Cultural Heritage Digitisation Checklist: Rights Assessment, Diligent Search Records, Access Tiers, Community Consultation, and Takedown

By Casey Scott McKay · 2025-04-22

This checklist runs a cultural heritage digitisation programme from collection scoping to annual reporting, in the order that reduces the work rather than the order that feels logical. It front-loads the public domain determination and the deed of gift audit because both remove material from the harder analysis entirely. It then builds the tiered diligent search record, the written fair use assessment, and the accessibility programme that carries two independent legal bases. Phases five through seven assign access tiers, structure the rights metadata, and stand up the takedown operation that resolves most claims before they become disputes. It closes on community consultation, governance, and the five numbers worth reporting.

IP and Technology > Copyright | Checklist | Published 22 April 2025 - Updated 16 June 2025 | Casey Scott McKay - marksy.us

Summary. This checklist runs a cultural heritage digitisation programme from collection scoping to annual reporting, in the order that reduces the work rather than the order that feels logical. It front-loads the public domain determination and the deed of gift audit because both remove material from the harder analysis entirely. It then builds the tiered diligent search record, the written fair use assessment, and the accessibility programme that carries two independent legal bases. Phases five through seven assign access tiers, structure the rights metadata, and stand up the takedown operation that resolves most claims before they become disputes. It closes on community consultation, governance, and the five numbers worth reporting.

Keywords: digitisation checklist · collection scoping · public domain determination · renewal research · deed of gift audit · diligent search record · tiered search standard · fair use assessment · section 108 preservation · accessible formats · access tier assignment · rights metadata · standardised rights statements · community consultation · cultural protocols · takedown policy · claim handling · board approval · funder conditions · programme metrics

How to use this checklist

| Field | Detail | |---|---| | Who runs it | Digitisation lead, with the collections manager and the named rights owner | | When | Before scanning begins on any collection, and annually thereafter | | Time required | Two to six weeks per collection for phases one to four; ongoing thereafter | | Gates | Deed position confirmed; written assessment approved; consultation concluded; takedown live | | Output | A per-collection rights file, an assigned access tier, and structured metadata | | Companion documents | Running a Digitisation and Access Programme and The Rights You Cannot Trace |

The matter. A regional historical society holds forty thousand items across eighteen accessioned collections: glass plate negatives from a commercial studio that closed in 1961, the papers of a state senator including four decades of incoming correspondence, a run of a defunct weekly newspaper, oral history recordings from a 1970s community project, and photographs documenting a residential school. A grant will fund digitisation of all of it. The board wants to know whether they can put it online. Nobody has read a deed of gift since the accessions were made.

Phase 1. Scope the collection and run the public domain sweep

[ ] Define the unit of assessment. Assess by accessioned collection, not by item and not by programme. Why. Item-level assessment does not scale to forty thousand objects and programme-level assessment is too coarse to be honest. Trap. A single accession containing materially different categories — the senator's papers hold correspondence, photographs, printed ephemera, and audiovisual material — needs sub-units.
[ ] Inventory by format and date range. Count items, note formats, and record the earliest and latest creation dates in each sub-unit. Why. The date range determines which duration rules apply and which sweeps are worth running. Trap. Estimating a date range from the collection title. Sample the boxes.
[ ] Identify what is published and what is unpublished. Why. 17 U.S.C. § 303 governs unpublished works and the terms differ sharply from published material. Trap. Assuming that everything in an archive is unpublished. Newspapers, trade catalogues, postcards, and commercial studio photographs distributed to sitters were published.
[ ] Run the expiry sweep first. Anything published before the current public domain cutoff is clear on date alone. Why. Cheapest determination available; needs only a date.
[ ] Run the no-notice sweep. Examine objects for a copyright notice. Published material distributed without notice during the relevant period entered the public domain. Why. Highest visual-inspection yield in the entire programme; the determination is made by looking. Trap. Checking only the front of a photograph. Check versos, mounts, and packaging.
[ ] Run the renewal sweep. Search Copyright Office renewal records for works in the renewal window under 17 U.S.C. § 304. Why. Most works in that window were never renewed, which makes this the highest-yield research task available. Trap. Treating it as a background task. Budget it as a project with a deliverable.
[ ] Identify federal government works. Why. Not subject to copyright. Trap. Extending this to contractors, or to state and municipal government, neither of which is covered.
[ ] Segregate foreign-origin material. Why. Golan v. Holder upheld restoration of copyright in foreign works that had lost protection through formalities defects, and restoration was automatic rather than registered. Trap. Applying the no-notice shortcut to a foreign photograph. It does not work.
[ ] Calculate terms for anonymous and unidentified-author unpublished works. Why. These carry fixed terms from creation that do not require identifying anyone, which converts a large part of an "orphan" backlog into a calculation. Trap. Filing an unattributed 1930s snapshot as an orphan when its term is determinable.
[ ] Record each determination with its basis and date. Why. So the work is done once and is auditable. Trap. Recording the conclusion without the reasoning; the next cataloguer will redo it.
[ ] [Gate] Report the public domain yield before proceeding. Why. It usually removes a substantial share of the collection from every subsequent phase, and it changes the scope of the phases that follow.

Phase 2. Audit the deeds of gift

[ ] Locate the deed for every collection in scope. Why. It determines what the institution may do, and it is the single most consequential document in the programme. Trap. Assuming physical custody carries copyright. It does not.
[ ] Categorise each deed into assigned, licensed, or nothing granted. Why. Three positions, three workflows. Trap. Recording that a deed exists without reading its operative language.
[ ] Verify assignments satisfy 17 U.S.C. § 204. Why. A transfer of copyright ownership requires a signed writing. Trap. Treating an accession receipt or a covering letter as an assignment.
[ ] Record the scope of any licence. Research use, exhibition, non-commercial purposes, term, territory. Why. The licence sets the ceiling on the access tier. Trap. Reading "for the Society's purposes" as unlimited.
[ ] Ask whether the donor owned what they purported to give. Why. A family donating a relative's papers holds the objects; copyright in incoming correspondence belongs to its writers and copyright in studio photographs belongs to the studio. Trap. Accepting an assignment from a donor who had nothing to assign and treating the collection as cleared.
[ ] Extract every non-copyright restriction. Embargoes, access limitations, attribution requirements, conditions on characterisation, and any requirement to consult the donor. Why. These bind contractually and no fair use analysis reaches them. Trap. Publishing material under a live embargo because the copyright analysis came out clean.
[ ] Flag deeds that are missing or ambiguous. Why. An absent deed is a different risk posture from a deed granting nothing. Trap. Treating "we cannot find it" as "there were no restrictions."
[ ] Sample rather than exhaustively review legacy deeds. Why. Volume. Trap. Sampling a heterogeneous accession and generalising from the sample.
[ ] [Gate] Confirm the deed position per collection before scanning. Why. Discovering a restriction after digitisation wastes the scanning budget and creates a live problem.
[ ] Fix the deed template prospectively. Assign copyright where the donor holds it, take a broad licence where they do not, address digitisation and online access expressly, state what happens to material the donor did not own, and refuse restrictions the institution cannot operationally honour. Why. Every future accession is cheap to get right and expensive to fix. Trap. Accepting a restriction to close a donation and discovering the collections system cannot express it.

Phase 3. Build the diligent search record

[ ] Understand what the search does and does not do. Why. There is no United States orphan works safe harbour; a search confers no permission. It finds owners, it supports the fair use analysis on market effect, and it evidences the reasonable grounds that support remission of statutory damages under 17 U.S.C. § 504(c)(2). Trap. Telling a board the search makes the programme legal.
[ ] Set the tiered standard in writing before searching. Why. Effort should scale to intended use and risk. Trap. An undocumented standard that varies by whoever is doing the work.
[ ] Apply tier one to everything. Institutional records, finding aids, the object itself and its markings, Copyright Office online records for post-1978 material, and a general web search on any name present. Why. Fast, and it resolves more than expected.
[ ] Apply tier two before open online access. Renewal records, collective licensing databases, publisher and successor research, trade directories, and contact with known heirs, estates, or professional associations. Why. This is the level that supports a public posting.
[ ] Apply tier three to commercially valuable material and anything to be licensed out. Genealogical and probate research, correspondence with related repositories, direct outreach, and published notice of intent. Why. The exposure justifies the cost. Trap. Applying tier three uniformly and running out of budget in the first collection.
[ ] Use documented sampling for homogeneous collections. Search a defined sample, record the method and the yield, extrapolate, and search item-level only where the sample flags an identifiable creator. Why. It is the only mechanism that scales. Trap. Sampling a collection that is not homogeneous.
[ ] Record sources consulted, dates, search terms, results, and reasoning. Why. A record of what was done is worth far more than an assertion that a search was diligent. Trap. A one-line note reading "searched, no owner found."
[ ] Set a re-search interval for high-value orphans. Why. Records improve and a failed search may succeed later. Trap. Treating an orphan determination as permanent.
[ ] Log every owner found and what happened next. Why. Permission beats every exception, and the conversion rate justifies the search budget. Trap. Finding an owner and never contacting them because the workflow had no next step.

Phase 4. Write the fair use assessment

[ ] Write one assessment per collection, before publication. Why. It is what supports the programme and what a board approves. Trap. A single programme-wide assessment that does not engage with any actual collection.
[ ] Work all four factors of 17 U.S.C. § 107 explicitly. Why. A conclusory assessment is not evidence of reasonable grounds. Trap. Writing three sentences about transformation and stopping.
[ ] Apply Andy Warhol Foundation v. Goldsmith honestly. Why. It narrowed transformative use by directing attention to shared purpose and substitution. Trap. Asserting that digitisation is transformative in itself. Display of an archival photograph so people can look at it serves the photograph's original purpose.
[ ] Lean on the search and accessibility holdings. Authors Guild v. HathiTrust and Authors Guild v. Google are the strongest available ground. Why. Search genuinely serves a purpose different from reading, and both courts said so. Trap. Reading them as authorising full public display.
[ ] Address unpublished status where relevant. Why. It weighs against fair use under the second factor without barring it, particularly for historical documentary material. Trap. Ignoring it because the outcome is inconvenient.
[ ] Separate amount copied from amount displayed. Why. Whole-work copying can be necessary for preservation and search while display remains limited. Trap. Conceding the fourth factor by displaying everything you copied.
[ ] Make the fourth factor the centre of the analysis. Commercial unavailability, absence of a licensing market, and the documented failed search. Why. This is where orphan material is strongest. Trap. Sweeping in a subset of commercially exploited works and destroying the analysis for the whole collection.
[ ] State exclusions expressly. Why. An assessment that carves out what it does not cover is more credible than one that pretends the collection is uniform.
[ ] [Gate] Have the named rights owner approve and date it. Why. An unapproved assessment is a draft, and a draft is not a record of reasonable grounds.
[ ] Set a review trigger. New licensing market, a claim, a change in the case law, or a change in access tier. Why. The assessment describes conditions that change.

Phase 5. Section 108, accessibility, and the access tiers

[ ] Use 17 U.S.C. § 108 for preservation copying only. Why. It permits preservation and replacement reproduction but bars making those digital copies available outside the premises. Trap. Citing section 108 as authority for a public website. It is not.
[ ] Check whether section 108(h) applies. Published works in the last twenty years of term, not subject to normal commercial exploitation, no copy available at a reasonable price, no rightsholder notice. Why. Real, underused, and it permits distribution and display. Trap. Skipping the ongoing availability checks it requires.
[ ] Confirm the institution is an authorised entity under 17 U.S.C. § 121. Why. The accessible formats exception permits reproduction and distribution for eligible print-disabled users with no permission and no search. Trap. Assuming any nonprofit qualifies without checking the definition.
[ ] Launch the accessibility programme first. Why. It rests on a statutory exception and on the independent fair use holding in Authors Guild v. HathiTrust — two independent bases. Trap. Producing image-only scans and calling them accessible. Structured text, described images, tagged documents.
[ ] Consider 17 U.S.C. § 121A for cross-border exchange. Why. Marrakesh implementation permits sharing with authorised entities abroad.
[ ] Publish tier zero across the entire collection immediately. Catalogue records, finding aids, descriptions. No images, no text. Why. Discoverable and citable with zero copyright exposure, because descriptions are not the work. Trap. Holding the finding aid back because the contents are unresolved. Findability is most of the value.
[ ] Assign a tier per collection with recorded reasoning. On-premises access, authenticated access, limited public access, or open access. Why. Publication is not binary and the middle tiers carry most of the material. Trap. A single institutional default applied to everything.
[ ] Use the limited public tier as the workhorse. Thumbnails, reduced resolution, snippets, excerpts, with full resolution on request. Why. It is the Authors Guild v. Google posture and it is defensible for a great deal of material. Trap. Publishing a "thumbnail" at a resolution that substitutes for the original.
[ ] Reserve open access for public domain material, institution-owned rights, and collections the assessment supports. Why. It is irreversible in practice. Trap. Opening a collection because a funder asked, without checking the assessment.
[ ] [Gate] Design tier promotion as a metadata edit. Why. Items move up as terms expire, searches succeed, and licences arrive. Trap. A system where promotion means re-digitising.

Phase 6. Rights metadata and honest rights statements

[ ] Store the rights record in the collections management system, in structured fields. Why. A spreadsheet on a shared drive will not survive one migration. Trap. Treating rights as a free-text note.
[ ] Capture rights status, basis, determination date, determiner, search tier and date, access tier and reason, deed restrictions, claims and resolutions, and review date. Why. These are the fields that answer every later question. Trap. Capturing status without basis.
[ ] Use the standardised rights statement vocabulary. Why. Interoperable, aggregator-compatible, and honest about uncertainty. Trap. Inventing local terms that no aggregator can interpret.
[ ] Do not assert copyright in faithful reproductions of public domain works. Why. Bridgeman Art Library v. Corel held such reproductions lack the originality Feist Publications v. Rural Telephone Service requires. Trap. A blanket "© the Society" watermark across a public domain collection. It is unenforceable and it damages credibility.
[ ] Separate contract terms from copyright claims. Why. Conditions accepted at the point of access can bind contractually even where copyright is absent. Trap. Dressing a contractual condition up as a copyright notice.
[ ] Register the institution's own original outputs. Catalogue essays, original finding aid content, commissioned photography, exhibition materials. Why. Fourth Estate Public Benefit v. Wall-Street.com confirmed suit requires a completed registration. Trap. Registering after the copying.

Phase 7. Stand up the takedown operation

[ ] Publish the policy prominently on every item page and every collection landing page. Why. Visibility is what converts a potential claimant into a correspondent. Trap. Burying it in terms of use.
[ ] Word it plainly. We have made reasonable efforts to identify rightsholders; if you believe you hold rights, contact this address; we will respond promptly and will restrict access while we review. Why. Plain language gets used. Trap. Legalistic phrasing that reads as a challenge.
[ ] Staff a monitored address with a named owner and an hours-not-weeks service level. Why. An unanswered message is how a resolvable situation becomes a demand letter. Trap. A shared inbox nobody owns.
[ ] Restrict access first, analyse second. Why. Moving an item to metadata-only costs nothing; defending its display during a live claim costs a great deal. Trap. Litigating the merits before pressing the button.
[ ] Verify the claimant's chain of title. Why. A substantial share of claims come from people who assume they hold rights and do not. Trap. Removing permanently on an unverified claim and losing material that was clear.
[ ] Offer the full range of resolutions. Attribution, licence, reduced tier, permanent removal. Why. Attribution alone resolves a great many claims. Trap. Treating removal as the only option.
[ ] Register a designated agent and run a repeat infringer policy where the institution hosts user-contributed material. Why. 17 U.S.C. § 512 safe harbour requires it. Trap. Conflating the voluntary policy for the institution's own material with the statutory process for third-party uploads. Both should exist and they are different documents.
[ ] Log every claim, response time, and outcome. Why. It is the empirical answer to the board's risk question. Trap. Handling claims informally and having no data a year later.

Phase 8. Community consultation and cultural protocols

[ ] Run this review alongside the rights review, not after it. Why. It can change what gets digitised at all, and discovering that after scanning is expensive. Trap. Sequencing consultation after publication.
[ ] Identify material subject to community protocols. Indigenous cultural materials, sacred and ceremonial items, human remains and funerary material in documentary form. Why. These protocols do not map onto copyright and are not extinguished by it. Trap. Concluding that clear copyright means clear to publish.
[ ] Treat NAGPRA consultation as the model even outside its direct scope. Why. Its architecture is the established practice.
[ ] Flag records of institutional harm. Residential schools, asylums, hospitals, carceral and child welfare records. Why. They contain deeply personal information about identifiable people who did not consent. Trap. Publishing a name-bearing register because the copyright expired.
[ ] Screen for personal data about living identifiable people. Why. Privacy and sector-specific statutory obligations run independently of copyright. Trap. A copyright-only screen on a twentieth-century medical or student record series.
[ ] Consult before digitisation, and accept that the answer may be no. Why. Consultation that cannot change the outcome is not consultation. Trap. Presenting a completed plan for endorsement.
[ ] Build the technical capacity for community-set restrictions before promising them. Why. A restriction the system cannot express will be reversed at the next migration. Trap. Committing to seasonal or role-based access the platform cannot implement.
[ ] Support Traditional Knowledge labels or equivalent community notices. Why. They give communities a way to attach protocols to digital objects and are widely respected. Trap. Assuming they have legal force in themselves.
[ ] Record consultation outcomes in the same rights metadata. Why. Otherwise the restriction lives in someone's email. Trap. Treating consent as permanent; revisit.

Phase 9. Governance, funders, and the numbers

[ ] Take a policy to the board, not a project. Methodology, tier definitions, search standard, takedown commitment, escalation path. Why. A programme requiring board approval per collection stops. Trap. Bringing them a collection and no framework.
[ ] Set explicit escalation triggers for counsel review. Commercially exploited works, a refused permission, an ambiguous deed restriction, intended licensing out, unconcluded consultation, or personal data about living people. Why. Everything else runs under the policy. Trap. Escalating everything, which is the same as escalating nothing.
[ ] Name one accountable owner with authority to hold material back. Why. Rights decisions need a decision-maker. Trap. Distributed responsibility with no name attached.
[ ] Ask the insurance broker specifically about copyright infringement from published collections. Why. Some policies respond and many do not; the standard advertising injury grant is narrower than assumed. Trap. Assuming coverage from the existence of a policy.
[ ] Read the funder's conditions before accepting the grant. Open access mandates, required licence terms, repository deposit. Why. An open access requirement is hard to reconcile with a collection tiered at authenticated access. Trap. Negotiating the carve-out at the reporting stage instead of the application stage.
[ ] Separate any commercial licensing function from the open access programme. Why. The public offer and the commercial offer should not contradict each other. Trap. A collection that is simultaneously openly licensed and offered for sale.
[ ] Train the people making item-level calls. A decision tree, a clear escalation route, and explicit permission to hold something back without justifying it to a committee. Why. Cataloguers and technicians make most of the decisions. Trap. A policy document nobody applies.
[ ] Report five numbers annually. Items digitised; items published by tier; public domain determinations; searches completed and owners found; claims received and resolved. Why. They convert an anxious hypothetical debate into a factual one. Trap. Reporting throughput only.
[ ] Re-run the public domain sweep annually. Why. Material enters the public domain every year without anyone doing anything. Trap. Treating the sweep as a one-off.
[ ] [Gate] Adjust tiers on evidence. Collections generating claims move down; collections generating requests and no claims move up. Why. The programme should learn from its own data.

Phase 10. Format-specific handling

[ ] Split audiovisual material by rights layer. Screenplay, score, sound recording of the score, cinematographic work, and any licensed-in footage or artwork. Why. A film is a stack of separately owned copyrights and the music is usually the blocker, because synchronisation and master rights are separately administered. Trap. Assessing "the film" as one work and concluding it is an orphan when the score is actively licensed. A silent print is often the deliverable that ships.
[ ] Analyse sound recordings by recording date. Why. Pre-1972 recordings were governed by state law until federal treatment brought them into a scheme with staged expiry, and the answer genuinely turns on the date. Trap. Applying the published-work analysis for musical compositions to the recording, or the reverse.
[ ] Check oral history releases against the intended use. Why. The interviewee holds rights alongside the interviewer, and a 1974 release for "scholarly use" is not obviously a release for global streaming. Trap. Treating a signed release as unlimited. Read the scope, and where it is narrow, tier the access to match.
[ ] Treat incoming correspondence as a per-author problem. Why. Letters are unpublished literary works owned by their writers, so an archive of a public figure's incoming mail is an archive of other people's copyrights. Trap. Publishing the senator's incoming correspondence because the deed assigned the senator's copyrights. It did not reach the writers'.
[ ] Assess periodicals at contribution level as well as issue level. Why. Freelance contributions may never have transferred, and the periodical's renewal status does not determine the status of individual pieces inside it. Advertisements are separately owned again. Trap. Clearing a newspaper run on the masthead's renewal record and publishing the syndicated columns with it.
[ ] Check the deposit agreement for theses and dissertations. Why. Owned by their authors and governed by institutional policies that vary enormously by era. Trap. Assuming a deposit requirement transferred copyright.
[ ] Run a privacy review before a rights review on born-digital accessions. Why. Email, drives, and social media exports arrive with personal data, third-party material, and credentials mixed in. Trap. Ingesting a hard drive into a discovery system and finding out later what was on it. Appraisal and disposal are legitimate answers.
[ ] Separate the photograph from the artwork it depicts. Why. Bridgeman Art Library v. Corel governs the reproduction; the underlying work has its own term. Trap. A recent photograph of a mid-century painting does not make the painting available.
[ ] Verify the provenance of maps, plans, and technical drawings. Why. Frequently government-produced and public domain, and frequently misfiled because a commercial publisher issued the edition. Trap. Clearing on the imprint rather than the source.
[ ] Assess ephemera as a class. Posters, tickets, menus, packaging, flyers. Why. Highest yield for the no-notice determination and lowest yield for owner identification; item-level search on it is wasted budget. Trap. Running tier three searches on a box of restaurant menus.
[ ] [Gate] Give each format-specific category its own named assessment. Why. One elegant framework applied across every format is how a programme acquires its worst surprises.

Phase 11. Budget, staffing, and the work that is actually expensive

[ ] Budget rights work as staff time, separately from scanning throughput. Why. Scanning is the predictable line and it is rarely the constraint. Trap. Building the grant application around images-per-hour and absorbing the rights work into a cataloguer's existing workload.
[ ] Fund the renewal research as a named project with a deliverable. Why. Highest return per hour in the programme, because every non-renewal it finds converts directly into publishable material. Trap. Cutting it because it looks like overhead.
[ ] Fund structured metadata capture up front. Why. Determinations recorded outside the collections system are lost at the next migration and the work gets redone. Trap. Deferring metadata to a later phase that never gets funded.
[ ] Schedule community consultation in months, not weeks. Why. It involves travel, relationship-building, and community timelines, and it cannot be compressed by adding staff. Trap. A grant timeline that assumes consultation fits in a sprint.
[ ] Fund the takedown address permanently, not for the grant period. Why. It is the cheapest control in the programme and the obligation outlives the funding. Trap. A monitored address that stops being monitored when the project officer's contract ends.
[ ] Do not fund blanket permission campaigns for untraceable material. Why. A mailing where most of it returns undeliverable is postage, not diligent search. Trap. Reporting the mailing as evidence of diligence.
[ ] Train the technicians on three questions. Is there a name on it. Is there a date on it. Does anything about it feel like it should not be public. Why. The people making item-level calls are cataloguers and digitisation staff, not counsel, and they need something they can apply at speed. Trap. Issuing a policy document instead of a decision tree.
[ ] Give staff explicit permission to hold an item back without justifying it to a committee. Why. The failure mode is a technician who quietly publishes something they were unsure about because flagging felt like an imposition. Trap. An escalation culture that punishes flagging.
[ ] Answer the standing objections with data rather than argument. Waiting for orphan works legislation has a cost of its own; 17 U.S.C. § 504(c)(2) remits statutory damages for a qualifying institution acting on reasonable grounds; the access position rests on 17 U.S.C. § 107 rather than on section 108; and licensing revenue for most collections is smaller than the cost of administering it. Why. Boards respond to numbers. Trap. Meeting an anxiety with a doctrine lecture.

Phase 12. Working the sample matter

[ ] Take the glass plate negatives first. A commercial studio that closed in 1961 produced published portraits distributed to sitters, which makes the no-notice and non-renewal sweeps directly applicable, and a corporate successor search is a bounded task. Why. Highest yield, cleanest analysis, and a fast public win. Trap. Treating a studio's output as a single work. The negatives are individual photographs with individual publication histories.
[ ] Take the newspaper run second. Check the paper's own renewal record, then separately assess syndicated columns, wire copy, photographs, and advertisements. Why. The masthead's status does not clear the contributions. Trap. Publishing the full page images and with them a syndicated feature that is still administered.
[ ] Split the senator's papers into four sub-units. Outgoing correspondence and drafts, where the deed may reach; incoming correspondence, where it does not; photographs, owned by their photographers; and printed ephemera, which the no-notice sweep may clear. Why. One accession, four rights positions. Trap. Assessing the accession as a unit and tiering everything at the level the hardest sub-unit requires.
[ ] Publish the finding aid for all four sub-units immediately at tier zero. Why. Findability is most of the value and description is not the work. Trap. Holding the finding aid until the contents resolve.
[ ] Route the oral histories through the release review before anything else. Why. The 1970s consent language governs, and it will not have contemplated online streaming. Trap. Publishing on the strength of the project's institutional sponsorship.
[ ] Route the residential school photographs to consultation before scanning. Why. They document identifiable people in an institution of harm, and the copyright position is irrelevant to whether they should be public. Trap. Scanning first and consulting about publication afterwards, which presents the community with a completed fact.
[ ] Report the phase one yield to the board before scanning anything. Why. It reframes the conversation from forty thousand unknown items to a much smaller residual. Trap. Asking the board to approve online access before anyone knows how much of the collection is already public domain.
[ ] [Gate] Confirm the grant's open access condition against the tier assignments. Why. A mandate to publish openly cannot be satisfied by a collection tiered at authenticated access, and that conflict is negotiable at application and painful at reporting. Trap. Signing first and reading the conditions when the report is due.

[ ] Set the sequencing rule and hold to it. Public domain sweep, then deed audit, then accessibility, then tiering, then the residual. Why. Each step reduces the volume the next has to handle, and by the time the residual is reached — in-copyright, untraceable, culturally unrestricted, commercially unexploited — it is a fraction of the original and a manageable one. Trap. Starting with the hardest category because it is the most interesting, exhausting the budget on it, and never reaching the material that could have gone up on day one.
[ ] Diarise the annual review. Re-run the expiry sweep, re-search the flagged high-value orphans, revisit consultation outcomes, and re-read the claim log. Why. Every input to the assessment changes on a schedule. Trap. Filing the rights assessment as a completed deliverable.

[ ] Write down what the programme deliberately is not doing. Item-level clearance, blanket permission campaigns, and waiting for legislation. Why. An accepted risk that is documented as accepted is a governance decision; an accepted risk that nobody wrote down is an oversight when it surfaces. Trap. Leaving the omissions implicit and having them read later as things nobody thought of.

[ ] Name the successor for every ongoing obligation. The takedown address, the consultation relationships, the annual sweep, and the review diary. Why. Every one of these outlives the person who set it up, and an obligation with no named successor lapses silently. Trap. A handover document that lists the systems and not the promises.

Outcome. A programme that ships. The public domain sweep and deed audit remove most of the collection from the hard analysis. The accessibility programme launches on two independent bases. The remainder publishes at a tier the written assessment supports, with structured rights metadata, an honest rights statement, a live takedown address, and consultation concluded where it was required. Residual risk is accepted deliberately, documented, and managed operationally — which is a defensible institutional posture and the only one under which the material reaches the people it was preserved for.

The five things people get wrong

Attempting item-level clearance. It cannot be completed at collection scale. A programme premised on it is a decision not to proceed, dressed as diligence.

Skipping the public domain sweep. It is the cheapest work in the programme and it consistently clears more than institutions expect — particularly the no-notice and non-renewal categories.

Reading section 108 as a publication authority. It permits preservation and replacement copying and expressly bars making those copies available off-premises. The access position rests on 17 U.S.C. § 107.

Treating the diligent search as a permission mechanism. There is no safe harbour. The search finds owners, supports the fourth factor, and evidences reasonable grounds for remission of statutory damages. Those are worth having; permission is not what it delivers.

Leaving the takedown address unstaffed. It is the cheapest and most effective control in the programme, and an unanswered rightsholder message is the most reliable way to convert a correction into a claim.

Key Authorities at a Glance

| Authority | Proposition | |---|---| | 17 U.S.C. § 102 | Subject matter of copyright | | 17 U.S.C. § 106 | Exclusive rights | | 17 U.S.C. § 107 | Fair use | | 17 U.S.C. § 108 | Libraries and archives; on-premises limit | | 17 U.S.C. § 109 | First sale and display | | 17 U.S.C. § 121 | Accessible formats | | 17 U.S.C. § 121A | Cross-border accessible copies | | 17 U.S.C. § 201 | Ownership | | 17 U.S.C. § 204 | Signed writing for transfers | | 17 U.S.C. § 302 | Duration | | 17 U.S.C. § 303 | Unpublished works | | 17 U.S.C. § 304 | Renewal terms | | 17 U.S.C. § 411 | Registration precondition to suit | | 17 U.S.C. § 504 | Damages; nonprofit remission | | 17 U.S.C. § 512 | Safe harbour and takedown | | Authors Guild v. HathiTrust | Search and accessibility as fair use | | Authors Guild v. Google | Snippet display for search | | Andy Warhol Foundation v. Goldsmith | Transformative purpose narrowed | | Campbell v. Acuff-Rose Music | Transformative use framework | | Bridgeman Art Library v. Corel | Faithful reproductions lack originality | | Feist Publications v. Rural Telephone Service | Originality requirement | | Golan v. Holder | Restoration of foreign works | | Eldred v. Ashcroft | Term extension upheld | | Fourth Estate Public Benefit v. Wall-Street.com | Registration means registration | | NAGPRA | Repatriation and consultation | | Standardised rights statements | Interoperable rights metadata | | Orphan works diligent search proposals | The unenacted framework |