Agricultural Technology IP Checklist: Farm Data Ownership, Equipment and Repair Access, Agronomic Model Terms, Grower and Cooperative Agreements, and Trial Data Provenance

By Casey Scott McKay · 2024-03-24

This checklist audits an agricultural technology position in the order the exposure sits, and it opens by setting aside the question the client will raise. Data ownership settles nothing, because facts about a field are not property; five contract terms settle everything, and almost nobody reads them. The checklist works through the data category map, those five terms, the model and calibration enumeration that protects the supplier's real asset, the university collaboration agreements sitting behind most agronomic science, and the trial provenance supporting every product claim. Later phases cover equipment interfaces and the repair position, imagery and third-party licences with their derived-work clauses, diligence readiness, production system variants, and the collective negotiation that is the only real leverage a small grower has. Gate items mark where work should stop.

IP and Technology > Information Technology | Checklist | Published 24 March 2024 - Updated 7 July 2025 | Casey Scott McKay - marksy.us

Summary. This checklist audits an agricultural technology position in the order the exposure sits, opening by setting aside the question the client will raise. Data ownership settles nothing, because facts about a field are not property; five contract terms settle everything and almost nobody reads them. It covers the data category map, those five terms, the model and calibration enumeration protecting the supplier's real asset, the collaboration agreements behind most agronomic science, and the trial provenance supporting every claim. Later phases cover repair positions, imagery licences, diligence readiness, production system variants, and collective negotiation.

Keywords: agtech checklist · data category map · five terms · licence scope · aggregation thresholds · export specification · change of control · calibration register · collaboration agreements · trial provenance · repair position · imagery derived works · cooperative negotiation · diligence readiness · production system variants

How to use this checklist

| Phase | What it produces | Who runs it | Gate | |---|---|---|---| | 1. Categories | A six-way data map | Counsel and engineering | No single "farm data" definition | | 2. Five terms | Two pages of drafted clauses | Counsel | Aggregation has a number in it | | 3. Models | A calibration register and access matrix | Counsel and data science | Separated from collaborator material | | 4. Collaborations | A register per project | Counsel and research partnerships | Retained licences identified | | 5. Trials | A claims substantiation file | Counsel and regulatory | Every claim mapped to a trial | | 6. Equipment | A repair and interface position | Counsel and field | Defensible parts separated | | 7. Imagery | An external data register | Counsel and product | Derived-work column completed | | 8. Diligence | A one-page ownership summary | Counsel | Maintained, not reconstructed | | 9. Systems | A note on which phases apply | Counsel | Adjusted for production type | | 10. Collective | A grower explainer and mandate | Counsel | Written once, given away |

The matter. An agronomy platform serving eleven hundred growers across four regions, licensing satellite imagery from two vendors, running calibration models developed with three universities, making four comparative yield claims in its marketing, integrating with three equipment manufacturers' telematics, and preparing for a financing round.

Phase 1. Map the data categories

[ ] Set the ownership question aside first. Agree that the grower owns the data, note that this settles nothing because facts are unprotectable under Feist Publications v. Rural Telephone Service, and move to the licence.
[ ] Machine data. Engine hours, fuel burn, fault codes, component wear. The manufacturer's claim is strongest; the grower's interest is repair access.
[ ] Agronomic data. Yield by location, moisture, soil samples, as-applied rates. The grower's claim is strongest and this is what the ownership debate concerns.
[ ] Prescription and recommendation data. Created by an advisor applying judgment, and an authored work in a way raw measurement is not.
[ ] Aggregated and benchmarked data. Derived from many growers, nobody's individually, and where the commercial value concentrates.
[ ] Imagery and remote sensing. Arrives under third-party licences with their own derived-work restrictions.
[ ] Trial data. A separate category with regulatory as well as commercial significance.
[ ] Record the client's position on each, since the drafting differs by category.
[ ] [Gate] A single "farm data" definition covering all six is the source of most disputes.

Phase 2. Draft or audit the five terms

[ ] Licence scope, enumerated positively. Operate the service; support equipment; produce analytics for this customer; improve products and models; produce aggregated outputs for others; supply data to identified third parties. Each is a separate permission and the standard drafting collapses them into "any purpose."
[ ] Close the list, stating that no other use is permitted, rather than granting broadly with carve-outs.
[ ] Aggregation definitions with three components: a minimum contributor threshold expressed as a number of distinct farms, a small-cell suppression rule, and a prohibition on attempted re-identification binding supplier and recipient.
[ ] Third-party sharing by named category: input suppliers, buyers, lenders, insurers, research institutions, government — with separate consent stated where required.
[ ] Export specification, not an export right. Format by reference to an industry standard, schema documentation, derived layers included or excluded, timeframe, and cost.
[ ] Test the export before signing, since a specification nobody has exercised is an aspiration.
[ ] Termination and change of control: what the customer receives, what the supplier may continue doing, retention period, and whether the customer may exit on a change of control of the supplier.
[ ] Draft for farmers, not lawyers. Short sentences, consistent defined terms, no cross-references, and a plain-language summary alongside.
[ ] [Gate] "Aggregated" without a number in it does commercial work it cannot support.

Phase 3. Enumerate the model and the calibration

[ ] Enumerate at artefact level: this model, these coefficients, this feature set, this regional calibration, this validation dataset, this tuning history.
[ ] Identify the actual moat. A model tuned to a specific soil type, climate, and cropping system over years is genuinely hard to replicate; the published agronomy underneath it is not.
[ ] Apply reasonable measures under 18 U.S.C. § 1839: access control, marking, segregation, and onboarding and exit discipline.
[ ] Separate published from unpublished, since a commercial model built on public science is protectable only as to its own contribution.
[ ] Check what circulates to collaborators, since the calibration is frequently stored alongside material shared with universities.
[ ] Address the improvement loop, since a platform permitted to train on user outcomes improves faster than one that is not — which makes the Phase 2 licence a technical decision as well as a commercial one.
[ ] Be realistic about patents. Recommendation engines and prescription methods face Alice Corp. v. CLS Bank International and 35 U.S.C. § 101; sensing hardware and machinery are different and worth filing under 35 U.S.C. § 103 analysis.
[ ] Record what a departing data scientist could reach, because that record is the claim under 18 U.S.C. § 1836.
[ ] [Gate] A calibration on an open shared drive is not a protected secret.

Phase 4. Read the collaboration agreements

[ ] Check publication rights, which universities require and which may reach material the client treats as secret.
[ ] Check retained licences, since institutions frequently keep a research-use licence the commercial partner has forgotten.
[ ] Check federally funded invention obligations, including disclosure, election of title, and government licence rights.
[ ] Check student and postdoctoral contributions, which raise ownership questions the agreement may not address.
[ ] Check material transfer terms where germplasm, soil, or biological material moved.
[ ] Establish what was published and when, which determines both secrecy and novelty.
[ ] Find the agreements, which live with individual scientists rather than in a contract system.
[ ] Renegotiate ongoing collaborations; document historic ones.
[ ] [Gate] No model is treated as wholly owned until the collaboration register is complete.

Phase 5. Build the trial provenance and claims file

[ ] Distinguish regulatory trials from marketing trials. The first support registration and attract data protection periods; the second support claims and must satisfy substantiation standards.
[ ] Record conductor, protocol, and raw data holder for each trial, since third-party providers hold the data and the terms rarely address publication.
[ ] Address grower-sited trials with consent, confidentiality, and a right for the grower to receive results from their own land.
[ ] Map every marketing claim to a supporting trial, and expect at least one gap.
[ ] Substantiate under 15 U.S.C. § 45, treating comparative claims against a named competitor as inviting challenge under 15 U.S.C. § 1125.
[ ] Record negative results and the publication decision, since a regulator or court may later ask what was known.
[ ] Protect the trial network and protocols as trade secrets, since these are what a small competitor cannot match.
[ ] Install a marketing gate: every comparative or numerical claim goes to regulatory before it publishes.
[ ] [Gate] No claim ships without a trial reference.

Phase 6. Set the equipment and repair position

[ ] Identify the firmware and diagnostic software position, since these channel repair to authorised dealers.
[ ] Assess circumvention exposure under 17 U.S.C. § 1201, tracking the rulemaking exemptions that have covered agricultural equipment on a renewable cycle.
[ ] Separate the defensible from the indefensible: safety-critical software and emissions compliance are one thing, ordinary mechanical repair another.
[ ] Concede the second early enough that the concession counts.
[ ] Address the memoranda of understanding that have partly displaced litigation, which are contractual rather than statutory.
[ ] Recognise that data access is bundled with repair access, converting a service question into an ownership one.
[ ] Address mixed-fleet interoperability, where implementing another supplier's format runs through Google LLC v. Oracle America, Inc..
[ ] Check post-sale restrictions against Impression Products, Inc. v. Lexmark International, Inc. and the repair line from Aro Manufacturing Co. v. Convertible Top Replacement Co..
[ ] Note the access statute at 18 U.S.C. § 1030 as construed in Van Buren v. United States for disputes about extracting data from equipment.
[ ] [Gate] Review the position annually against the exemption cycle.

Phase 7. Audit the imagery and third-party licences

[ ] List every external layer: satellite, aerial and drone, public earth observation, weather, soil survey, field boundaries, and commercial enrichment.
[ ] Record the derived-work clause per source, since an index computed from licensed imagery may or may not be distributable.
[ ] Check coverage, refresh, and redistribution terms, which platforms routinely exceed.
[ ] Distinguish public programmes supplying base layers free of copyright, where value is entirely in the processing.
[ ] Address drone capture, including the difficulty that a flight over one farm records the neighbours.
[ ] Reconcile to the strictest terms, since an engine consuming several sources inherits the tightest restriction.
[ ] Register the client's own imagery and processed layers in batches under 17 U.S.C. § 412, noting 17 U.S.C. § 411 conditions suit.
[ ] Take contractor assignments under 17 U.S.C. § 204, since 17 U.S.C. § 201 leaves ownership with the author.
[ ] [Gate] No product ships on a derived output the imagery licence does not permit.

Phase 8. Prepare for diligence and consolidation

[ ] Maintain a one-page ownership summary, updated rather than reconstructed.
[ ] Answer question one: does the aggregation licence support the products already sold? Which turns on Phase 2.
[ ] Answer question two: is the calibration protected? Which turns on Phase 3.
[ ] Answer question three: do the collaboration agreements leave the company owning what it thinks it owns? Which turns on Phase 4.
[ ] Answer question four: can customers leave? Which is the export specification, and which a buyer asks because churn risk is priced.
[ ] Expect acquisition, since the sector has consolidated harder than almost any adjacent one.
[ ] On the grower side, negotiate change of control at signature, since a platform chosen because it was independent may be acquired by an input supplier and the licence transfers with the business.
[ ] Watch the competition dimension, since data access has featured in merger analysis where a small number of suppliers hold both the inputs and the data about how they perform.
[ ] [Gate] The summary should be producible in a day, not assembled during a process.

Phase 9. Adjust for the production system

[ ] Arable. The default assumption behind Phases 1 to 8.
[ ] Livestock and dairy. Individual animal monitoring produces data about identified animals rather than land. The animal is property with a traceable identity, herd records are regulatory as well as commercial, breeding genetics is a distinct asset, equipment is more integrated, and treatment records raise veterinary and food safety questions.
[ ] Controlled environment. Closer to industrial process control. The growing recipe is the product, is genuinely proprietary, and is protectable as a trade secret because the environment is closed and the process invisible. This is the one part of the sector where conventional strategy applies straightforwardly.
[ ] Aquaculture. Animal-level monitoring plus process control, with environmental and siting regulation generating its own data obligations.
[ ] Traceability and provenance. Increasingly mandatory, with claims requiring substantiation and frequently certification through certification marks under 15 U.S.C. § 1054 with governance requirements.
[ ] Carbon and ecosystem services. A credit is a contractual construct rather than property in most jurisdictions; the proving data is farm data; baselines depend on exportable history; verification bodies see everything; reversal risk allocation is the commercial fight; and stacking is unresolved.
[ ] [Gate] Record which phases apply and which do not before quoting the work.

Phase 10. Advise the small grower and the collective

[ ] Recognise the imbalance. A family enterprise presented with standard terms has one decision — accept or decline — and declining means operating without tools its neighbours use.
[ ] Negotiate collectively where possible, since a cooperative or association obtains terms an individual cannot and one negotiation serves a thousand growers.
[ ] Ask for the three that cost the supplier nothing: export specification, change of control, and small-cell suppression. They are declined because nobody asks.
[ ] Prefer disclosure to prohibition. A grower who understands that participation means an input supplier sees their yield history may still participate; what is objectionable is that almost nobody knows.
[ ] Write the translation. A one-page plain-language explanation of the five terms, written once and given away, is the intervention with the widest reach in this sector.
[ ] [Gate] A grower who cannot state what the licence permits has not been advised.

Where agtech positions fail

[ ] The ownership argument consumed the budget, and the licence was never read.
[ ] "Aggregated" was never defined, and the benchmark products rest on a permission that may not support them.
[ ] The calibration sits on a shared drive alongside material circulated to collaborators.
[ ] A collaboration agreement has a retained licence nobody flagged, discovered in diligence.
[ ] A marketing claim has no trial behind it, or one whose provider's terms do not permit publication.
[ ] The platform exceeded an imagery derived-work clause building the product it now sells.
[ ] The grower cannot leave, because the export is an undocumented dump in a proprietary format.
[ ] A change of control moved the counterparty, and the data now sits with the input supplier the grower avoided.
[ ] Patents were filed on recommendation methods for two years before anyone ran the eligibility analysis.
[ ] Contractor-written integration code was never assigned, discovered when the contractor could no longer be found.

The documents an audit should produce on request

[ ] The data category map, six categories, with the client's position on each.
[ ] The five drafted terms, with the plain-language summary.
[ ] The aggregation specification, including the threshold number and suppression rule.
[ ] A tested export, with the schema documentation it produced.
[ ] The calibration register at artefact level, with the access matrix.
[ ] The collaboration register, per project, with publication rights and retained licences.
[ ] The claims substantiation file, mapping every claim to a trial.
[ ] The trial provenance records, with conductor, protocol, and data holder.
[ ] The repair and interface position paper, dated and reviewed against the exemption cycle.
[ ] The external data register, with the derived-work column completed.
[ ] The contractor assignment file.
[ ] The one-page ownership summary, current.

If a business can produce all twelve, the position is defensible. If it can produce the five terms and the calibration register, it is ahead of most of the sector. If it can produce only an assertion that the grower owns the data, it has been having the wrong conversation for years.

Drafting the five terms in detail

Because Phase 2 is where the whole audit lands, it deserves specification rather than description.

[ ] The licence scope should be an enumerated list, not a broad grant with carve-outs. A grower reading six enumerated permissions understands what has been agreed; a grower reading "any purpose related to the services" does not, and neither does a court in a dispute.
[ ] Distinguish operation from improvement. Operating the service for this customer and improving the supplier's models for all customers are different permissions with different commercial value, and collapsing them is the sector's most common drafting failure.
[ ] Distinguish aggregation from sale. Producing an aggregated benchmark available to other customers is one thing; supplying data or derived outputs to an input company is another, and growers react very differently to each.
[ ] The aggregation threshold should be a number of distinct farms, not a percentage, because a percentage in a thin region produces the same disclosure problem the threshold exists to prevent.
[ ] The suppression rule should be operational, meaning that outputs derived from fewer than the threshold are withheld automatically rather than reviewed case by case.
[ ] The re-identification prohibition should bind recipients, flowing down through any onward supply, since the supplier's obligation is worthless if its customers are unconstrained.
[ ] The third-party list should name categories, not describe them, so that a grower can see at a glance whether its lender or its grain buyer is included.
[ ] The export specification should name a standard format where the industry has one, and should state explicitly whether derived layers travel with the raw data.
[ ] Schema documentation should be required, since an undocumented file is not portability.
[ ] The export should be tested during negotiation, not after signature, because the gap between what a supplier promises and what its systems produce is where switching fails.
[ ] Termination should address four things: what the customer receives, what the supplier may continue doing, how long retention runs, and whether a change of control triggers a right to exit.
[ ] [Gate] Every one of the five should be readable by a farmer without a lawyer present.

Three audits worked through

A grower consolidating four platforms. A substantial arable enterprise has used four systems over a decade. Consolidating means extracting a decade of records, and the exercise reveals that two platforms export only to proprietary formats, one charges for bulk export, and one takes the position that derived layers it created are its own. None of that is a breach; all of it is what the agreements say. What can be exported is; derived layers are rebuilt from raw where raw survives; and the new agreement is negotiated with an export specification, a change-of-control right, and a termination obligation.

A platform preparing for a financing. Diligence asks whether the aggregation licence supports the benchmark products already sold, whether the calibration is protected, and whether the collaboration agreements leave the company owning what it thinks it owns. The answers are: not clearly, because "aggregated" was never defined; no, because the calibration sits on a drive shared with two universities; and partly, because one agreement retains a research licence broad enough to matter. The patent portfolio, six applications on recommendation methods, is the least valuable item and the only one the founders had considered.

A manufacturer facing a repair campaign. The legal questions are the exemption cycle, the enforceability of restrictions on diagnostic tool distribution, and the exposure of tying data access to repair access. Counsel's contribution is to separate the defensible parts — safety-critical software and emissions compliance — from the indefensible parts, and to concede the second early enough that the concession counts for something.

Working with the functions that hold the answers

[ ] Product and engineering know which data categories the platform actually holds and which it merely displays, and can produce the Phase 1 map faster than any contract review.
[ ] Agronomy and data science hold the model and the calibration and regard access controls as friction. The argument that lands is the departure scenario, not the compliance one.
[ ] Commercial and customer success own the grower relationship and resist anything making terms look more restrictive. Show them that positively drafted permissions read better than blanket reservations — true, and the easiest sale in this sector.
[ ] Research partnerships sit with a scientist rather than a lawyer, and the collaboration register is built by asking them rather than by searching files.
[ ] Regulatory affairs hold the trial data and understand data protection periods better than most counsel; the gap is that nobody has asked them about the marketing claims.
[ ] Marketing makes the claims and has never seen the trial file. The one-line gate is the highest-yield hour available.
[ ] Field and dealer organisations hold the repair position in practice, and their view of what growers actually need beats any policy discussion.
[ ] Finance will fund the programme if it is framed as diligence readiness rather than as compliance, which is accurate.

Timelines

[ ] Data category map: one week, produced by product and engineering.
[ ] Five terms: two to three weeks to draft, then applied to every agreement. Highest return, fastest delivery.
[ ] Calibration register: four to eight weeks, with the difficulty in persuading data science rather than in drafting.
[ ] Collaboration register: three to four weeks, most of it finding agreements.
[ ] Trial provenance file: four to six weeks for an active portfolio, and it usually finds a claim with no trial.
[ ] Imagery audit: two to three weeks, with the findings in the derived-work column.
[ ] Repair position paper: a fortnight, reviewed annually.
[ ] Who does it. One commercially minded technology lawyer with access to product, data science, and regulatory affairs. The failure mode is a patent counsel who never reads the platform terms and a commercial lawyer who never sees the model.
[ ] With no budget, do two things: the five terms and the calibration register. Five weeks, and together they decide whether the business is defensible.

Cross-border adjustments

[ ] Database rights exist in several jurisdictions, protecting investment in obtaining, verifying, or presenting contents independently of copyright — which reaches compilations unprotectable domestically and means a platform may have more to assert abroad than it realises.
[ ] Personal data rules apply where growers are individuals, which in many markets is most of them, bringing notice, basis, rights, and transfer obligations a corporate-farm assumption misses.
[ ] Repair legislation differs sharply and is moving, with some jurisdictions enacting statutory access rights and others relying on voluntary commitments.
[ ] Product registration and trial data protection periods vary, affecting both what must be submitted and what a competitor may rely upon.
[ ] Certification and provenance schemes are jurisdiction-specific, so a claim substantiated for one market may be unusable in another.
[ ] Carbon and ecosystem credit regimes differ fundamentally in whether a credit is property, a contractual right, or a regulatory permission.
[ ] Produce a two-page matrix: global standards, local instruments, and genuinely divergent analyses — with database rights and repair legislation dominating the third column.

First ten days, for a practitioner with other work

[ ] Day one: install the marketing gate. Every comparative or numerical claim goes to regulatory before publishing. An hour, and it stops the problem growing.
[ ] Day two: read the licence clause, not the ownership clause, in the client's standard terms.
[ ] Day three: ask how many farms sit behind the smallest published benchmark. If unknown, the anonymisation claim is untested.
[ ] Day four: ask where the calibration lives and who can open it. Usually a shared drive, usually everyone.
[ ] Day five: attempt an export as a customer would, and see what arrives.
[ ] Days six to seven: list the external data sources and pull the derived-work clause for each.
[ ] Day eight: ask research partnerships for the collaboration agreements, and expect them to be with individual scientists.
[ ] Day nine: pull the marketing claims and ask regulatory which have trials.
[ ] Day ten: write the one-page brief — what was found, what is fixable this quarter, and why the five terms come before the filings.

Auditing from each side of the table

[ ] The platform. Its instinct is to reserve everything in the licence and file patents on its recommendation logic — both wrong, in opposite directions. The broad reservation alarms customers and may be unenforceable; the patents will not issue. Redirect: enumerate the calibration, define the aggregation, draft the export specification the customer will eventually demand, and file only on genuinely technical sensing and machinery.
[ ] The grower. Its instinct is to argue about ownership, which it will win and which changes nothing. Redirect to the five terms, and be honest that leverage is limited unless the enterprise is large or negotiating collectively.
[ ] The equipment manufacturer. Its instinct is to defend the whole repair position, which is losing ground legislatively and reputationally. Redirect to separating the defensible from the indefensible.
[ ] The advisor or agronomist. Its instinct is to assume its recommendations are its own. They are, as authored works, unless the platform terms say otherwise — and frequently they do.
[ ] The cooperative. Its instinct is to treat the data programme as a member service. It is also a negotiating position of real weight, and this is the single highest-leverage client type in the sector and chronically under-advised.
[ ] The input supplier. Its instinct is to gather as much as possible and condition commercial terms on participation, both of which attract competition attention where market position is concentrated. Voluntary and separately compensated aggregation is more defensible than aggregation bundled into pricing.
[ ] [Gate] Identify which chair the client occupies before quoting the work, because the same ten phases produce different priorities from each.

What good looks like

[ ] The data category map exists and the agreement does not treat six different things as one.
[ ] The five terms are drafted properly, in plain language, with a summary alongside.
[ ] "Aggregated" has a number in it, plus a suppression rule and a re-identification prohibition.
[ ] The export specification has been tested, not merely drafted.
[ ] The model and calibration are enumerated at artefact level, access-controlled, and separated from collaborator material.
[ ] Collaboration agreements have been read, with publication rights and retained licences recorded per project.
[ ] Every marketing claim maps to a trial, with a recorded protocol and a known data holder.
[ ] The imagery register has a derived-work column, and the product respects it.

Businesses with those eight survive diligence, retain customers who read their terms, and can act when a data scientist leaves. Businesses without them own an argument about ownership and very little else.

A closing note on sequencing

The sector spent a decade answering the wrong question and building an industry on the answer. Everyone agrees the grower owns the data; nothing follows, because facts about a field are not property.

What does exist is five contract terms — two pages properly drafted, three lines in most agreements — deciding what the supplier may build, what the grower may take away, who else sees a farm's operations, and what happens when the counterparty changes.

On the other side of the table the insight inverts. If the data is not property, the supplier's defensible assets are the model, the regional calibration, the integration estate, and the trial network, and almost nobody has enumerated any of them.

Work the phases in order where there is time. Where there is a fire, start where it is burning and come back — but come back, because a repair dispute, a diligence process, and a departing scientist all eventually ask the same question, which is what the client actually owns.

The carbon and ecosystem services annex

[ ] Treat a credit programme as a data licence with a payment attached, and apply the five terms to it.
[ ] Establish the credit's legal character, since in most jurisdictions it is a contractual construct rather than property, which determines assignability, security, and what happens on reversal.
[ ] Recognise the proving data as farm data, subject to every phase of this checklist, with aggregator terms frequently drafted by parties who have never seen a grower agreement.
[ ] Check the baseline dependency, since additionality and permanence require historic records and a grower unable to export from a previous platform cannot demonstrate the baseline.
[ ] Read the verification body's confidentiality obligations rather than assuming them, since verifiers see everything.
[ ] Allocate reversal risk expressly: if the carbon returns to the atmosphere, who bears the loss and for how long does the obligation run.
[ ] Address stacking, since whether one practice can generate carbon, water quality, and biodiversity credits simultaneously is a market convention question with no legal answer.
[ ] Check the registry's terms, since double counting is the integrity issue and the registries addressing it are private bodies with rules of their own.
[ ] Record which answers are convention rather than authority, because a grower signing a twenty-year commitment over land they farm deserves to know which parts are settled.
[ ] [Gate] No long-term credit commitment is signed without the baseline data secured and exportable.

[ ] Diary the four events that change this advice: a repair statute passing, a credit registry changing its rules, a database right decision in a market the client sells into, and a change of control at a platform the client depends on. None will be announced to counsel by the client, and a calendar reminder is cheaper than an annual review nobody has budgeted for.
[ ] Re-run the smallest-benchmark question annually, since a region that was thin when the threshold was set may have grown and one that was viable may have shrunk as growers consolidate.

[ ] Re-test the export annually as a customer would, since platforms change formats and a specification that worked at signature may not describe what the system now produces.
[ ] Re-run the collaboration register whenever a new research partnership begins, because these accumulate quietly and each one carries publication rights that reach the calibration.

[ ] Refresh the imagery register whenever a new data source is added, since sources are added by engineers and the derived-work clause is checked by nobody.
[ ] Keep the whole audit as a living register rather than a report, because a report describes a moment and this sector changes faster than reports are reissued.

[ ] Name an owner for the register, since one that belongs to nobody stops being current within a season.

Key Authorities at a Glance

| Authority | Phase | |---|---| | Feist v. Rural Telephone | 1 — facts are unowned | | 17 U.S.C. § 102 | 1, 7 — code, imagery, and prescriptions as works | | 17 U.S.C. § 103 | 1 — compilations and derived datasets | | 17 U.S.C. § 106 | 7 — reproduction and derivative rights | | 17 U.S.C. § 201 | 7 — contributor ownership default | | 17 U.S.C. § 204 | 7 — the signed assignment | | 17 U.S.C. § 411 | 7 — registration before suit | | 17 U.S.C. § 412 | 7 — timely registration and remedies | | 17 U.S.C. § 1201 | 6 — access controls and repair exemptions | | Google LLC v. Oracle America | 6 — interfaces and mixed fleets | | Impression Products v. Lexmark | 6 — post-sale restrictions | | Aro Manufacturing v. Convertible Top | 6 — repair versus reconstruction | | 35 U.S.C. § 101 | 3 — eligibility for recommendation engines | | Alice Corp. v. CLS Bank | 3 — the abstract idea framework | | 35 U.S.C. § 103 | 3 — obviousness in machinery | | KSR v. Teleflex | 3 — combination obviousness | | 35 U.S.C. § 112 | 3 — enablement across conditions | | Amgen v. Sanofi | 3 — enabling a claimed range | | 18 U.S.C. § 1839 | 3 — reasonable measures over calibration | | 18 U.S.C. § 1836 | 3 — the claim on departure | | 18 U.S.C. § 1030 | 6 — extracting data without permission | | Van Buren v. United States | 6 — authorised access, narrowed | | 15 U.S.C. § 45 | 5 — substantiation of yield claims | | 15 U.S.C. § 1125 | 5 — competitor challenges | | 15 U.S.C. § 1054 | 9 — certification marks for provenance | | 15 U.S.C. § 1 | 8 — conditioning terms on data contribution | | FRCP 26 | 5 — discovery into trial data |

Search the underlying materials directly for farm data licence aggregation definition, agricultural equipment repair exemption, agronomic model trade secret university, crop protection trial data protection period, and precision agriculture platform export portability.