esign: ink-reproduction consent gate + patent-risk research
Consent gate (the legal linchpin from the wet-signature memo): - migration 092 adds ink_consent/ink_consent_at/ink_consent_text to esign_records - extract pure, unit-tested gate logic into esign-ink-consent.ts (DRY single source for route + signing page): isInkReproduction / inkConsentRequired / inkConsentSatisfied + verbatim client-safe INK_CONSENT_TEXT - portal-esign-generic.ts: GET surfaces ink_reproduction + consent text; POST gates DRAWN signatures on ink-path docs on explicit consent, stores it - signing page locks the signature block until consent is checked (drawn only) - npi_provider marks cms855/cms10114 esign metadata ink_reproduction=true - 33 unit checks: gate truth table + consent text omits all internal mechanics (plotter/machine/CMS/MAC/etc) and keeps required legal reassurances Patent-risk memo (docs/legal/patent-risk-mechanical-wet-signature.md): - prior-art-dated risk analysis (autopen 1803/1942, plotters, CNC = public domain => low risk on core concept; e-sign workflow space litigious) - firsthand recent-grant sweep (1.58M USPTO grants 2021-2025, queried via DuckDB): ZERO patents on machine-applies-signature-in-ink; e-sign players hold only electronic-workflow patents. Not an FTO; flags where attorney search is needed
This commit is contained in:
justin
parent
f8d2a7f01f
commit
a4bad723bc
7 changed files with 452 additions and 5 deletions
214
docs/legal/patent-risk-mechanical-wet-signature.md
Normal file
214
docs/legal/patent-risk-mechanical-wet-signature.md
Normal file
|
|
@ -0,0 +1,214 @@
|
|||
# Patent-risk analysis: remote signature capture + mechanical wet-ink reproduction
|
||||
|
||||
Status: internal patent-RISK research. **Not legal advice and not a freedom-to-
|
||||
operate (FTO) opinion.** A real FTO requires a patent attorney to search live
|
||||
claims and read them against our implementation. This memo identifies where risk
|
||||
is low (mature/expired prior art) vs. where a targeted FTO search is warranted.
|
||||
|
||||
Companion docs: `docs/legal/remote-mechanical-wet-signature-precedent.md`,
|
||||
`docs/plans/remote-wet-signature-products.md`.
|
||||
|
||||
What we do (the accused "system," broken into components):
|
||||
1. Capture a hand-drawn signature in a browser as **vector strokes** (x, y, t).
|
||||
2. Store it + an audit trail (`esign_records`), send a JWT no-login signing link.
|
||||
3. Generate/fill a PDF form; stamp a digital signature image (existing path).
|
||||
4. **Reproduce the captured strokes in real ink on paper with a pen plotter /
|
||||
Line-us arm** (the novel-feeling part).
|
||||
5. Mail/file the document; track fulfillment.
|
||||
|
||||
## TL;DR
|
||||
|
||||
- **Components 1-3 and 5** are ubiquitous, decades-old techniques (signature
|
||||
capture, e-sign workflow, PDF fill, mailing automation). **Patent risk: low**,
|
||||
but the *e-signature workflow* space is litigious (DocuSign, RPost, etc.), so a
|
||||
targeted FTO on any **specific workflow feature** is prudent before scaling.
|
||||
- **Component 4** (drawing a signature in ink with a machine) sits on **very old,
|
||||
expired prior art**: signature-duplicating machines (Hawkins polygraph, 1803),
|
||||
autopens (1930s-1942), pen plotters (1950s-1980s), CNC/G-code motion (1950s+).
|
||||
The *idea* of a machine writing a signature is firmly in the public domain.
|
||||
**Patent risk on the core concept: low.**
|
||||
- The realistic risk is **not** the broad concept but **narrow, recent claims**
|
||||
on a *specific combination/method* (e.g. "capture an e-signature online, then
|
||||
transmit it to a robotic pen to apply wet ink for a regulated filing"). A
|
||||
firsthand title sweep of 1.58M recent grants (2021-2025) found **no such
|
||||
patent**; the e-sign players (DocuSign/Adobe/Silanis) hold only *electronic*
|
||||
workflow patents. This reduces, but does not eliminate, the need for an
|
||||
attorney FTO (which must also cover pre-2021 and pending claims).
|
||||
|
||||
## 1. Signature capture (vector strokes in a browser)
|
||||
|
||||
- **Prior art is overwhelming and old.** Electronic signature pads (Topaz,
|
||||
Wacom), tablet/stylus capture, and HTML5 `<canvas>` signature capture have been
|
||||
standard for 15-25+ years. Capturing (x, y, t) stroke data is basic digitizer
|
||||
technique predating the web.
|
||||
- **Risk: low** for the act of capturing strokes. Any patent broad enough to
|
||||
cover "capture a signature on a touchscreen" would almost certainly be invalid
|
||||
over prior art. Avoid copying a *specific patented capture algorithm* (e.g. a
|
||||
particular biometric/pressure-analysis method) - we use plain stroke capture.
|
||||
|
||||
## 2. E-signature workflow + audit trail + no-login link (JWT)
|
||||
|
||||
- **This is the litigious zone of the three software components.** Companies with
|
||||
e-signature/e-delivery patent portfolios have actively asserted them:
|
||||
- **RPost** (Registered Email / e-signature patents) has a history of
|
||||
licensing/litigation around proof-of-delivery and e-signature.
|
||||
- **DocuSign** holds many patents on signing *workflows*, "envelopes," routing,
|
||||
and tamper-evidence.
|
||||
- **But:** post-*Alice v. CLS Bank* (573 U.S. 208 (2014)), many abstract
|
||||
"do-it-on-a-computer" software/business-method patents are vulnerable to
|
||||
invalidation under 35 U.S.C. § 101. Generic e-sign workflow claims have been
|
||||
narrowed/invalidated.
|
||||
- **Risk: low-moderate.** Our workflow (JWT link, store signature + IP/UA/time,
|
||||
resume a pipeline) is generic and well-trodden. The risk is asserting a
|
||||
*specific patented feature* (e.g. a particular tamper-evidence/seal method, a
|
||||
specific routing/orchestration claim). **Do a targeted FTO before building any
|
||||
distinctive workflow feature** (sequential multi-party routing, a proprietary
|
||||
audit-seal, etc.).
|
||||
|
||||
## 3. PDF generation / form fill / digital stamp
|
||||
|
||||
- **Ubiquitous, mature.** Filling AcroForm fields and overlaying an image onto a
|
||||
PDF is standard library functionality (pypdf, reportlab, PDFBox, etc.) used by
|
||||
millions of applications. **Risk: low.**
|
||||
|
||||
## 4. Mechanical wet-ink reproduction (pen plotter / robotic arm)
|
||||
|
||||
This is the component that *feels* novel but is built on the oldest prior art:
|
||||
|
||||
- **Signature-duplicating machines: 1803.** John Isaac Hawkins patented the
|
||||
"polygraph" pantograph signature copier (used extensively by Jefferson).
|
||||
- **Autopen: 1930s-1942.** The "Robot Pen" (1930s) and De Shazo's commercially
|
||||
successful autopen (1942, Navy order) mass-reproduced signatures in ink. By
|
||||
mid-century there were 500+ autopens in Washington.
|
||||
- **Pen plotters: 1950s-1980s.** X-Y pen plotters (Calcomp 1959; HP 7470A 1980s)
|
||||
draw arbitrary vector paths in ink - exactly our motion problem.
|
||||
- **CNC / G-code motion control: 1950s+.** Numerically-controlled multi-axis
|
||||
motion (which a CR-10/3D printer is) is decades old.
|
||||
- (Sources: Wikipedia "Autopen," retrieved firsthand; well-documented hardware
|
||||
history.)
|
||||
- **Conclusion:** "a machine that applies a person's signature in ink by
|
||||
following a stored path" is **firmly in the public domain.** No one can hold a
|
||||
valid patent on the *concept*. **Risk on the core mechanism: low.**
|
||||
|
||||
Residual risk is only on **narrow modern improvements**, e.g.:
|
||||
- A *specific* signature-reproduction algorithm (pressure/velocity modeling to
|
||||
mimic natural handwriting) that someone recently patented. We use plain
|
||||
fit-to-box stroke replay; avoid copying a specific patented "natural-motion"
|
||||
method without checking.
|
||||
- A *specific combined method claim* tying online e-signature capture to a
|
||||
robotic ink applicator for filings. We found no such patent, but this is the
|
||||
one area an attorney FTO should specifically probe (handwriting-robot startups,
|
||||
e.g. companies in the "robot-written mail" space, may hold method patents on
|
||||
*their* pipelines - those would be narrow to their implementations).
|
||||
|
||||
## 5. Mailing / fulfillment automation
|
||||
|
||||
- **Mature and generic.** Batch mailing, address routing, print-and-mail
|
||||
workflows are old and widely practiced. **Risk: low.** (Note RPost-style
|
||||
proof-of-delivery patents exist; we are not claiming a proprietary e-delivery
|
||||
receipt mechanism.)
|
||||
|
||||
## Recent-grant prior-art sweep (2021-2025, firsthand query)
|
||||
|
||||
To probe the one real risk zone (a recent method patent tying online signature
|
||||
capture to robotic ink application, and e-sign workflow patents our flow might
|
||||
read on), we ran a firsthand title search against a full USPTO grant dataset
|
||||
(PatentsView-derived, **1,575,344 patents granted 2021-06 to 2025-09**). This is
|
||||
a **title-level** sweep of recent grants, **not** a claims-level FTO, and it does
|
||||
not cover pre-2021 art - but it directly tests "did anyone recently patent our
|
||||
specific combination?" The answer is essentially no:
|
||||
|
||||
- **Signature + (ink|pen|stylus|wet) + (robot|machine|apparatus|device): 0
|
||||
relevant hits.** All matches were unrelated (wireless-comms "signatures,"
|
||||
printer ink-color flushing). **No "machine applies a signature in ink" patent
|
||||
appears in 5 years of grants.**
|
||||
- **autopen: 0 hits. signature+plotter: 0. signature+reproduce: 0.**
|
||||
- **"writing/drawing robot": only ornamental DESIGN patents** (D1090657,
|
||||
D1058630 - cover a device's *appearance*, not a signing method) plus an
|
||||
unrelated Yaskawa robot-withdrawal utility patent. Nothing signature-specific.
|
||||
- **"handwriting + machine": only handwriting *extraction* (reading) via ML**
|
||||
(12175799), i.e. the opposite of writing.
|
||||
- **E-signature workflow space is owned by the expected players, all purely
|
||||
electronic (no ink/paper):**
|
||||
- **DocuSign, Inc.** - many recent grants on document models, recipient
|
||||
routing/notification, OCR, contract generation (e.g. 12411896 "Document
|
||||
graph"; RE50043 associating third-party content with online signing).
|
||||
- **Adobe Inc.** - electronic-signature *capture UI* (11132105, 11159328,
|
||||
11750670 "signature collection within an online conference").
|
||||
- **Silanis Technology, Inc.** - 11093652 "Web-based method and system for
|
||||
applying a legally enforceable signature on an electronic document" (the
|
||||
single closest hit by name) - but it is **purely electronic**; it does not
|
||||
apply ink to paper, so it does not read on our mechanical reproduction.
|
||||
- **Tamper/seal:** 11743053 "Electronic signature system and tamper-resistant
|
||||
device" - electronic; relevant only if we built a proprietary tamper-seal
|
||||
(we do not).
|
||||
|
||||
**Interpretation:** the recent-patent landscape is consistent with the prior-art
|
||||
analysis above. The litigious e-sign players hold **electronic-workflow** patents
|
||||
(read them before building any distinctive workflow feature), but **no one holds
|
||||
a recent patent on reproducing a captured signature in ink with a machine for a
|
||||
filing** - our specific, novel-feeling combination. This *reduces* (does not
|
||||
eliminate) the residual risk flagged in section 4.
|
||||
|
||||
**Caveats:** (1) title-only, recent-grant (2021-2025) - misses pre-2021 patents
|
||||
(many e-sign core patents predate 2021) and any in-force patent whose title does
|
||||
not contain these words; (2) does not read claims; (3) does not cover pending
|
||||
applications. **Still not an FTO** - a patent attorney must read live claims
|
||||
(including pre-2021 and pending) against our implementation before launch.
|
||||
|
||||
Method: queried the patent grant parquet directly via DuckDB (title LIKE
|
||||
filters). Dataset coverage and the specific hits above were retrieved firsthand
|
||||
2026-06-07.
|
||||
|
||||
## What would actually create exposure (avoid these without FTO)
|
||||
|
||||
1. Copying a **named competitor's specific feature** (a particular tamper-seal,
|
||||
a specific multi-party routing UX, a specific "robot handwriting realism"
|
||||
algorithm).
|
||||
2. Implementing a **distinctive method** and marketing it as such (the more
|
||||
specific/novel our claim of novelty, the more likely it overlaps a method
|
||||
patent - and the more attractive we are as a target).
|
||||
3. Building the **RON / e-notary** layer without checking notary-tech patents.
|
||||
|
||||
## Recommendations
|
||||
|
||||
1. **Now (low cost):** keep the implementation **generic and built on
|
||||
public-domain techniques** (plain stroke capture, standard PDF libs, standard
|
||||
G-code/CNC motion, autopen-class reproduction). This is our best defense - we
|
||||
are practicing decades-old art.
|
||||
2. **Before scaling / raising / public launch:** commission a **freedom-to-operate
|
||||
search** from a patent attorney focused on two questions:
|
||||
(a) any live method patent tying **online e-signature capture → robotic ink
|
||||
application** for documents/filings; (b) any e-signature **workflow** patent
|
||||
our specific features might read on.
|
||||
3. **Defensive posture:** document our **prior-art basis** (this memo + the
|
||||
hardware history) so that if anyone asserts a broad patent, we can argue
|
||||
invalidity/non-infringement. Consider a **defensive publication** of our
|
||||
pipeline (a public description) to create prior art preventing others from
|
||||
patenting the same combination later.
|
||||
4. **Do not patent-troll-bait:** avoid over-claiming novelty in marketing; frame
|
||||
the service as "we file your documents with an original ink signature," not as
|
||||
a patented invention.
|
||||
|
||||
## Honest limits of this memo
|
||||
|
||||
- We ran a **title-level sweep of recent grants (2021-2025)** (see section above)
|
||||
but did **not** run a full **claims-level** search across all years; reading
|
||||
live claims against our code requires a patent attorney. The recent-grant sweep
|
||||
supports the low-risk *concept* assessment but is **not an FTO clearance.**
|
||||
- Patent risk is jurisdiction- and claim-specific and changes as patents issue;
|
||||
re-check before any major launch.
|
||||
|
||||
## Sources
|
||||
|
||||
- Prior-art hardware history (autopen 1803/1930s/1942, plotters, CNC):
|
||||
Wikipedia "Autopen" (retrieved firsthand 2026-06-07), corroborated by
|
||||
general hardware history.
|
||||
- *Alice Corp. v. CLS Bank Int'l*, 573 U.S. 208 (2014) - § 101 abstract-idea
|
||||
invalidity framework (well-known; confirm citation before reliance).
|
||||
- Litigious-player context (RPost, DocuSign patent portfolios): general industry
|
||||
knowledge plus the firsthand recent-grant sweep above (DocuSign/Adobe/Silanis
|
||||
hits) - confirm specifics with counsel before relying.
|
||||
- Recent-grant prior-art sweep: USPTO patent grant dataset (PatentsView-derived,
|
||||
2021-2025, 1.58M patents), queried firsthand via DuckDB 2026-06-07. Specific
|
||||
patent numbers cited above are real hits from that dataset.
|
||||
Loading…
Add table
Add a link
Reference in a new issue