Additive Manufacturing IP Checklist: File Ownership, Patent and Design Coverage, Bureau Terms, Digital Controls, and Takedown Practice

By Casey Scott McKay · 2026-06-21

This checklist builds an additive manufacturing IP position in the order that eliminates work rather than the order that feels doctrinal. Phase one settles who owns the CAD files, because an ownership defect makes every later step pointless. Phase two asks what protection survives a competitor's 3D scan, which disposes of the copyright-centred strategies most programmes start with. The middle phases cover filings and registration, the licence and bureau stack that does the real controlling, the instrumentation that converts an unprovable leak into a named contract claim, and the trade secret controls that protect process parameters separately from geometry. The closing phases handle export and regulatory screening, evidence channels, and a four-bucket enforcement triage that keeps copyright takedown forms away from patent problems.

IP and Technology > Patent Counseling Transactions | Checklist | Published 21 June 2026 - Updated 20 July 2026 | Casey Scott McKay - marksy.us

Summary. This checklist builds an additive manufacturing IP position in the order that eliminates work rather than the order that feels doctrinal. Phase one settles who owns the CAD files, because an ownership defect makes every later step pointless. Phase two asks what protection survives a competitor's 3D scan, which disposes of the copyright-centred strategies most programmes start with. The middle phases cover filings and registration, the licence and bureau stack that does the real controlling, the instrumentation that converts an unprovable leak into a named contract claim, and the trade secret controls that protect process parameters separately from geometry. The closing phases handle export and regulatory screening, evidence channels, and a four-bucket enforcement triage that keeps copyright takedown forms away from patent problems.

Keywords: additive manufacturing checklist · CAD ownership audit · confirmatory assignment · design patent filing rule · collection registration · print bureau terms · quality annex · serialisation · file watermarking · copyright management information · process parameter controls · export classification · warranty return channel · enforcement triage · takedown sorting · parts lifecycle policy · escrow release conditions · clean room records · file register · bureau audit

How to use this checklist

| Field | Detail | |---|---| | Who runs it | IP counsel with engineering, operations, procurement, and the export function | | When | Before the first bureau engagement; on each new part family; annually thereafter | | Time required | Three to four weeks for a first pass over a twenty-part family | | Gates | Ownership confirmed; design filings made before publication; bureau terms executed | | Output | A file register, executed assignments and bureau agreements, and an enforcement triage log | | Companion documents | Managing IP in an Additive Manufacturing Programme and Printing the Part |

The matter. An industrial equipment manufacturer with a thirty-year installed base is standing up a digital inventory programme for spare parts. Most of its CAD work over the last decade was done by two outside design bureaus with purchase orders that say nothing about ownership. Marketing published renders of the new range four months ago. Its European distributor has asked for production files in escrow. Two print bureaus already hold files under a one-page NDA. A marketplace listing has appeared offering a printed version of a visible door handle, and the enforcement team has drafted a copyright takedown notice. Nobody has classified any file for export.

Phase 1. Settle who owns the files

[ ] List every modelling relationship of the last seven years. Employees, secondees, contractors, design bureaus, university collaborators, and customers who supplied geometry. Why. Ownership is determined by who created the file under what paper, and no other step matters if this one fails. Trap. Starting with the current supplier and assuming history is clean.
[ ] Confirm employee status where it is assumed. Why. A work prepared by an employee within the scope of employment is a work made for hire under 17 U.S.C. § 101; Community for Creative Non-Violence v. Reid supplies the agency test. Trap. Long-term contractors treated internally as staff, with no assignment on file.
[ ] Assume contractor files are not works made for hire. Why. Commissioned works qualify only within nine enumerated categories and only with a signed writing; a standalone part model fits none of them comfortably. Trap. A purchase order reciting "work for hire" and nothing else, which transfers nothing when the category test fails.
[ ] Check the assignment verb. Why. "Hereby assigns" transfers on signature; "agrees to assign" leaves title with the assignor pending performance. Trap. A template that uses the second form and has done so for years.
[ ] Confirm patent rights are covered, not only copyright. Why. An outside modeller who contributes to conception may be a joint inventor under 35 U.S.C. § 116, and joint owners may practise and license without accounting under 35 U.S.C. § 262. Trap. A copyright-only assignment leaving a co-inventor free to license a competitor.
[ ] Require native files, not exported meshes. Why. A programme holding only STL exports cannot revise its own parts and has bought a snapshot rather than an asset. Trap. Accepting delivery without opening the file in the native application.
[ ] Run a confirmatory assignment campaign over historic engagements. Why. It is cheap, it is uncontroversial while relationships are good, and it becomes impossible once they are not. Trap. Waiting until diligence forces it.
[ ] [Gate] Record ownership status per part family in the file register before any file leaves the building. Why. The register is the index for every later step. Trap. Ownership tracked in the contract system rather than against parts.

Phase 2. Ask what survives a scan

[ ] Assume the competitor buys the part and scans it rather than stealing the file. Why. This is what actually happens, and it eliminates the rights most programmes rely on. Trap. Building a strategy around file copyright.
[ ] Accept that a faithful scan carries no copyright. Why. Meshwerks v. Toyota Motor Sales held exact digital models of existing objects lack originality, applying Feist Publications v. Rural Telephone Service. Trap. Threatening a scanner with copyright infringement and receiving a declaratory judgment complaint.
[ ] Accept that the printed functional object is usually not copyrightable. Why. The useful article exclusion in 17 U.S.C. § 101 and the separability test in Star Athletica v. Varsity Brands leave functional geometry unprotected. Trap. Confusing protection of the drawing with protection of the article.
[ ] Accept that drawing copyright does not reach the article depicted. Why. 17 U.S.C. § 113(b), the modern form of Baker v. Selden. Trap. The single most common misunderstanding in this field.
[ ] Identify what does survive: utility patents, design patents, marks on the part, and trade secret in the process. Why. These are indifferent to how the geometry was obtained. Trap. A protection memo that never mentions design patents.
[ ] [Gate] For each of the top twenty parts, record which surviving right applies. If the answer is "none", say so. Why. Knowing a part is unprotected is a commercial decision input, not a failure. Trap. Recording aspiration rather than status.

Phase 3. File and register

[ ] Adopt a bright-line design filing rule. Visible in use, or shape not wholly dictated by mating geometry, or distinctive surface treatment, means file. Why. 35 U.S.C. § 171 protection is cheap and fast, infringement is tested by the ordinary observer under Egyptian Goddess v. Swisa, and 35 U.S.C. § 289 awards total profit. Trap. Case-by-case debate that resolves as "not this time" every time.
[ ] Check the publication clock before filing. Why. 35 U.S.C. § 102(b) gives a twelve-month domestic grace period and absolute-novelty jurisdictions give none; a published render is a disclosure. Trap. Marketing renders posted before the design application.
[ ] Review design filings against LKQ Corp. v. GM Global Technology Operations. Why. Design obviousness now follows the flexible KSR International v. Teleflex approach, so trivially varied designs are more vulnerable. Trap. A portfolio of near-identical filings that looks impressive and invalidates together.
[ ] Register the CAD collection with the Copyright Office. Why. Registration is a precondition to suit under 17 U.S.C. § 411, and Fourth Estate v. Wall-Street.com requires completed examination; registration opens 17 U.S.C. § 504 statutory damages and fees. Trap. Discovering the registration requirement after finding a byte-identical copy online.
[ ] Record part numbers as trademarks where they identify source. Why. It converts copying into counterfeiting, with 15 U.S.C. § 1117 statutory damages and 15 U.S.C. § 1116 seizure available. Trap. Treating part numbers as purely descriptive without testing actual use.
[ ] [Gate] No part family is released to a bureau or storefront before its filings are made or a documented decision not to file is recorded. Why. Release is publication in practice. Trap. Pilot programmes exempted from the gate.

Phase 4. Build the licence and bureau stack

[ ] Draft the grant as a right to produce units, not a right to use a file. Why. Unit count is the commercial term; a file-use licence is silent on how many parts exist. Trap. Software-style licence templates applied to physical production.
[ ] Specify part and revision, material specification, machine class, territory of production, and territory of installation. Why. These are separately meaningful and separately breached. Trap. A single "territory" definition that does not distinguish where the part was made from where it went.
[ ] Address modification expressly. Why. Customers adapt parts; manufacturers do not want to warrant adapted parts. Trap. Silence, which is read as permission by one side and prohibition by the other.
[ ] Prohibit reverse modelling from printed output, not only file copying. Why. This is the route a licensee takes to become a competitor without touching a file. Trap. A no-copying clause that only mentions the file.
[ ] Require serial-level reporting, record retention, and an audit right with a cost-shifting trigger. Why. Reporting is what makes metering real. Trap. An audit right with no notice mechanics and no consequence.
[ ] Carve confidentiality and IP indemnities out of the liability cap and require insurance certificates before first transmission. Why. An uncapped indemnity from a small bureau is a paragraph, not a risk transfer. Trap. Certificates requested after the incident.
[ ] Write a quality annex. Inspection, acceptance, non-conformance, destruction of rejects, and record retention. Why. Rejected prints leaving the facility are a leading source of unauthorised genuine-looking parts. Trap. Quality handled in a separate document nobody links to the IP terms.
[ ] Resist unqualified residuals language. Why. It permits a bureau to use anything its people remember, which is most of what you gave them. Trap. Accepting a residuals clause because it appears in the counterparty's standard form.
[ ] [Gate] No file transmission before the bureau agreement and the quality annex are executed. Why. Files sent under a one-page NDA cannot be recovered contractually. Trap. Pilot jobs run on an NDA "just to test the process".

Phase 5. Instrument the programme

[ ] Serialise every printed unit. Why. It ties each part to a licence, a bureau, and a job ticket, converting an unprovable leak into a named contract claim. Trap. Retrofitting traceability onto a deployed population, which does not work.
[ ] Watermark the file. Why. Steganographic marking survives format conversion better than metadata and links a surfaced file to a recipient. Trap. Relying solely on filename or header metadata that any exporter strips.
[ ] Choose a format that carries ownership information. Why. STEP and 3MF can hold copyright management information within 17 U.S.C. § 1202; removal or alteration is independently actionable with statutory damages. Trap. Distributing STL because it is convenient and then having no metadata claim.
[ ] Log every download and access event. Why. It is both a detection channel and evidence of reasonable measures. Trap. Logs retained for thirty days, which is shorter than the time it takes to notice a leak.
[ ] Prefer streamed jobs to transmitted files where the technology allows. Why. A file that never persists locally cannot be retained after termination. Trap. Treating the delivery method as an IT decision.
[ ] Authenticate at the machine, not only at the portal. Why. A job ticket bound to a machine identity and a unit count is enforceable; a downloaded file plus a promise is not. Trap. Portal-level controls presented as digital rights management.

Phase 6. Protect the process, not only the geometry

[ ] Separate process parameters from geometry in storage and in contracts. Why. Build orientation, support strategy, energy density, scan strategy, and post-processing schedules are usually the more valuable secret. Trap. Parameters travelling in the same package as the model, to the same recipients.
[ ] Test the programme against 18 U.S.C. § 1839. Why. Reasonable measures are an element, not a best practice. Trap. A policy document with no access controls behind it.
[ ] Include the 18 U.S.C. § 1833 whistleblower notice in agreements. Why. Omission forfeits exemplary damages and fees against that employee under the Defend Trade Secrets Act. Trap. Notice in the handbook but not in the agreement.
[ ] Extend NDAs to the counterparty's subcontractors. Why. Bureaus subcontract capacity routinely, particularly at peak. Trap. A confidentiality clause binding only the named entity.
[ ] Run exit forensics for anyone with production file access. Transfer logs, cloud sync records, device imaging. Why. CAD files are small, portable, and look like ordinary work product. Trap. Exit process triggered by seniority rather than access.
[ ] Say who owns qualification and test data. Why. For regulated parts it is often worth more than the geometry, and customers and bureaus both claim it. Trap. Silence, resolved later in the counterparty's favour.

Phase 7. Export, safety, and regulatory screening

[ ] Classify every file for export before it enters a distributed network. Why. A CAD file for a controlled item is controlled technical data under the EAR or the ITAR. Trap. Classifying the finished part and not the file.
[ ] Screen bureaus by personnel nationality and system location, not merely facility address. Why. A deemed export happens where a foreign national gains access, wherever the printer sits. Trap. A supplier questionnaire that asks only for a registered office.
[ ] Treat published files as a publication decision. Why. Defense Distributed v. U.S. Department of State shows how sharply publication of controlled technical data is contested. Trap. Engineering posting sample files to a public repository for marketing reasons.
[ ] Check whether printing makes the customer a manufacturer. Why. Point-of-care and patient-matched printing can trigger device obligations under the Federal Food, Drug, and Cosmetic Act. Trap. Licensing a hospital to print without addressing regulatory status.
[ ] Confirm qualification requirements for safety-critical parts. Why. In aerospace, rail, and marine, a part is only lawful in service if machine, material, process, and records all qualify. Trap. A licence that permits printing on "equivalent" equipment.

Phase 8. Build the evidence channels before you need them

[ ] Monitor marketplaces and repositories against part numbers, product names, geometry keywords, and render image hashes. Why. Detection is the hardest part of enforcement here. Trap. Monitoring brand names only, which misses every unmarked replica.
[ ] Route non-genuine warranty returns to legal. Why. A failed part that is dimensionally right and materially wrong identifies an unauthorised source and arrives with a customer, a serial, and a dealer attached. Trap. Non-genuine returns sent straight to scrap.
[ ] Record customs recordations and keep them current. Why. Printed counterfeits crossing a border are seizable like any other goods. Trap. Expecting customs to intercept a file.
[ ] Commission failure analysis on suspect parts. Material testing, dimensional comparison, root cause. Why. It converts a commercial complaint into a safety case, which regulators, marketplaces, and courts treat differently. Trap. Asserting safety risk without test data.
[ ] Preserve the chain on test buys. Documented purchase, packaging photographs, shipping records, sealed retention. Why. Additive cases turn on physical exhibits. Trap. An engineer's disassembled sample with no provenance.

Phase 9. Triage enforcement by right

[ ] Sort every target into one of four buckets before acting. Why. The most common error in this field is a copyright notice for a patent problem. Trap. A single enforcement template.
[ ] Bucket one — file copies, derivatives, or stripped metadata. Send a 17 U.S.C. § 512 notice; consider a 17 U.S.C. § 1202 claim. Why. This is genuine copyright infringement. Trap. Failing to document the byte-level or structural comparison.
[ ] Bucket two — parts bearing your marks or part numbers. Counterfeiting practice under 15 U.S.C. § 1114. Why. Platform and seller-identification machinery applies directly. Trap. Treating it as a copying case and losing the statutory damages.
[ ] Bucket three — unmarked replicas of patented articles. Patent notice letter identifying patent and claims. Why. It also builds the knowledge element for inducement under Global-Tech Appliances v. SEB; Commil USA v. Cisco Systems forecloses the invalidity-belief answer. Trap. Sending a copyright takedown, exposing the company to 17 U.S.C. § 512(f) and to the reputational response Lenz v. Universal Music Corp. makes predictable.
[ ] Bucket four — independently modelled commodity geometry. Take no action. Why. Assertion here produces misuse and invalidity attacks that contaminate the parts that matter. Trap. An enforcement queue prioritised by ease rather than value.
[ ] Do not sue end users. Why. The direct infringers are the installed base. Trap. A test case against a customer that becomes the story.
[ ] [Gate] Every notice is signed off against its bucket, with the right identified and the evidence attached. Why. It is the control that prevents the misrepresentation problem. Trap. Delegating notice-sending to a vendor with volume incentives.

Phase 10. Governance and review

[ ] Name a single owner of the programme. Why. Additive sits across engineering, operations, procurement, and legal, and unowned programmes produce inconsistent contracts and unfiled designs. Trap. A steering committee with no accountable individual.
[ ] Maintain the file register. Part, revision, native location, registration, filings, bureaus holding it, licences granted, serialisation scheme, export classification. Why. It is the deliverable that makes diligence, enforcement, and underwriting straightforward. Trap. A spreadsheet last updated at launch.
[ ] Publish a parts-lifecycle policy. When files are released, on what terms, to whom. Why. It sets the terms of the repair conversation before a legislature does. Trap. An unwritten practice of ignoring discontinued parts.
[ ] Exercise one bureau audit a year and document it. Why. Unexercised rights are not reasonable measures. Trap. Audit rights negotiated hard and never used.
[ ] Report metrics to the board. Parts under design protection, bureaus audited, reported versus expected unit counts, enforcement actions by bucket. Why. Additive IP is invisible until it fails. Trap. Reporting filing counts alone.

Phase 11. If you are the one printing

[ ] Clear the part, not the file. Why. 35 U.S.C. § 271(a) asks whether the article reads on a claim and is indifferent to how it was made. Trap. A clearance memo that analyses file provenance and never charts a claim.
[ ] Search design patents by classification and image, not by text. Why. Design patents are invisible to keyword search and carry 35 U.S.C. § 289 total-profit exposure that can exceed the margin on the whole line. Trap. A utility-only freedom-to-operate search presented as complete.
[ ] Ask whether the original article was sold. Why. Impression Products v. Lexmark International exhausts patent rights on an authorised sale regardless of post-sale restrictions, and foreign sales exhaust too. Trap. Reading a printed single-use restriction as legally operative.
[ ] Run the repair analysis before conceding anything. Why. Aro Manufacturing v. Convertible Top Replacement permits replacement of worn unpatented components of a purchased patented combination. Trap. Settling a claim that repair doctrine would have defeated outright.
[ ] Never start from a downloaded file. Why. The moment a third-party file enters the process, independent creation is unavailable and the case becomes substantial similarity. Trap. An engineer using a repository model "as a reference".
[ ] Keep a clean-room record. Separate examiner from modeller, record the specification handed across, retain measurement data. Why. Sega Enterprises v. Accolade and Sony Computer Entertainment v. Connectix show the documented process is what carries the defence. Trap. A clean room asserted after the fact with no contemporaneous records.
[ ] Strip marks and part numbers from scanned geometry. Why. Reproducing them because they were on the original converts a defensible parts business into counterfeiting under 15 U.S.C. § 1114. Trap. A moulded-in logo nobody noticed in the mesh.
[ ] Obtain a written opinion where the exposure is material. Why. Opinions remain relevant to wilfulness and the enhanced damages discretion described in Halo Electronics v. Pulse Electronics. Trap. Relying on an oral clearance from an engineer.
[ ] Document acquisition of every competitor sample. Invoice, date, seller, condition. Why. Provenance of the examined article is contested routinely. Trap. Samples bought on a personal account with no record.

Phase 12. Diligence, insurance, and transactions

[ ] Ask for the file register first in any acquisition. Why. Its existence and currency say more about the programme than any warranty. Trap. Accepting a data room folder of files as evidence of a programme.
[ ] Trace ownership on a sample of ten commercially important parts. Why. Ownership defects cluster; a two-in-ten failure rate means the population fails. Trap. Testing the parts the seller selected.
[ ] Compare design filings against the visible product range. Why. Forty visible parts and three registrations is quantified aftermarket exposure. Trap. Counting filings without mapping them to parts.
[ ] Read bureau agreements for change-of-control termination and residuals. Why. Both can strand or devalue the acquired programme on day one. Trap. Reviewing only the customer contracts.
[ ] Schedule every escrow arrangement as an encumbrance. Why. Files in escrow are contingent disclosures of the asset being bought. Trap. Escrow treated as a service arrangement rather than an IP encumbrance.
[ ] Confirm export classification exists for the file library. Why. An unclassified library in a distributed network is a compliance liability an indemnity does not cure. Trap. Classification of products presented as classification of technical data.
[ ] Check the insurance position honestly. Why. General liability advertising injury grants typically reach copyright and trade dress and exclude patent; specialty IP policies underwrite on the strength of assignment and registration records. Trap. Assuming the general policy responds to the claims this field actually generates.
[ ] Look at the professional services exclusion in bureau policies. Why. Printing to a customer specification may be characterised as a professional service, and the exclusion can remove cover exactly where the exposure sits. Trap. Accepting a certificate without reading the form.

Phase 13. The community question

[ ] Separate enthusiasts, non-commercial repositories, and commercial sellers before acting. Why. Only the third is usually a business problem, and treating all three alike guarantees the worst outcome. Trap. An enforcement queue sorted by ease of takedown.
[ ] Consider publishing a non-commercial repair tolerance. Why. It costs almost nothing, forecloses a category of reputational risk, and preserves every argument against commercial sellers. Trap. An unwritten practice that varies by who receives the complaint.
[ ] Give a reason when you do act against a community file. Why. Safety in unqualified material is a reason people accept; ownership alone is not, and produces mirrors within days. Trap. A form letter with no explanation.
[ ] Consider publishing official files for cosmetic and non-critical parts. Why. It converts an adversarial relationship into a marketing asset and gives a documented answer on repairability. Trap. Publishing without a licence limiting commercial use.
[ ] Keep the enforcement queue proportionate to loss. Why. Time on hobbyist files is time not spent on the seller shipping thousands of units in the wrong polymer. Trap. Measuring the programme by notices sent.

Phase 14. Working the example matter

[ ] Deal with the design bureau purchase orders first. Why. Two bureaus with a decade of modelling work and no ownership language means the manufacturer may own almost none of its own geometry. Confirmatory assignments while relations are good cost days; the same conversation after a dispute costs the programme. Trap. Assuming payment implies ownership, which it does not in copyright.
[ ] Accept the loss on the published renders and reprice the range. Why. Renders published four months ago have started the domestic clock under 35 U.S.C. § 102(b) and defeated absolute-novelty rights abroad. Domestic design applications are still available within the grace period; file them now. Trap. Filing abroad anyway and paying for rights that will fall on examination or in opposition.
[ ] Recall the files held under the one-page NDA. Why. Those bureaus hold geometry with no unit metering, no reporting, no destruction obligation, and no quality annex. Replace the NDA with a full agreement or terminate access and require certified deletion. Trap. Leaving the pilot arrangement in place because production depends on it, which converts a temporary lapse into the operating model.
[ ] Negotiate the distributor escrow carefully rather than refusing it. Why. Escrow is a reasonable customer demand for long-life equipment; the risk is in the release conditions. Tie release to discontinuation of supply or insolvency, define the field of use narrowly, and require confidentiality that survives release. Trap. A release condition triggered by "material breach", which any commercial dispute can be argued into.
[ ] Stop the copyright takedown on the door handle listing. Why. A visible external door handle is exactly the part that should have a design patent, and a copyright notice against a printed functional article is a misrepresentation exposed to 17 U.S.C. § 512(f). Check whether a design registration exists; if it does, send a patent notice; if it does not, this is a bucket-three or bucket-four target. Trap. Sending the notice because the form was already drafted.
[ ] Classify the file library for export before the digital inventory programme goes live. Why. Industrial equipment for regulated end uses frequently carries controlled technical data, and a distributed bureau network is a deemed-export exposure from the first transmission. Trap. Classifying at go-live, after files have already moved.
[ ] Write the parts-lifecycle policy now, not after the first enquiry. Why. A thirty-year installed base guarantees a large discontinued-parts population and therefore an aftermarket. The policy decides whether that aftermarket is licensed or unlicensed. Trap. Deferring the decision, which decides it in favour of the unlicensed option.
[ ] [Gate] Do not release the first part family to the storefront until ownership, filings, licence terms, serialisation, and export classification are recorded against it. Why. Five checkboxes, once per family, prevent every failure described above. Trap. Exempting the launch family because the launch date is fixed.

Phase 15. Annual review

[ ] Re-run the ownership sample. Ten parts, traced from file to assignment to filing. Why. New engagements accumulate defects at a steady rate, and an annual sample catches drift before it becomes a population problem. Trap. Reviewing only new suppliers.
[ ] Compare reported unit counts against expected demand. Why. A bureau reporting materially fewer units than the installed base consumes is the clearest signal of unreported production available. Trap. Accepting reports without a demand model to test them against.
[ ] Audit one bureau properly. Facility, machines, material lots, file handling, deletion evidence, subcontracting. Why. One documented audit a year is the cheapest available evidence of reasonable measures under 18 U.S.C. § 1839. Trap. A questionnaire returned by email described as an audit.
[ ] Review the design portfolio against the current range. New parts released without filings, filings maintained on discontinued parts. Why. Portfolios drift out of alignment with products within two release cycles. Trap. Renewing everything by default.
[ ] Re-classify for export after any change to the network or the product. Why. Classification is a point-in-time conclusion that changes with end use, destination, and personnel. Trap. Treating the initial classification as permanent.
[ ] Review the enforcement log by bucket. Why. A log dominated by bucket one and four suggests the programme is chasing what is easy rather than what matters. Trap. Reporting volume rather than composition.
[ ] Refresh the parts-lifecycle policy against current repair legislation. Why. Statutory duties to supply parts, tools, and documentation are expanding, and whether documentation reaches production files is unsettled. Trap. A policy written once and never revisited.

Phase 16. Documents this checklist should produce

[ ] A file register. One row per part family: revision, native file location, ownership status, registration number, design and utility filings, bureaus holding it, licences granted, serialisation scheme, export classification, and review date. Why. Every other output is a view of this table. Trap. Three partial registers maintained by three functions.
[ ] An executed assignment pack. Current template plus confirmatory assignments for historic engagements, indexed to the register. Why. It is the first thing a buyer, an insurer, or a court asks for. Trap. Signed copies stored in a contract system with no link to the parts.
[ ] A bureau agreement with a quality annex, executed before any transmission, with insurance certificates attached. Why. It is the instrument that actually controls printing. Trap. A framework agreement with the commercial annex left "to be agreed".
[ ] A parts-lifecycle policy, published, stating when files are released, on what terms, and to whom. Why. It sets the repair conversation on your terms. Trap. An internal draft never published.
[ ] An enforcement triage log, one row per target, recording the bucket, the right asserted, the evidence, and the sign-off. Why. It is the control that keeps copyright notices away from patent problems. Trap. A log started after the first complaint.

Outcome. A file register that states, for every commercially significant part, who owns the model, what rights subsist, which bureaus hold it, on what licence terms, with what serialisation and export classification — and an enforcement queue in which each item is sorted by the right that actually applies.

The five things people get wrong

One. Building the strategy on copyright in the CAD file. It is the most intuitive right and the least useful one. It does not reach the article depicted, it does not survive an independent scan, and it is not what a competitor takes. Copyright in the file protects against one thing — someone distributing your actual file — and that is the least common form of loss in the field.

Two. Publishing renders before filing designs. Marketing publishes images as a matter of routine, and every published image of an unfiled design forfeits absolute-novelty rights abroad and starts the domestic clock under 35 U.S.C. § 102(b). The fix is one line in the launch checklist, and the loss is permanent.

Three. Sending files under an NDA and calling it a programme. A one-page NDA does not specify unit counts, materials, machines, reporting, audit, destruction, or quality. Files sent that way cannot be metered, cannot be traced, and cannot be recovered. Every mature programme has a story about the pilot job that was run "just to test the process".

Four. Using copyright takedowns for patent problems. It is fast, the form exists, and the platform complies — which is exactly why it happens. It is also a misrepresentation exposed to 17 U.S.C. § 512(f), it forfeits the good-faith posture Lenz v. Universal Music Corp. requires, and in maker communities it generates mirrored files faster than the notice removes them.

Five. Ignoring discontinued parts and then complaining about the aftermarket. When supply stops, owners obtain parts elsewhere. A manufacturer that declined to supply and said nothing has surrendered the safety argument, the goodwill, and the legislative narrative. A published parts-lifecycle policy with a licensed route costs little and preserves all three.

Key Authorities at a Glance

| Authority | Proposition | |---|---| | 35 U.S.C. § 101 | Patentable subject matter | | 35 U.S.C. § 102 | Novelty; grace period; publication as disclosure | | 35 U.S.C. § 112 | Written description and enablement | | 35 U.S.C. § 116 | Joint inventors | | 35 U.S.C. § 171 | Design patents | | 35 U.S.C. § 262 | Joint owners practise without accounting | | 35 U.S.C. § 271 | Direct and indirect infringement | | 35 U.S.C. § 284 | Damages; reasonable royalty | | 35 U.S.C. § 289 | Total profit for design infringement | | 17 U.S.C. § 101 | Useful article; work made for hire | | 17 U.S.C. § 102 | Originality; systems excluded | | 17 U.S.C. § 113 | Drawings of useful articles | | 17 U.S.C. § 411 | Registration precondition | | 17 U.S.C. § 504 | Statutory damages | | 17 U.S.C. § 512 | Notice, takedown, and misrepresentation | | 17 U.S.C. § 1201 | Circumvention and trafficking | | 17 U.S.C. § 1202 | Copyright management information | | 15 U.S.C. § 1114 | Registered mark infringement | | 15 U.S.C. § 1116 | Injunctions and seizure | | 15 U.S.C. § 1117 | Damages and counterfeiting statutory damages | | 15 U.S.C. § 1125 | False designation; trade dress | | 18 U.S.C. § 1833 | Whistleblower immunity notice | | 18 U.S.C. § 1836 | DTSA civil action | | 18 U.S.C. § 1839 | Trade secret definition; reasonable measures | | 37 C.F.R. § 1.84 | Patent drawing standards | | Baker v. Selden | Description is not the system | | Gorham Manufacturing v. White | Ordinary observer | | Mazer v. Stein | Art in a useful article | | Feist Publications v. Rural Telephone Service | Originality standard | | Community for Creative Non-Violence v. Reid | Employee status | | TrafFix Devices v. Marketing Displays | Functionality bar | | Wal-Mart Stores v. Samara Brothers | Product design secondary meaning | | KSR International v. Teleflex | Flexible obviousness | | Microsoft Corp. v. AT&T Corp. | Abstract software not a component | | Global-Tech Appliances v. SEB | Inducement knowledge | | Commil USA v. Cisco Systems | Invalidity belief no defence | | Life Technologies v. Promega | Single component not substantial portion | | Samsung Electronics v. Apple | Article of manufacture | | Star Athletica v. Varsity Brands | Separability | | Fourth Estate v. Wall-Street.com | Examination required | | Impression Products v. Lexmark International | Exhaustion on authorised sale | | Aro Manufacturing v. Convertible Top Replacement | Permissible repair | | Egyptian Goddess v. Swisa | Design infringement standard | | LKQ Corp. v. GM Global Technology Operations | Design obviousness | | Meshwerks v. Toyota Motor Sales | Exact models lack originality | | Lenz v. Universal Music Corp. | Good-faith consideration before notice | | Sega Enterprises v. Accolade | Intermediate copying | | Defense Distributed v. U.S. Department of State | Technical data export control | | Additive part qualification | Qualification regimes | | Right to repair documentation | Statutory supply duties |