magic parent is rendered "5,953,369" in one USPTO-derived listing and "5,953,392" elsewhere; the family listing supports 5,953,392. Not an error in the '370 itself.
5. Entity discrepancy for US 5,875,436: specification calls it Kikinis; the citation table names Data Link Systems, Inc. Reported both ways; inventor of record not independently verified.
6. Verification coverage. I verified substance and dates from primary/mirror sources for Wise '262, Imielinski '448, Barbará '789, General Magic '650 and Kikinis '159. The remaining eight ([record-only]: '1280, '9063, '8682, '3077, '5436, '298129, '6443621, '621502) are described from the '370's own front-page citation table plus the specification's own prior-art narrative — the authoritative sources supplied — and were not independently re-pulled in this session.
7. Statutory/IPR context (cross-reference, not repeated): consistent with the earlier sections, the patent lapsed for non-payment of maintenance fees effective 2008-11-23, and no AIA trial proceeding exists — so the only § 102 record on this patent is the prosecution-time citation list analysed above.
8. Date coordination: the task header states "Current Date: April 26, 2026" while the operating framework gives 2026-09-28. This analysis uses the latter. All dates are YYYY-MM-DD.

I'll verify the cited prior-art references before building the § 103 combinations, since my analysis must rest on what those references actually disclose rather than on titles alone.

Good — I now have the substance of Wise, Kikinis, Talk Web, and Bellcore. Let me pull the remaining cited references and check the prosecution history for any § 103 rejections and applicant arguments.

Obviousness Analysis Under 35 U.S.C. § 103 — U.S. Patent No. 6,823,370 B1

"System and method for retrieving select web content" — Kredo, Kohl & Knight; Nortel Networks Ltd.
Priority date 1999-10-18 · Filed 2002-07-10 · Granted 2004-11-23 · Expired (fee lapse) 2008-11-23

0. Scope, standard, and a threshold caveat

This section builds on — and does not repeat — the identification, litigation, and PTAB sections above. It uses the 13 references of record in the patent's own "Citations" list, supplemented by the applicant's characterizations of them in the Background and by the full-text excerpts I retrieved. It applies the Graham v. John Deere, 383 U.S. 1 (1966) framework (scope/content of prior art; differences; PHOSITA level; secondary considerations) and the KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) rationales, with MPEP § 2143 as the operative examiner's checklist.

Caveat on completeness of the record. All 13 citations appear to be examiner citations (Google's list marks third-party citations with "†"; none are so marked). I was not able to complete a search of the file wrapper for a § 103 rejection and any applicant traverse, so I cannot tell you whether the examiner applied a combination and the applicant argued it away. That matters: the fact that the patent issued over individually-considered references does not immunize the claims from a § 103 combination, because issuance is consistent with either (a) no combination having been articulated, or (b) the applicant having distinguished individual references one at a time. A KSR-style combination challenge is a separate question. I flag this as an open verification item, not a finding.

One more caveat. Two of the 13 cited references are, on the dates of record, not prior art at all against this patent — see § 2.3. I do not use them in any ground.

1. The claims, decomposed into limitations

Reproduced in full in the earlier summary; here only the limitation set that any rejection must meet:

Dependents add only: caller-ID-keyed storage (2, 3, 10); a software program for selecting + de-texting (4, 11); live-operator administration/performance (5, 7, 12); voice inquiry + output voice signal (9).

The entire delta over the art reduces to three things: (i) a human operator in the select-and-store loop; (ii) an intermediate audio server that holds human-curated page fragments keyed to a caller; and (iii) the handoff of the live call from the operator to an IVR. Nothing else in the claim set is new; every remaining element is a bare-bones telephony/IVR component.

2. Prior art: scope, content, and status

2.1 The references that carry the analysis

2.2 What the references do not disclose

Nothing in the set discloses a live human operator choosing which fragment of a Web page a specific caller should hear, nor the operator→IVR live-call handoff. Those are the only true gaps.

2.3 References that are not § 102 prior art (flagging, not auto-correcting)

Both were nonetheless cited. Their presence in the "Cited By" list of later patents does not convert them into prior art. If an examiner's rejection had rested on either, that rejection would be improper on its face.

The remaining eleven references are § 102(a)/(b) art (those published >1 yr before 1999-10-18: Kikinis '159, Noonen '280, Talk Web '063, Dekelbaum '682, Kanoh '077, Kikinis '436) or § 102(e) art (US patents whose applications predate 1999-10-18: Wise '262 (filed 1996-03-28), Bellcore '789 (filed 1996-12-19), Rutgers '448 (filed 1995-12-22), MCI '129 (filed 1998-03-11), General Magic '650 (filed 1996-03-01)).

3. Person having ordinary skill in the art (PHOSITA)

A bachelor's degree in electrical engineering or computer science (or equivalent), with 2–4 years of experience in one or more of: telephony switching/ACD systems, interactive voice response design, or Web/HTTP application development — sufficient to understand call routing, caller-ID delivery, TTS, DTMF menus, and HTTP/markup-based content. This is a low-to-moderate skill level. At this level, the radio/telephony "operator assisted" idiom (directory assistance, live-agent call centers, the operator-to-automation handoff) was universally familiar, and the markup-page constructs (Rutgers' ATML, Bellcore's HAML, General Magic's pagelets, HDML/WML) were an active, crowded design space.

4. The core § 103 finding

The asserted advance — "facilitates high-quality search request interpretation and highly-efficient Web searches by a trained operator … while utilizing the existing efficiencies of automated announcement IVR systems" (patent Summary) — is on its face the reason a PHOSITA would make the substitution. The specification supplies the motivation itself:

1. The Background attacks the automated alternative at length (Wise's ASR error rates, database-query skill requirements, PSTN band-limiting). A PHOSITA reading those criticisms is being told why to replace the automated front end.
2. Rutgers '448 independently confirms the same deficiency ("this process cannot be fully automated," cannot handle tables/databases) — an express statement in the prior art that full automation had failed, which is a teaching toward a human-in-the-loop solution, not away from it.
3. Kikinis '436 discloses the human-in-the-loop network processing pattern — a server assigning content-interpretation jobs to human technicians and receiving their transcribed output back.
4. Wise '262 already discloses nearly all of the machine plumbing (call-in, retrieve HTML, parse/select portions, TTS, DTMF navigation, caller-ID-driven profiles).
5. Rutgers '448 and Talk Web '063 already disclose the intermediate audio server + telephone client architecture; Kikinis '159 already discloses transform-and-store-for-a-limited-client.

Under KSR, "when a patent claims a structure already known in the prior art that is altered by the mere substitution of one element for another known in the field, the combination must do more than yield a predictable result" and "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." Substituting a trained operator for Wise's speech recognizer / Rutgers' automated authoring step is precisely such a substitution, and the result (better interpretation accuracy at the cost of a human agent) is exactly the predictable result the applicant claims.

5. Grounds of rejection

Ground A (strongest) — Wise '262 + Rutgers '448 + Kikinis '159, in view of Kikinis '436

Applied to claim 1:

Motivation to combine: Rutgers expressly teaches that automated HTML→audio "cannot be fully automated" and fails on numeric/tabular data; Wise's field and the applicant's own Background confirm the ASR/skill problems; Kikinis '436 shows the server-dispatch-to-human-technician solution. Combining these to obtain accurate interpretation of a specific caller's request is the predictable use of a known technique (human content interpretation) applied to the known voice-browser system, yielding the claimed, expected benefit.

Applied to claim 6 (method): step (a) selecting portions — Wise's parser / Rutgers claim 23; step (b) removing non-textual indicia — inherent in any audio-only rendering and expressly performed by Wise's parser (which "selects what portions … are converted to speech") and Kikinis '159's density reduction; step (c) storing on an audio server in accordance with caller-ID data — Rutgers' audio web server + Wise's AIN caller-ID user profiles + Talk Web's subscriber identification; step (d) IVR session — Wise/Rutgers; step (e) playing files per session results — Wise's DTMF/voice navigation and Rutgers' ATML LINK/JUMPTO menus.

Applied to claim 8 (workstation): (a) telephone voice I/O over PSTN or packet-switched network — Wise (POTS/ISDN) and Dekelbaum '682 (dual data-network + PSTN); (b) browse/select pages including text and non-text — Wise's browser 250 + parser 230; (c) remove non-text indicia — Wise's parser selection + Kikinis '159 transposition; (d) store — Kikinis '159 caching; (e) convert text to audio and transmit to a caller — Wise's TTS convertor 260 + audio file player 270.

Ground B — Wise '262 + Talk Web '063 + Kikinis '159 + operator-services art

Talk Web supplies the server-hosts-audio-page + VRU-plays-it-over-the-telephone-link architecture with index/comparison logic and DTMF/voice navigation ("pause, fast-forward, replay or skip"), and explicitly contemplates "a Speech Recognition platform … a telephone keypad platform … or a combination platform."

Motivation: Talk Web's user must know the site name in advance; Wise's users are subject to ASR error. A PHOSITA seeking to let a caller reach arbitrary, unfamiliar Web content with high accuracy would combine Talk Web's server/VRU delivery node with Wise's Web-retrieval-and-TTS capability, and would be motivated by Rutgers' express statement that the retrieval-to-audio step "cannot be fully automated" to put a human in the content-selection loop.

Ground C — Noonen '280 + Wise '262 (+ Kikinis '159) — directed mainly at claim 8 and claim 9

Noonen discloses the telephone-as-Web-browser interface (DTMF menu navigation, attached display); Wise supplies the audio translation and telephone-network audio access; Kikinis '159 supplies content transposition for a capability-limited terminal. Motivation: Noonen is display-dependent, so it fails for the ordinary POTS caller the invention targets (a fact the applicant concedes in the Background). Adding Wise's TTS to Noonen's DTMF browse-with-menu stack is the combination of two references addressing the same problem (telephone access to Web content) by complementary known means — the classic KSR "familiar elements according to known methods."

Ground D — Rutgers '448 + Bellcore '789 + Kikinis '436 — a "prior-art-only" route to claims 1/6

Bellcore '789 supplies the telephone-implemented virtual client, hyperaudio/directory pages, and a markup language for audio pages (HAML) — the functional analogue of the claimed HDML/WML/[any format] storage. Rutgers supplies the audio web server, page authoring by selecting linked data, and translation into audio. Kikinis '436 supplies the human content-interpretation step. Motivation: both Rutgers and Bellcore address the same problem by the same means; Bellcore's motivation section ("the first three pages retrieved were directory pages that were 'stacked'"; the complexity of navigating by audio) supplies the reason to pre-select and structure content for the caller.

Ground E — Ground A + Dekelbaum '682, where the "switch" and VoIP alternatives are at issue

Dekelbaum teaches coordinating a PSTN line with a data-network connection — useful against any argument that the "switch" must be PSTN-specific, and against the packet-switched-network alternative in claim 8. The patent's own specification concedes the point: "Those skilled in the art understand that the invention may employ connection, switching, and voice-over-IP capabilities of the Internet to deliver the same services."

6. Claim-by-claim conclusion

All twelve claims are, in my assessment, invalid under § 103 over the references of record. The strongest, cleanest ground is Ground A against claims 1, 4, 6 and 8 — those claims contain no human-operator limitation and therefore fall to the machine art alone once the "audio web server" element is read onto Rutgers' disclosure.

7. Anticipated patent-owner counterarguments and rebuttals

• "The references teach automation, so combining them with an operator teaches away." No. Neither Wise nor Rutgers disparages human assistance; Rutgers expressly states automated conversion "cannot be fully automated," which affirmatively motivates human involvement. Teaching-away requires a reference that "criticizes, discredits, or otherwise discourages" the combination — Rutgers does the opposite.
• "The examiner considered all 13 references and the patent still issued." Issuance over references considered individually does not defeat a combination rejection. I could not verify whether a combination was ever articulated or traversed (open item, § 0).
• "The claims recite a specific architecture (workstation / IVR / audio web server / switch) not found in one reference." KSR forecloses this: "the combination of familiar elements according to known methods is likely to be obvious when it does no more than yield predictable results." Distributing known functions across known network nodes (operator PC → intermediate audio server → VRU) is the definition of a predictable architecture.
• "'Means for selecting' requires a specific human-operator structure." Two problems for the patent owner. First, the specification discloses no such structure — it states the selection/de-texting program's "construction and operation … are not … disclosed because one skilled in the art may implement such programs in a variety of ways." For a § 112(f) means-plus-function limitation, the claim is then construed to cover the corresponding structure disclosed in the specification and equivalents — and there is essentially none beyond a generic programmed computer — so the limitation is broad enough to read on Wise's parser and Kikinis '159's transposer. Second, if one instead reads "means for selecting" as requiring a human, that limitation appears only in claims 5, 7, 12, which are themselves obvious over Kikinis '436 and the operator-services art. Either construction supports invalidity.
• "Secondary considerations." See § 8.

8. Secondary considerations (objective indicia)

I found no probative secondary-consideration evidence, and the record cuts against it:

• No commercial success evidence is in the record; the patent lapsed for non-payment of maintenance fees effective 2008-11-23 — roughly four years after grant — which is inconsistent with a commercially significant, copied, or widely licensed product.
• No nexus: the only associated litigation is the 2007 Vonage declaratory-judgment action (D. Del. 1:07-cv-00507), which settled in March 2008 without any merits ruling, and — as the litigation section above flags — the case's own AO 120 report and settlement exhibit list three different Nortel patents, so even that link is unconfirmed. No infringement judgment, no royalty, no injunction.
• No long-felt-but-unmet-need weighing toward patentability: the need was recognized and was addressed by the cited art (Wise, Rutgers, Talk Web, Bellcore all confront it), which is evidence the solution was apparent, not that it was nonobvious.
• No unexpected results are disclosed; the stated benefits (accuracy improves with a trained human; cost drops by using existing IVR infrastructure) are the expected consequences of the substitution.

9. Verification gaps (stated rather than papered over)

1. File wrapper not searched — I could not confirm whether the examiner made a § 103 rejection or what the applicant traversed. Recommend pulling the prosecution history via USPTO PatentCenter (application 10/191,081) and its parent 09/420,154.
2. Full texts not read for every reference. I verified Wise '262, Rutgers '448, Kikinis '159, Talk Web '063, and Bellcore '789 from full text/excerpts. For Noonen '280, Dekelbaum '682, Kanoh '077 and MCI '129 I relied on the applicant's own Background characterizations (which are party admissions as to those references' disclosures) plus titles — not the full specifications. Every mapping above that rests on those four is flagged as such and should be confirmed against the reference texts before it is relied on in a filing.
3. General Magic US 6,366,650 ("Method and apparatus for telephonically accessing and navigating the internet") is a § 102(e) reference (filed 1996-03-01) that I identified but did not read in full; it is the patent counterpart of the Web-On-Call™ product the applicant discusses, and it is a strong candidate additional reference for the "telephone navigation of Web content" element.
4. Claim-construction assumptions. I treated the "means for …" limitations under § 112(f) (structure + equivalents) rather than as pure functional language. If a court instead held them indefinite under § 112(b) for lack of corresponding structure, the claims would be invalid on that independent ground (a point already flagged in the identification and PTAB sections above) — which only strengthens the bottom line.

Bottom line

On the references of record, every claim of US 6,823,370 is vulnerable under § 103. The single strongest ground is Wise (US 5,884,262) + Rutgers (US 6,240,448) + Kikinis (US 5,721,159), with Kikinis '436 for the human-operator limitation — a combination whose motivation is supplied by the prior art itself (Rutgers: automatic HTML-to-audio "cannot be fully automated") and echoed in the patent's own Background. Claims 1, 4, 6 and 8 do not even recite a human operator, so they stand or fall on the machine art alone once Rutgers' audio web server is read as the claimed "audio web server." Claims 5, 7 and 12 — the operator claims — are obvious over the technician-in-the-loop transcription art and the long-established operator-services practice. There are no probative secondary considerations, and the patent's early lapse (2008) removes any practical exposure: this analysis is therefore historical/diligence-oriented rather than a live enforcement matter.