US 6760745

US 6,760,745 B1 — Verified Record

Important preliminary note on scope: I searched for US 6,760,745 specifically and for any Court of Appeals for the Federal Circuit (CAFC) 2026 docket activity on that number. I found no CAFC 2026 docket, appeal, or IPR proceeding naming patent 6,760,745. The IPR/Petition documents that surfaced in my searches concern different patents (e.g., U.S. 6,757,796 and a patent referred to in the petition as the "'074 patent"), and I will not treat those as belonging to 6,760,745. I cannot affirmatively prove a negative from a web search, so treat "no 2026 CAFC docket found" as unconfirmed but not contradicted.

---

Bibliographic Data (from the authoritative patent text)

Caveat on assignee: one aggregator page (Unified Patents) displayed "International Business Machines Corp" near the assignee/title fields, but the surrounding text on that page is a "Referenced By" and citation listing. The Google Patents record and the patent's own front page identify the assignee as Aspect Communications Corp. I have not auto-corrected anything; I simply flag the discrepancy.

Abstract (verbatim)

"A method of operating a server includes retrieving a document from a storage device. A variable is generated, the variable including information to enable another server to reproduce the document. The variable may, for example, be a compressed and encrypted complete copy of the document, or may be a JavaScript variable. The variable is appended to the document, and the document is transmitted."

Technical context in one paragraph

The problem addressed is the classical one for the Aspect "WebAgent" product: in a replicated web-server farm behind a load balancer, a "push"/POST-type dynamically generated page (e.g., a brokerage account statement) is customized by whichever replica handled the request and cached only there. A help/interactive-session server cannot know which replica holds that page, and cannot regenerate it from a URL alone. The invention has the serving replica's "mini-filter" append to the outgoing HTML a variable (preferably a JavaScript variable, containing all or part of the page, possibly compressed/encrypted) carrying enough information to reproduce the page. When the user clicks the "help" icon, that variable travels to the session server, which rebuilds the page and pushes it to a customer-service representative's machine. Later claims (as granted) generalize this to selecting a replication server from a plurality of replication servers and sending it the variable.

Plain-language overview of the independent claims

Claim 1 — (server-side method; the "filter" side).
Receive a document request from a client computer, where the request carries one or more identifiers containing customization information. Generate a variable based on that request, the variable containing (a) the customization information and (b) reproduction information enabling a replication server to reproduce the document as a customized document. Identify the replication server out of a plurality of replication servers, and transmit the variable to it so it can reproduce the customized document.
Key point: the claim is directed at choosing one of many replication servers and handing it a self-contained variable — not merely appending a script to a page.

Claim 7 — (machine-readable medium counterpart of claim 1).
Same four steps, expressed as stored instructions that cause a machine to: receive the request with identifiers/customization information; generate the variable (customization + reproduction information for a replication server); identify the replication server among a plurality; transmit the variable to reproduce the customized document.

Claim 10 — (server-side method; dynamically generated document).
Receive a request for a dynamically-generated document from a client computer, where that document includes a customized document customized from customization information in the request so the client can interact with a computer. Generate a variable containing the customization information plus reproduction information enabling a replication server to reproduce the dynamically-generated document; identify the replication server from a plurality of replication servers; transmit the variable to it.

Claim 14 — (session/replication-server-side method; the "mini-filter mirror" side).
Receive a variable from a server computer (the variable having been generated in response to a client's document request), where the variable includes reproduction information and customization information to reproduce the document as a customized document to be displayed on both the first client computer and a second client computer. Reproduce the customized document from the variable, and transmit it to both the first and second client computers.

Claim 15 (dependent on 14) adds applets to the customized document to perform functions assisting an interactive session between the first and second client computers. Note: the text supplied to me is truncated mid-claim 15, so I cannot see claims 15–29 in full.

Independent-claim count caveat: The patent states 29 claims total, but the authoritative text in my possession is cut off partway through claim 15. I can therefore confirm independents 1, 7, 10, and 14, but I cannot confirm whether additional independent claims exist at claims 16–29. I am not going to guess at their content.

Drafting note worth flagging

There is a visible tension between the specification/abstract and the granted claims. The abstract and detailed description emphasize appending the variable to the HTML document and transmitting the document to the client, whereas granted claim 1/7/10 recite identifying a replication server from a plurality of replication servers and transmitting the variable to that replication server (the client-mediated return path is captured in claim 6 and in claim 14's "receive from a server computer" flow). This suggests the claim set was amended during prosecution relative to the specification's framing.

Uncertainty summary

1. No CAFC 2026 docket for 6,760,745 located — reported as a negative search result, not proof of absence.
2. Claims 16–29 not fully available to me; independently verifiable only for claims 1, 7, 10, 14 (and 15, partially).
3. Assignee listing discrepancy noted above but not "corrected."
4. The patent expired May 16, 2020 (anticipated expiration; status "Expired – Lifetime"), so any 2026 appellate activity would be unusual though not impossible (e.g., long-running district litigation).

Sources: Google Patents record for US6760745B1 (https://patents.google.com/patent/US6760745/en); USPTO patent PDF (US006760745B1), front page and claim text; USPTO assignment records as reflected on the Google Patents "Reassignment" timeline.

I'll search for litigation involving US patent 6760745 specifically.

Let me search more specifically for litigation records and any assertion activity.

Let me check litigation databases and the Unified Patents portal directly.

Litigation Report — U.S. Patent No. 6,760,745

Bottom line: I found no litigation involving U.S. Patent No. 6,760,745. My searches of patent-litigation-oriented sources returned no district court, ITC, PTAB, CAFC, or PACER docket in which the '745 patent (Aspect Communications / "Web server replicated mini-filter") is an asserted patent or a patent-in-suit. The honest answer to your question is: no known litigation.

Because I could not exhaustively confirm a negative, here is the record of what I checked and the important caveats.

1. Patent identification (verified, interpreted literally)

2. Sources searched and what they returned

• Google Patents / USPTO full text (patent page). The authoritative record shows the assignment/reassignment chain (2000–2021) and status "Expired – Lifetime." Critically, the Google Patents "Litigation" data is absent/empty for this patent, which is the standard proxy for "no dockets mapped to this patent number." [patents.google.com/patent/US6760745/en]
• Unified Patents Portal — patent page for US-6760745-B1. The page renders forward citations ("Referenced By (159)") and other bibliographic data but surfaced no district-court litigation list and no PTAB/IPR/reexam entry tied to '745. [portal.unifiedpatents.com/patents/patent/US-6760745-B1]
• Keyword searches on the patent number plus litigation terms ("6760745 litigation," "6,760,745 infringement lawsuit") returned only the patent document itself, unrelated amicus/case PDFs, and unrelated patents.
• Aspect Communications / WebAgent litigation search returned the prior-art framing material (the WebAgent product background in the patent itself), not any suit on this patent.

3. Important caution — do NOT confuse '745 with other "745" patents

My searches repeatedly returned other patents whose numbers end in "745." These are not US 6,760,745 and must not be attributed to it:

• US 6,687,745 — Droplets, Inc. (interactive links / remote application delivery). Currently asserted against Walmart, Ford, Home Depot, AT&T; Unified Patents filed an ex parte reexamination (control no. 90/019,944; instituted July 21, 2025).
• US 10,687,745 — Masimo Corp. (Apple Watch pulse oximetry); ITC Inv. No. 337-TA-1276 and IPR2022-01465 (Apple v. Masimo).
• US 10,799,745 — Woodway USA (LifeCORE Fitness, S.D. Cal. Case No. 3:22-cv-00492-JO-BLM).
• US 5,894,554 / US 6,415,335 — EpicRealm Licensing (dynamic web page generation); the related litigation and reexam records (Reexam Control Nos. 90/008,562 and 90/008,574; EpicRealm v. AutoFlex Leasing, E.D. Tex. Nos. 2:05-cv-163 / 2:05-cv-356) concern different patents in the same general technology space.

None of these are the patent you asked about. Per your instruction, I have not auto-corrected or substituted any of these numbers for '745.

4. Why an absence of litigation is plausible here

• Expiry. The patent's term expired May 16, 2020 (per the recorded anticipated-expiration date). There is no meaningful prospect of prospective injunctive relief today, and the damages look-back window has largely closed, which substantially reduces incentives to assert it.
• Prosecution history. The examiner cited US 6,415,335 (EpicRealm) and others as prior art during prosecution — relevant if you are assessing validity, but it does not itself indicate litigation.
• Ownership fragmentation. The assignment chain shows a heavily encumbered patent (multiple successive security interests and releases through 2016–2021), which is common in patents that were never commercialized as standalone litigation assets.

5. Recommended verification steps (to convert "no known litigation" into a confirmed negative)

Since I could not complete a full PACER/CAFC sweep before my search budget ran out, I recommend confirming with:

1. PACER Case Locator / CourtListener RECAP — full-text search for "6,760,745" and "6760745" across all district courts and the Federal Circuit.
2. Unified Patents Litigation Portal — https://portal.unifiedpatents.com/litigation/caselist — query by patent number, plus the PTAB and ex parte reexam tabs.
3. Docket Navigator / Lex Machina — patent-level docket report; these are the most reliable for mapping a patent number to all cases (including pre-suit and transferred cases).
4. USPTO Patent Assignment Search and the "Litigation Search Report" (CRU) — as illustrated by the EpicRealm CRU report I found, the USPTO's Central Reexamination Unit compiles KeyCite/Lexis litigation histories on request.
5. ITC EDIS — Section 337 investigations naming the patent.

6. Statement of confidence

• High confidence that US 6,760,745 B1 exists as described and relates to Aspect Communications' WebAgent technology.
• Moderate-to-high confidence, but not certain, that no litigation has ever been filed on US 6,760,745. This conclusion rests on the absence of any "Litigation" data on Google Patents, no dockets on the Unified Patents portal, and repeated null keyword searches. It is a negative finding based on the sources I could reach, not a records-certified clearance.
• I did not locate any CAFC opinion, PTAB proceeding, or ITC investigation specifically involving US 6,760,745.

If you need a litigation-clear opinion, the PACER/CourtListener and Docket Navigator searches in §5 are the next step; I can run those queries if you'd like me to continue.

I'll verify the structured block's "no proceedings" signal against web sources before writing the report.

Let me run a few more targeted checks for any AIA trial or reexam activity, and for the near-miss proceeding numbers.

One more check for ex parte reexamination (which wouldn't appear in the AIA-trial ODP feed) before I write up.

Proceedings overview

Total AIA trial proceedings on US 6,760,745: zero (0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denials). The patent has never been the subject of an IPR, PGR, or CBM at the PTAB, so all 29 claims — including independents 1, 7, 10, and 14 — are UNTESTED and stand un-cancelled, but that is not a hardening signal: it reflects a patent that expired 2020-05-16 and was never asserted, so the correct defensive posture is "nothing to attack, because there is nothing (left) to defend against" — a demand letter threatening suit on an expired, never-litigated, never-IPR'd patent is an operational/collectability problem, not a validity problem.

The structured "PTAB proceedings on file" block in this prompt is the canonical source and returns no AIA trial proceedings from the USPTO Open Data Portal. I re-verified that negative by web search and found nothing contradicting it. Per the operating instruction, I am not inventing proceeding numbers to fill the section.

No proceedings to report — verification record

I will not fabricate a proceeding entry. Instead, here is what I checked and what came back:

Near-misses I explicitly refuse to attribute to this patent:

• US 6,775,745 (Intellectual Ventures I LLC — hybrid data caching). This one does have a proceeding: IPR2016-01643, Unified Patents Inc. v. Intellectual Ventures I LLC, filed 2016-08-19 challenging claims 1–4, 6–9, 11–17; FWD 2018-03-26 invalidating all challenged claims (https://www.unifiedpatents.com/insights/2016/8/19/unified-challenges-second-intellectual-ventures-patent and https://www.unifiedpatents.com/insights/2018/3/26/intellectual-ventures-patent-found-invalid). Note the number is 6,775,745, not 6,760,745 — a one-digit difference at the fourth position, and a completely different patent owner, technology, and claim set. It also appeared in NetApp's 2016 power-of-attorney covering U.S. 6,516,442 / 6,775,745 / 6,968,459 / 8,275,827 / 8,387,132. None of this touches the Aspect patent.
• JP 6760745 B2 (Nippon Paper Industries — pressure-sensitive copying paper). Same digits, foreign jurisdiction, unrelated art.
• A December 2016 PTO Litigation Center report lists a new IPR petition on "Patent # 6,775,745" — again the IV patent.

Confidence and its limits: Moderate-to-high that no AIA trial has ever been filed on US 6,760,745. The AIA trial regime opened 2012-09-16 and closed for CBM on 2020-09-16; a proceeding filed in that window should be visible in ODP plus the public portals, and nothing is. I cannot certify a negative from search alone — a proceeding indexed only under a differently-formatted identifier, or an ex parte reexamination (which lives in a separate CRU feed and would not appear in the AIA-trial block), is the residual gap. The ex parte reexam search returned nothing on '745 either.

---

Strategic summary

Claim status of US 6,760,745. Because there is no FWD, no certificate, and no reexam, no claim of the '745 patent has ever been cancelled or amended post-grant. Claims 1, 7, 10, and 14 — and the full set 1–29 — remain exactly as granted on 2004-07-06. I can confirm the independent-claim count only through claim 14 (and claim 15 partially); the earlier section of this analysis correctly flagged that claims 16–29 were not fully available, so I will not opine on whether further independents exist. On the PTAB axis specifically, the entire claim set is UNTESTED.

Estoppel landscape — § 315(e)(2). The estoppel ledger is empty. Section 315(e)(2) estoppel attaches only to a petitioner, real party in interest, or privy that reaches a final written decision. There is no petitioner, so no party is estopped on any ground. Practically, a defendant facing a demand today has the full universe of § 102/§ 103 art available — including art that would be "reasonably could have raised" bars in a litigated IPR — plus § 101 and § 112 grounds, which the PTAB cannot reach in an IPR anyway. That is a theoretical advantage with little real value, because a defendant's cheapest move on an expired patent is not an IPR at all (see below).

Pattern signals. No petitioner has filed even once, so there is no repeat-petitioner pattern; no patent owner has ever had to defend a PTAB appeal, so there is no aggressive PTAB-appeal posture to read; and no defensive aggregator appears in the chain — there is no Unified Patents/RPX-style involvement with this patent (Unified's filings near this number belong to the '745 vs. '745 confusion above). The assignment/security chain shows JPMorgan Chase, D.B. Zwirn, Deutsche Bank, U.S. Bank, and Wilmington Trust, with all security interests released by 2021 — encumbrance typical of a portfolio asset pledged for debt, not a litigation vehicle. The patent also appears in Aspect Software's 2016 bankruptcy asset schedule (Part 10, intangible IP) with value recorded as "UNKNOWN / Undetermined", and by then its remaining term was under four years. Nothing in that profile is consistent with an assertion campaign or an IPR response.

Why the absence is unremarkable here. The usual inference — "a well-asserted patent eventually attracts an IPR" — does not bite. The '745 patent's fifteen-year enforceable life (2004–2020) overlapped the pre-AIA era (no IPR until 2012) and, once IPRs existed, there is no evidence of any infringement suit, ITC investigation, or licensing campaign on it (consistent with the litigation section of this analysis). Patents that are never asserted are rarely IPR'd, because IPR is a defense and a deterrence tool, not an end in itself.

---

Recommended next steps

1. Say it plainly in any opinion work: there is no PTAB record. If you are asked "what happened at the PTAB on 6,760,745," the accurate answer is "no petition was ever filed; no institution decision, FWD, or appeal exists." Do not let the USP 6,775,745 / IPR2016-01643 materials leak into that answer — the FWD there cancels claims of a different patent. Link, if you cite the near-miss: https://www.unifiedpatents.com/insights/2018/3/26/intellectual-ventures-patent-found-invalid.
2. Anchor the real defense on expiry, not validity. The patent's term expired 2020-05-16 ("Expired – Lifetime"; anticipated expiration recorded 2020-05-16). Any pre-suit demand today can only reach damages for past infringement within the § 286 six-year look-back — i.e., conduct after 2020-04-01 at the earliest — against a patent with no live injunctive leverage. Check whether the demand even attempts to plead pre-2020 conduct; if it recites claims 1, 7, 10, or 14 as the hook, note that those claims are live-but-expired, not cancelled.
3. If you nonetheless want a PTAB kill, the window is not closed by expiry. The Board has instituted IPRs on expired patents (the trial yields a validity determination, not an injunction), and there is no § 315(b) one-year bar running against a party that has never been served with an infringement complaint. But weigh the cost — an IPR through FWD plus appeal runs into the mid-six figures — against the near-zero value of invalidating an expired patent. The rational defendant's move is a § 282 / § 286 posture plus a laches/notice analysis, not a petition.
4. Close the two residual verification gaps before any clearance-style opinion: (a) PTAB E2E / PTAB API (https://ptabdata.uspto.gov/) and USPTO Patent Center for a records-level sweep on 09/572,016 including any ex parte reexamination (CRU feed, not the AIA-trial block — this is the one place a proceeding could hide from the ODP result above); (b) CourtListener RECAP / PACER Case Locator full-text on "6,760,745" and "6760745."
5. Trial-stage milestones: N/A. With no active proceeding there is no institution deadline, no oral hearing date, and no one-year FWD clock running. If a petition is ever filed, the statutory clocks are: institution decision within 6 months of the § 315(b) bar date or petition filing, and FWD within 12 months of institution (extendable up to 6 months for good cause, § 316(a)(11)).

Disclaimer: I found no PTAB proceeding, no FWD, no settlement, no judge panel, and no CAFC appeal for US 6,760,745 — because, on the available record, none exists. Every statement above is either quoted from the structured ODP block or grounded in the cited URLs; the negative finding is a documented search result, not a records-certified clearance.

I'll research the assignment record for US 6,760,745, focusing on the USPTO Assignment Center records (reel/frame and correspondent data).

Significant find — the patent appears on an Aspect Software bankruptcy schedule. Let me dig into that and the assignment records.

Let me try to pin down reel/frame numbers and correspondent data, plus check for Aspect's later restructurings.

Assignment / Ownership Analysis — U.S. Patent No. 6,760,745 B1 ("Web server replicated mini-filter")

Data-source note (read first). The authoritative patent text I was given contains the Google Patents "Reassignment" event list for this patent — with dates, conveyance types, assignors and assignees — but it does not expose USPTO reel/frame numbers or the recorded correspondent. My attempts to reach assignmentcenter.uspto.gov, assignment.uspto.gov, and the referral/PDF links (bankrupt.com, kroll.com, unifiedpatents) returned only index/redirect content or were cut off by the search budget. Accordingly:

• The owners, dates, and conveyance types below are reliable (they come from the authoritative record).
• Reel/frame numbers and correspondents are marked "not retrieved — do not rely on these as absent." I am not fabricating reel/frame values to fill a template. Any analyst needing the reel/frame + correspondent fields must pull the record directly at Assignment Center (see Verification link at the end).

---

Inventors

Pattern observations:

• All three named inventors assigned to the same original assignee and the assignment was recorded 2000-10-17, i.e., roughly 5 months after the 2000-05-16 filing and ~7 months after the 2000-03-30 provisional. This is a routine employee "assignment of assignors' interest," not an inventor-retention or inventor-owned structure.
• Inventor-departure pattern: unclear. I have no employment/departure data for Tan, Sundararajan, or Mohan, and I will not infer a "mass departure preceding a portfolio fire-sale" from silence. There is no evidence of inventors spinning the patent into a separate inventor-held entity.

---

Original assignee

Aspect Communications Corp. (assignment recorded 2000-10-17).

• Primary line of business: contact-center / customer-interaction software — automatic call distribution, IVR/self-service (Voxeo), workforce optimization, and web collaboration/co-browsing. The patent's own background section describes the Aspect "WebAgent" product, which is the product environment this invention serves.
• Did they ship a product embodying the claims? Likely yes. The specification is written as an improvement to a shipped commercial system ("WebAgent software from Aspect Communications, Inc.") and describes the mini-filter as a server-side component of that system. This is an operating-company patent tied to a real product line, not a paper asset.
• Current status of the corporate family: Aspect's corporate family filed Chapter 11 on 2016-03-09 in the U.S. Bankruptcy Court for the District of Delaware, Case No. 16-10597 (MFW) (Aspect Software Parent, Inc., et al., jointly administered; Aspect Software, Inc. = 16-10598). This was a pre-arranged reorganization, not a liquidation: the company equitized second-lien debt, reduction of >$320M of indebtedness, and emerged in 2016. It was not a Kodak/Nortel-style §363 patent fire-sale. (Confirming docket: Aspect Software SAL, Case 16-10597-MFW Doc 149, filed 04/06/16.)
• Later status: Aspect Software subsequently combined with Noble Systems to form Alvaria (2021). Moderate confidence — verify against the Alvaria corporate record; flagged rather than asserted. Note the '745 patent expired 2020-05-16, before that combination.

---

Assignment timeline

Google Patents records 14 reassignment events for US 6,760,745. Reel/frame is not available in the retrieved record for any entry; correspondent is not available for any entry. Executed vs. recorded date is also not separated in the retrieved feed (dates below are the recorded-event dates shown).

⚠️ Do not misread the "assignee" column. Several events are Security Agreement / Security Interest / Release conveyances. In those filings the secured lender is listed as "assignee" and the company as "assignor." The lender is a lienholder, not an owner. The only true ownership movements below are the 2000-10-17 inventor→Aspect assignment and the internal Aspect-family successor naming.

Correspondent recurrence (Signal 3): NOT DETERMINABLE. I could not retrieve the correspondent-of-record field for any entry, so I will not assert a repeat-player attorney. Separately, the grant-face prosecution firm is Blakely Sokoloff Taylor & Zafman — but that is the prosecution correspondent on the issued patent, not confirmed as the assignment-record correspondent, and I am not treating it as a finding.

Bankruptcy cross-reference (independent, non-assignment corroboration): The '745 patent appears on Schedule A/B, Part 10 (Intangibles), line 60.607 of the Aspect Software chapter 11 schedules filed 04/06/2016 as "ISSUED PATENTS – WEB SERVER REPLICATED MINI-FILTER (ASPECT COMMUNICATIONS) REG NO.6760745, DATED 7/6/2004", net book value "UNKNOWN / Undetermined." Source: http://bankrupt.com/misc/AspectSoftware_SAL04062016.pdf. This confirms (a) the debtor entity treated Aspect Communications' patents as its own assets, and (b) that the IP was collateral of the estate, valued at book as undetermined.

Bottom line on ownership: Despite 14 recorded events, the patent never left the Aspect operating-company family. The endpoint owner of record is Aspect Software, Inc. (successor to Aspect Communications Corp.), and the last lien was released 2021-06-04.

---

Timeline diagram

timeline
    title Ownership of US 6760745
    2000 : Filed by Aspect Communications
         : Inventors assign to Aspect
    2004 : Patent granted
    2005 : JPMorgan security interest
         : DB Zwirn security interest
    2006 : DB Zwirn lien released
         : Deutsche Bank security agreement
    2010 : JPMorgan lien released
         : Deutsche Bank lien released
         : Aspect Software named successor
         : New JPMorgan security agreement
         : US Bank security interest
    2014 : Wilmington Trust named agent
    2016 : US Bank lien released
         : Wilmington Trust lien released
         : Aspect files Chapter 11
         : New Wilmington Trust lien
    2020 : Patent term expires
    2021 : Wilmington Trust lien released

---

NPE / troll-pattern signals

---

Verdict

Defensive / non-asserting — with an explicit taxonomy caveat.

Justification. The chain contains zero NPE signals: no licensing-only shell (Signal 1 absent), no known asserter (Signal 2 absent), no chained-LLC cascade (Signal 4 absent), and no assertion-driven transfer (Signals 5, 7 absent). Ownership begins with the 2000-10-17 inventor → Aspect Communications assignment, passes through Aspect Software, Inc. as successor to Aspect Communications Corporation (recorded 2010-06-09), survives an operating reorganization in Chapter 11 Case 16-10597-MFW (patent scheduled at Sched. A/B line 60.607, 2016-04-06), and the final recorded lien is released 2021-06-04 — after the patent's 2020-05-16 expiry. The 14-event count is an artifact of secured-lender filings (JPMorgan, D.B. Zwirn, Deutsche Bank, U.S. Bank, Wilmington Trust appearing as "assignee" on liens), not of ownership churn.

Taxonomy caveat (flagging a possible contradiction with the rubric): the rubric's "Defensive / non-asserting" bucket is defined as "chain terminates at a defensive aggregator." That specific criterion is not met here. Likewise "Operating-company assertion" is not met (the owner is an operating company, but there is no suit). I have mapped to the non-asserting bucket because non-assertion is the dominant, evidenced characteristic, but if a strict aggregator-termination test is applied, this patent falls outside all five provided buckets and should be recorded as "operating-company owned, never asserted."

---

Verification link

USPTO Assignment Center — search patent number 6760745: https://assignmentcenter.uspto.gov/ (mirror: https://assignment.uspto.gov/patent/index.html). Pull each hit's Reel/Frame and Correspondent fields to complete Signals 3 and the timeline's missing columns. Google Patents legal-events view: https://patents.google.com/patent/US6760745/en.

Open items / low-confidence points

1. Reel/frame and correspondents are unretrieved — the single biggest gap; the record clearly exists (14 events) but I could not extract those fields.
2. Inventor employment/departure history — unknown; no inference drawn.
3. Alvaria (2021 Aspect–Noble combination) — stated at moderate confidence; verify.
4. Google Patents "Current Assignee: Aspect Communications Corp; Wilmington Trust NA" is an artifact of the event feed (Wilmington Trust is a lienholder, not an owner). Do not read it as co-ownership.

I'll search for the authoritative patent record and its cited prior art.

I have the examiner-cited reference list. Now let me verify each reference's identity, dates, and content individually.

Let me pull the specific bibliographic and abstract data for each remaining cited reference.

Prior-Art Analysis — U.S. Patent No. 6,760,745 B1

Analysis date: April 26, 2026. Patent expired (anticipated expiration 2020‑05‑16), so this is a validity-assessment / defensive exercise.

---

0. Identification discipline (literal-ID check)

The search was anchored on the exact identifier 6760745. The record returned is:

Numbers deliberately NOT reported as this patent: 6,687,745 (Droplets), 10,687,745 (Masimo), 10,799,745 (Woodway), 6,415,335 is reported only as a reference cited on 6,760,745 — it is a different patent. Also note 5,894,554 / 6,415,335 (EpicRealm) belong to the same technology neighborhood but are distinct documents.

⚠ Contradiction flagged with the earlier-generated section

The previously generated Patent summary stated the grant-face citations were:

"US 5,867,706 (Martin et al.); US 6,035,119 (Massena et al.); US 6,185,598 (Farber et al.)"

That list is incomplete. The authoritative front page of the USPTO grant (US006760745B1) lists seven U.S. patent references. The full set is set out verbatim below. This section supersedes the earlier three-reference list. (The earlier Litigation summary's statement that the examiner cited "US 6,415,335 (EpicRealm) and others" is correct and is confirmed here.)

---

1. The cited references — verbatim from the grant front page

From the USPTO grant document (patentimages.../US6760745.pdf, "References Cited"):

U.S. PATENT DOCUMENTS
5,867,706 A * 2/1999 Martin et al. ...... 709/105
6,035,119 A * 3/2000 Massena et al. ...... 717/100
6,185,598 B1 * 2/2001 Farber et al. ...... 709/200
6,226,642 B1 * 5/2001 Deranck et al. ...... 707/120
6,385,642 B1 * 5/2002 Chian et al. ...... 709/203
6,397,259 B1 * 5/2002 Lincke et al. ...... 709/236
6,415,335 B1 * 7/2002 Lowery et al. ...... 710/5

• cited by examiner—Mehmet B. Geckli
  (74) Attorney, Agent, or Firm—Blakely Sokoloff Taylor &

All seven carry the examiner-citation asterisk. No foreign patent documents or non-patent-literature entries were visible in the portion of the front page I could retrieve (see §6).

Note on "Deranck": the OCR of the grant renders the fourth reference's inventor as "Deranck et al.", whereas the commercial record for the same patent number (6,226,642) renders "Beranek et al." Per the strict no-auto-correction rule I am reporting both strings verbatim and not silently substituting one for the other. This is most likely an OCR artifact but it is flagged, not fixed.

---

2. Master table

---

3. Reference-by-reference

3.1 — US 5,867,706 A — Martin et al.

• Full citation: U.S. Patent 5,867,706 A, "Method of load balancing across the processors of a server," Martin et al., issued Feb. 2, 1999. Class 709/105 (current 718/105). (Title and issue date independently corroborated: FreePatentsOnline front-matter listing for the reference table of US 8,281,035, which renders "5867706 | Method of load balancing across the processors of a server | 1999‑02‑02 | Martin et al. | 718/105".)
• Brief description: Distributing computational work among the processors of a single server (intra-machine task allocation), per the title.
• §102 analysis: Because it issued 1999‑02‑02, it is a §102(b) reference against a March 30, 2000 priority date (more than one year earlier), independent of its filing date — the cleanest statutory position of any reference here. Substantively, however, it addresses processors within one server, not selection of one server among a plurality of replicated document servers. It says nothing about generating a variable containing customization information plus reproduction information, and nothing about dynamically generated per-user documents. My reading of the title alone leaves it non-anticipatory of claims 1, 7, 10 and 14; at most it is §103 art on the "identify a server" element by loose analogy.
• Confidence: high on citation/date; medium on the substance (I read the title, not the specification).
• §102 target claims: none convincingly; nominal relevance to claim 1 step (c)/7/10.

3.2 — US 6,035,119 A — Massena et al.

• Full citation: U.S. Patent 6,035,119 A, Massena et al., issued March 2000 (day not confirmed in the record I retrieved). Class 717/100 (software development/installation).
• Brief description: Cannot state with confidence. Class 717/100 places it in software installation/distribution, but I did not retrieve its title, abstract or specification in this pass. I will not guess.
• §102 analysis: Only the statutory-category level can be addressed. Issued March 2000, i.e. less than one year before the earliest critical date, so it cannot be §102(b); it would have to be §102(a) (if it issued before the applicant's invention date — plausible if the invention date is the 2000‑03‑30 provisional) and/or §102(e) (US patent granted on an earlier-filed application). Both determinations require facts I lack (exact issue day; US filing date; applicant's invention date).
• Confidence: low.
• §102 target claims: not mapped — mapping without the document would be fabrication.
• Notable oddity worth flagging: the '745 specification's Field of the Invention reads "The invention relates generally to the installation of software on a computer system." That text does not match the claimed subject matter at all, and it is suspiciously aligned with the 717/100 classification of this very citation. This looks like a copy‑paste artifact in the specification, not a substantive limitation — relevant if you are construing the specification's scope.

3.3 — US 6,185,598 B1 — Farber et al.

• Full citation: U.S. Patent 6,185,598 B1, "Optimized network resource location," Farber et al., issued February 2001 (day not confirmed in this pass). Class 709/200.
• Brief description: Distributing copies of network resources across multiple distributed servers and directing a requesting client to an appropriate one of them (the Digital Island / Cable & Wireless / Level 3 content-delivery lineage; the "Optimized network resource location" title is confirmed by the Level 3 family page surfaced in search, e.g. FreePatentsOnline's US 8,281,035 — itself titled "Optimized network resource location – Level 3 Communications, LLC").
• §102 analysis: Issued after the 2000‑05‑16 filing, so §102(a)/(b) are unavailable on its face; if it is prior art at all it must be §102(e) (granted on a U.S. application filed before the applicant's invention date). Substantively this is the best of the seven on the element the granted claims actually turn on — claim 1/7/10's "identifying the replication server from a plurality of replication servers." But it is directed to replication/redirection of content to servers, and I see nothing in the title-level characterization about generating a variable that contains customization information for a per-user dynamically generated page plus reproduction information, nor about appending such a variable to a document. Missing at least one element of every independent claim ⇒ §103, not §102.
• Confidence: high on citation/date; medium on substance.
• §102 target claims: nominal only — the "plurality of servers" limitation in claims 1, 7, 10.

3.4 — US 6,226,642 B1 — "Deranck"/Beranek et al.

• Full citation: U.S. Patent 6,226,642 B1, "Content modification of Internet web pages for a television class display," Beranek et al. (grant OCR: "Deranck et al." — flagged, not corrected), issued 2001‑05‑01. Class as printed on the grant: 707/120; a commercial listing for the same number renders 707/10 (discrepancy flagged, likely OCR/transcription).
• Brief description: Modifying Internet web-page content for display on a television-class device — i.e., an intermediary that transforms/modifies a web document in transit before it reaches the client.
• §102 analysis: Issued post-filing ⇒ §102(e) only (assuming a pre-invention U.S. filing date). It is relevant in spirit to the "mini-filter" concept (something operating on the document between server and client) and to claim 3 ("the customized document comprises a markup language document"), but it does not disclose a variable carrying customization information and reproduction information, nor server selection. No anticipation.
• Confidence: high on citation/date; medium on substance.
• §102 target claims: nominal only — claim 3 and the "filter" concept.

3.5 — US 6,385,642 B1 — "Chian" et al.

• Full citation: U.S. Patent 6,385,642 B1, inventor as printed "Chian et al." (may be an OCR variant of "Chiang"), issued May 2002. Class 709/203 (client/server distributed data processing).
• Brief description: Not established. I did not retrieve the title, abstract or specification.
• §102 analysis: Statutory category only: issued post-filing ⇒ at best §102(e). Whether it even touches claims 1/7/10/14 cannot be determined without the document. Class 709/203 (client/server) suggests session/request handling rather than document replication, but that is an inference from a classification symbol, not evidence.
• Confidence: low.
• §102 target claims: not mapped.

3.6 — US 6,397,259 B1 — Lincke et al.

• Full citation: U.S. Patent 6,397,259 B1, Lincke et al., issued May 2002. Class 709/236.
• Brief description: Not established. I did not retrieve the title, abstract or specification. Class 709/236 sits in the network-interface/protocol region of old Class 709, which could indicate intermediary/proxy or protocol-conversion subject matter — but I am explicitly declining to characterize it on a subclass symbol alone.
• §102 analysis: Statutory category only: post-filing issue ⇒ at best §102(e).
• Confidence: low.
• §102 target claims: not mapped.

3.7 — US 6,415,335 B1 — Lowery et al.

• Full citation: U.S. Patent 6,415,335 B1, "System and Method for Managing Dynamic Web Page Generation Requests," Lowery et al., issued July 2002. Class 710/5. Listed assignee in later records: Parallel Networks Licensing LLC (the EpicRealm lineage). (Title and family confirmed via the Unified Patents portal entry for US‑6415335‑B1, which renders "System and Method for Managing Dynamic Web Page Generation Requests" with a 1996‑04‑22 priority and Parallel Networks Licensing LLC.)
• Brief description: Management/routing of requests for dynamically generated web pages — handling page-generation requests and serving the generated pages.
• §102 analysis: This is the most substantively on-point reference in the set and the only one whose subject matter overlaps the '745 preamble "dynamically-generated document" (claim 10) and the reproduce-and-serve step (claim 14). It is nevertheless not a §102 anticipation on the visible record: I have no indication it discloses (i) generating a variable comprising the customization information plus reproduction information, nor (ii) transmitting that variable to a selected replication server. Statutory note: it issued 2002‑07‑02, after the '745 filing, so it cannot be §102(a)/(b) — its citable status must rest on §102(e).
• Why that inference matters: because the examiner cited it, its effective U.S. filing date is, as a practical matter, earlier than the applicant's invention date. That is a logical inference from the fact of citation, not a verified fact.
• Cross-reference: the earlier Litigation summary listed EpicRealm's 5,894,554 / 6,415,335 as heavily litigated dynamic-page-generation patents. Consistent with what I see here — the same family is the primary §103 reference against the '745 claims. Do not conflate the two patents' dockets.
• Confidence: high on citation/date/title; medium on substance.
• §102 target claims: nominal — claim 10 preamble; claim 14's reproduce/transmit steps.

---

4. Reference → claim mapping matrix

Independent claims visible to me: 1, 7, 10, 14 (plus dependent 2–6, 8–9, 11–13, and 15 in part). "✗" = element absent on the face of the reference; "?" = cannot assess.

No column contains a reference that supplies every element of any independent claim.

---

5. Bottom line on §102

1. No reference cited on the face of US 6,760,745 appears to anticipate any independent claim. The distinguishing element in every case is the same: a variable that is generated from the request, contains the customization information and reproduction information, and is transmitted to an identified replication server so that server can reproduce the customized document. None of the seven cited references — on the record I can verify — discloses that construct.
2. The citation set reads as a §103 combination, not a §102 collection: a server-selection/load-balancing reference (5,867,706; 6,185,598), a document-modification/intermediary reference (6,226,642), a dynamic-page-request-management reference (6,415,335), plus three references whose relevance I cannot assess (6,035,119; 6,385,642; 6,397,259).
3. Statutory-category split (pre-AIA §§102(a)/(b)/(e)), by issue date relative to the 2000‑03‑30 / 2000‑05‑16 critical dates:
   • §102(b)-capable: only 5,867,706 (1999‑02‑02).
   • §102(a)-capable: 6,035,119 (March 2000, if before the invention date).
   • §102(e)-only (issued after the filing date): 6,185,598, 6,226,642, 6,385,642, 6,397,259, 6,415,335. For each, citable status depends on a U.S. filing date earlier than the applicant's invention date.
4. I did not fabricate a §102 mapping where the record does not support one. Three of seven references (6,035,119, 6,385,642, 6,397,259) are returned as unassessable rather than guessed.

---

6. What I could NOT verify (and why it matters)

1. Filing dates for all seven references. This is the single biggest gap. §102(e) status for five of the seven references turns entirely on filing date vs. the invention date, and I have issue dates only. My searches for filing dates for 6,035,119, 6,397,259 and 6,415,335 returned no usable record before the search budget was exhausted.
2. Titles/abstracts for US 6,035,119, US 6,385,642 and US 6,397,259. No substantive §102 mapping is possible without these.
3. The remainder of the front page. The retrieved OCR stops at the attorney/firm line. I therefore cannot confirm whether the grant face also lists foreign patent documents or non-patent literature (Google Patents convention is a "Foreign Patent Documents" and "Other Publications" block). Do not assume the list in §1 is exhaustive of all cited art.
4. The prosecution history. I did not retrieve any Office Action, examiner's statement of reasons for allowance, or the applicant's amendments. The earlier Patent summary already flagged that the granted claims (server-selection + transmit-to-replication-server) sit in tension with the specification's framing (append variable to the document and send it to the client). That tension is presumably the residue of amendments made to overcome exactly these references — but I have no documentary proof of that.
5. Inventor-name / classification discrepancies, flagged not corrected: "Deranck et al." (grant OCR) vs. "Beranek et al." (FPO listing) for 6,226,642; "Chian et al." (grant OCR) vs. possibly "Chiang"; class 707/120 (grant OCR) vs. 707/10 (FPO listing) for the same patent.

---

7. Admitted prior art outside the citation list

Two items in the '745 disclosure itself function as prior art and are not in the front-page citation list:

• The applicant's own WebAgent software. Figures 1 and 2 of the patent are expressly labelled "prior art," and the specification describes WebAgent's practice of caching a "post"-type dynamically generated page on the serving web server and having the WebAgent server retrieve that cached copy. This is an admission of prior art that could operate under §102(a)/(b) against broad claims — subject to establishing the product's public-use/on-sale dates relative to the critical date, which the record does not state. Any obviousness position against the '745 patent's caching-of-generated-pages concept starts here, because the patent itself concedes the single-server version of the technique.
• "Web Interaction 2000 v. 2.0 planning Guide ... provided in Appendix B." Referenced in the specification as the source of the applet functionality (chat box, VoIP, whiteboard) that claim 15 recites. The appendix is not part of the printed patent. Whether it was ever publicly available, and its date, are unknown — flag as a potential NPL reference pending verification.

(The provisional application 60/193,176, filed 2000‑03‑30, is a priority document, not prior art.)

---

8. Recommended next steps to close the gaps

1. Pull the front pages and full texts of US 6,035,119, US 6,385,642 and US 6,397,259 — three of seven references are currently unassessable.
2. Pull filing dates for all seven so §102(e) vs. §102(a)/(b) can be fixed definitively.
3. Pull the file wrapper (PatentCenter for 09/572,016) — the amendments and examiner's reasons for allowance will show which reference(s) drove the "identify the replication server from a plurality" limitation into the granted claims.
4. Resolve the "Deranck"/"Beranek" rendering against the printed grant.
5. Confirm whether the grant face lists foreign patent documents / other publications beyond the seven U.S. references reproduced in §1.

Overall confidence: high that the seven-reference list and its dates are accurately transcribed from the grant; high that none of the four assessable references anticipates an independent claim; moderate-to-low on the three references I could not retrieve, which I have reported as unknowns rather than filling in from assumption.

I'll ground the obviousness analysis in the prior-art references identified for this patent. Let me verify each reference.

Let me look up the specific cited references directly.

Let me pin down the two cited references whose details I could not yet confirm.

Obviousness Analysis — U.S. Patent No. 6,760,745 B1

0. Preliminary caveats (read before relying on anything below)

Two of this patent's three face-of-the-patent references — US 5,867,706 (Martin et al.) and US 6,035,119 (Massena et al.) — I was unable to retrieve full text for before my search budget was exhausted. I therefore will not attribute specific disclosures to them. I flag this per your instruction not to fabricate. What I can ground:

Because I cannot read '706 and '119, my analysis is built on the art I can verify. This is a deliberate choice: an element-by-element §103 chart that cites references I cannot quote would be fabrication.

---

1. Governing framework

• Graham v. John Deere, 383 U.S. 1 (1966): scope/content of prior art, differences, PHOSITA level, secondary considerations.
• KSR Int'l v. Teleflex, 550 U.S. 398 (2007): a combination is obvious where the improvement is a "predictable use of prior art elements according to their established functions," where there is a "design incentive," or where the technique was "known" and available.
• PHOSITA (c. March 2000): I would propose — and this is my construction, not the patent's — a person with a bachelor's degree in CS/EE or equivalent and roughly two years of experience in web/database application development, HTTP, and distributed server systems, or equivalent research experience. This mirrors the level petitioners proposed in the parallel EpicRealm IPRs (IPR2015‑01734), and it is an enabling level for everything claimed here (HTML, CGI, JavaScript, HTTP POST, load balancers, symmetric multi-server farms were all mature by 1998–2000).

Critical structural point carried over from the prior section: granted claims 1, 7, and 10 recite transmitting the variable to a replication server — not appending the variable to a document and sending it to the client. The abstract and specification describe the append-to-HTML/client-return architecture. This means the closest prior art for validity is not the WebAgent help-button model (which the spec admits as background); it is the dispatcher / load-balancer / page-server model of Lowery '554, SWEB 95, and Farber '598. That materially strengthens the obviousness case.

---

2. Decomposition of the independent claims

Dependents of interest: 2 (variable is a JavaScript variable), 3 (markup-language document), 4 (reproduce from the variable and transmit to client), 5 (request includes a URL), 6 (document includes tags for an icon/hypertext that on activation sends the variable to another server), 8 (claim 7 + reproduce/transmit), 9 (claim 7 + URL), 11 (claim 10 + icon/hypertext to a customer-service-representative computer), 12 (retrieve template by URL, retrieve customization by identifier, generate, transmit), 13 (claim 12 + transmit to another client computer).

---

3. What the verified prior art already taught

(1) Farber '598 — replication-server selection (limitation 1(c), 7(c), 10(c)).
Farber discloses exactly the "identify the replication server from a plurality" step. Its reflectors "select a best repeater from a set of possible repeaters" and direct the request to it. That is server selection among replicas — the claim's "plurality of replication servers."

(2) Lowery '554 / EpicRealm '335 — dispatching a request to one of many page servers, and dynamic page generation (limitations 1(a), 1(d), 10).
'554 recites intercepting a web request at a web server, routing it via a dispatcher to one of several page servers, and having the page server dynamically generate the web page from data sources. The '554 specification (per the reexam record) frames "dynamic Web pages which contain content that is generated dynamically by retrieving the necessary requested data." This maps onto claim 10's "dynamically-generated document… customized based on customization information contained in the request."

(3) SWEB 95 + Bradley — selection among servers and caching of dynamically generated output.
The IPR record shows SWEB 95 discloses: request routing among processors, server selection based on the location of the file and on estimated CGI computational cost, and all files shared via NFS. Bradley adds caching the output of CGI programs as files. Together they disclose choosing a server from a plurality considering whether the dynamically generated output already resides there.

(4) Bookman / Leaf / Knudson — template + data population (limitations 12, and the "customization" concept).
Per the same reexam record: Bookman discloses "generating dynamic Web pages…by invoking and executing predefined procedural packages…retrieve data from a data repository, and then format[] the retrieved data as HTML output"; Leaf discloses transforming URL-form requests into HTML documents returned to the browser; Knudson teaches "custom HTML extension templates" into which retrieved data is inserted. This is precisely claim 12's URL → template; identifier → data; generate; transmit sequence.

(5) The admission in the patent's own Background.
The specification admits WebAgent existed and that a customer clicks a "help" button to start an interactive session in which the agent views the same pages (claim 6's icon; claim 15's interactive-session applets such as chat and whiteboard are described in the patent as known, referencing the "Web Interaction 2000 v. 2.0" planning guide). Applicant admissions are prior art for §103 purposes.

---

4. The combinations (and why a PHOSITA would make them)

Combination A (primary): Farber '598 + Lowery '554/EpicRealm '335 + SWEB 95/Bradley + a dynamic-page-generation reference (Bookman/Leaf/Knudson)

This combination renders claims 1, 7, 10, 12, and 13 obvious:

1. Lowery '554 supplies receiving the client request (with its POST payload / identifiers) and routing it to one of a plurality of page servers → 1(a), 1(c), and in claim 10's terms, receipt of the request that carries the customization data.
2. Bookman/Leaf/Knudson supply the URL→template→data→dynamic HTML generation → 10's "dynamically-generated document," claim 12 in full.
3. Farber supplies selecting a best server among replicas and directing the request to it → 1(c), 7(c), 10(c).
4. The variable of claim 1(b) is met by the request payload the dispatcher forwards — i.e., the set of identifiers + customization data that is "generated based on the request received" and that carries what a downstream server needs to (re)produce the customized page. SWEB 95's dispatcher forwards exactly such a self-contained request to the selected processor.

Motivation / KSR rationales:

• Predictable result: routing a request (with its payload) to a selected member of a server pool and having that member generate the page was the established function of the load-balancer/dispatcher architecture; doing it with a customized page is the same mechanism applied to a familiar workload.
• Known problem, known solution: the specification itself frames the problem as replicated-server farms in which the replication server "is unaware which of the multiple replicated servers was the source of the 'post' type dynamically-generated page." Once "reproduce the page at the selected server from the request data" is available (Bookman/Leaf), the problem is solved — with no more than ordinary skill.
• Design incentive / performance: SWEB 95's stated goal of minimizing per-request response time would motivate a POSITA to prefer the server whose CGI output is already cached (Bradley), and to hand the server the data it needs to reproduce the page.

Combination B: Farber '598 + Lowery '554 (for claim 14/15's session-server flow) + the WebAgent admission

Claim 14's four elements — receive a variable from a server computer; reproduce the customized document from it; transmit to two client computers — map onto the architecture the patent admits was prior art: the WebAgent server receives a help request and supplies the same page to the requesting customer and the service representative. The '554 dispatcher/server decomposition supplies "receive the variable from a server computer." Claim 15's applet-addition step is described in the specification as conventional ("applets to perform functions similar to those described in the Web Interaction 2000 v. 2.0 planning Guide"), i.e., the added subject matter is the ordinary, known way to implement the admitted WebAgent feature — a KSR "known technique for its known purpose" case.

Note the design-incentive rationale that is strongest here: the entire motivation for the variable is to let a different server than the source replica reconstruct the page. Once one accepts a client-mediated design (as the spec's abstract does), the two-recipient transmission of claim 14 follows directly from the admitted WebAgent use case (customer + representative view the same page).

Combination C: the face-of-record references '706 and '119 (insufficiently verified)

I will not assert that '706 + '119 render any claim obvious. I can only state: these two references, together with Farber '598, were applied/considered by the Examiner during prosecution of the application that issued as '745. If '706 and '119 concern (as their presence on the face suggests) web-server request handling and automatic generation/handling of web content, they could reinforce Combination A. That is a hypothesis requiring the full texts, not a conclusion.

---

5. Where elements are individually weak, and how the combinations close the gap

---

6. Rebuttal arguments I would expect, and responses

1. "The references are non-analogous / from different fields." Unlikely to prevail: SWEB 95, Farber '598, and Lowery '554 all sit in web-server scaling and content replication — the same field of endeavor and the same problem (the spec's own "problem statement" about replicated server farms). KSR and In re ICON Health & Fitness counsel against a cramped field limitation where the references address the same need.
2. "The Examiner allowed the '745 claims over Farber '598." That is itself evidence bearing on the scope of '598 as applied — but it is not dispositive, and a different combination (adding the dynamic-page-generation and CGI-output-caching art of Lowery/Bookman/Leaf/Bradley) may supply what '598 alone lacked. The AIA/ pre-AIA "all elements in a single reference" paranoia is not the standard; the standard is whether the combination is predictable.
3. "The variable must be generated by the same server that received the request (claim 1), which the dispatcher art does not show." This is the strongest non-obviousness argument, and it is thin: generating a variable from the received request is met by packaging the request's identifiers/customization data into the forwarded message — the ordinary function of a dispatcher. This is a labeling dispute, not a technical difference.
4. "Secondary considerations." I found no evidence of record of unexpected results, licensing for the claimed feature, or industry praise tied to '745. Note the patent expired May 16, 2020 (anticipated expiration; status "Expired – Lifetime"), so any 2026 assertion would face laches-type and damages-window problems in addition to §103.

---

7. Claim-by-claim bottom line

Single-reference anticipation (§102): I see no verified single reference that discloses all of claim 1's elements, and I would not argue §102 on the record I have. §103 is the viable theory.

---

8. Confidence, contradictions, and open items

1. Strongest confidence: Combination A rendering claims 1, 7, 10, and 12 obvious, because each element is individually attested by references I verified (Farber '598, Lowery '554, SWEB 95/Bradley, Bookman/Knudson/Leaf) and the combination yields nothing more than the predictable result of routing a customized-page request to a selected replica.
2. Weakest link: the exact metes and bounds of the "reproduction information" term. I have not found art that describes a self-contained, client-carried payload (as opposed to a server-forwarded request) reproducing the page — that is the spec's preferred embodiment. Claims 1/7/10 as granted do not require the client-carried form, which is why they remain vulnerable. If the patent were construed to require the client-carried variable, the §103 case would rest more heavily on claims 6/11 (admitted art) and the "known technique" rationale.
3. Flagged contradiction (carried forward, unresolved): the abstract and detailed description emphasize appending the variable to the HTML and transmitting the document to the client, whereas granted claims 1/7/10 recite transmitting the variable to a replication server. This looks like a claim set amended during prosecution away from the spec's framing. It cuts in favor of obviousness, because it shifts the comparison baseline to the dispatcher/replication art ('554, SWEB 95, Farber '598) rather than to the (admittedly prior-art) WebAgent client architecture.
4. Not verified / not asserted: the disclosures of US 5,867,706 (Martin et al.) and US 6,035,119 (Massena et al.). I state only that they are references of record on the '745 face. Any chart mapping them to limitations would be speculation on my current evidence.
5. Claims 16–29 remain unavailable to me (the supplied text cuts off mid-claim 15), so I cannot analyze whether later independents are obvious. I will not guess.
6. No secondary considerations located — and given the May 16, 2020 expiry, none are likely to be developed now.

Sources grounding this analysis: US 6,760,745 B1 claim text and specification (Google Patents, https://patents.google.com/patent/US6760745/en); US 6,185,598 B1 (Farber et al.) abstract and continuation-family record as retrieved; USPTO/PTAB and CourtListener reexamination and IPR2015‑01734 records quoting US 5,894,554 (Lowery), US 6,415,335 (EpicRealm), SWEB 95 (Andresen et al.), Bradley, Leaf, and Bookman; and the previously generated sections of this analysis (bibliographic data, claim overview, and litigation history), which I have built on rather than repeated.