To digitize archives properly, you work through four gates in order: decide what to scan and in what shape it is, clear the copyright and privacy questions before anything goes online, capture the files at preservation-grade quality with usable metadata, and store them with enough redundancy and checking that they’re still readable decades from now. The scanning itself is the easy part. The legal and preservation decisions around it are where projects succeed or quietly fall apart.
Start With Triage and an Inventory
You almost certainly can’t scan everything at once. Rank materials by physical fragility, research value, and how often people ask for them. Deteriorating items and unique records go to the front; duplicates of widely available publications can wait.
Assess condition before anything reaches a scanner. Surface cleaning, careful flattening of folded pages, and minor repairs may be needed to prevent damage during handling. Tightly bound volumes, oversized maps, and photographs with flaking emulsion each need different handling than loose modern paper.
Build a detailed inventory of the physical items first. Document the original arrangement, assign temporary identifiers, and flag anything that needs conservation before it can be safely captured. That inventory becomes the bridge between the physical collection and the digital one, and it’s what lets you estimate scope and cost honestly.
Clear the Copyright Question Before You Publish
Copyright status is the single biggest legal variable, and it has to be settled before you post anything online. Scanning a document for internal preservation is one question; making it publicly accessible is another. The rules turn on when the work was created, whether it was published, and who created it.
Public Domain Works
If copyright has expired, you can digitize and share freely. As of January 1, 2026, all works published in the United States in 1930 or earlier are in the public domain, and that date rolls forward one year each January. Works published from 1931 onward may still be protected, depending on renewal.
Under the pre-1978 system, a published work received an initial 28-year term and the owner had to actively renew it. If they didn’t, the work fell into the public domain after those 28 years. If they did, total protection runs 95 years from publication.1Office of the Law Revision Counsel. 17 U.S.C. 304 – Duration of Copyright: Subsisting Copyrights A large share of mid-twentieth-century works were never renewed and quietly entered the public domain. The Copyright Office maintains registration records for checking renewal status.
Works Created After January 1, 1978
Copyright now lasts for the author’s life plus 70 years.2U.S. Copyright Office. 17 U.S.C. Chapter 3 – Duration of Copyright For anonymous works, pseudonymous works, and works made for hire, the term is 95 years from publication or 120 years from creation, whichever is shorter.3U.S. Copyright Office. How Long Does Copyright Protection Last? Most post-1978 material in your archives is still protected.
Unpublished Pre-1978 Works
This is where projects most often stumble. Historical collections are full of letters, diaries, manuscripts, and photographs that were never formally published. For those works, protection runs the author’s life plus 70 years, but federal law guarantees the copyright doesn’t expire before December 31, 2002. If the work was published before that date, protection extends to at least December 31, 2047.4Office of the Law Revision Counsel. 17 U.S.C. 303 – Duration of Copyright: Works Created but Not Published or Copyrighted Before January 1, 1978 A Civil War diary by an author who died in 1890 is in the public domain. A 1940 letter by someone who died in 1975 is not.
The Libraries and Archives Exception
Federal copyright law gives libraries and archives a specific carve-out for preservation copying. A qualifying institution can make up to three copies of an unpublished work for preservation, security, or deposit at another research library. For published works, it can make up to three copies to replace one that’s damaged, deteriorating, lost, stolen, or stored in an obsolete format, after first making a reasonable effort to find an unused replacement at a fair price.5Office of the Law Revision Counsel. 17 U.S.C. 108 – Limitations on Exclusive Rights: Reproduction by Libraries and Archives
The copying can’t be for commercial advantage, and the collections must be open to the public or to outside researchers. The catch that matters most for digitization: digital copies made under this exception generally can’t be distributed or made available to the public outside the library’s premises.5Office of the Law Revision Counsel. 17 U.S.C. 108 – Limitations on Exclusive Rights: Reproduction by Libraries and Archives You can digitize for on-site access and internal preservation. Posting the files on a public website needs a different legal basis.
Fair Use
Fair use allows reproduction of copyrighted material without permission in certain circumstances. Courts weigh four factors: the purpose and character of the use (nonprofit educational use favored over commercial), the nature of the original, the amount copied, and the effect on the market for the original.6Office of the Law Revision Counsel. 17 U.S.C. 107 – Limitations on Exclusive Rights: Fair Use Digitizing an entire document for a public archive weighs against fair use on the amount factor but may favor it on the others, especially where the original has no active commercial market and the purpose is scholarship. It’s a judgment call, not a rule, and it’s where archivists most often need legal counsel.
When You Need Permission
If a work is still under copyright and no exception clearly applies, you need a license or written permission from the rights holder before making the digital copy publicly available. The Copyright Office advises checking whether a limitation like fair use applies first, since permission isn’t required for every use.7U.S. Copyright Office. Circular 10 – How to Obtain Permission When you do get permission, document it: who granted the license, the scope of use, any restrictions. Keep those records alongside the files.
Check Donor Agreements Separately
Copyright isn’t the only constraint. Many collections arrive with deed-of-gift agreements imposing their own conditions. A donor may restrict access for a fixed period, retain intellectual property rights, or prohibit online publication entirely. A repository generally has the right to reformat materials it owns, including digitizing them, unless the gift agreement says otherwise. But “can we scan it” and “can we post it” are different questions. Review the transfer documents for every collection before publishing digital copies, and flag anything under time-limited restrictions.
Screen for Private Information
Historical records routinely contain personal information about people who may still be living. Social Security numbers, medical details, financial data, and criminal history scattered through old files can create real legal exposure if published without redaction. The National Archives uses an age-based screening framework that works as guidance for non-federal institutions too:
- Records older than 75 years generally don’t require screening for personally identifiable information.
- Records 30 to 75 years old should be spot-checked for dates of birth, medical history, criminal records, and photographs that could affect the privacy of the subject or surviving family.
- Records under 30 years old need screening for financial information such as bank account and credit card numbers.
- All records regardless of age must be screened for Social Security numbers, fingerprints, biometric data, and taxpayer identification numbers.
These thresholds apply to records going online, where the potential audience is unlimited.8National Archives. Before Screening Records If the collection includes medical records, HIPAA compliance adds another layer: any entity handling protected health information must limit access to the minimum necessary for each role and maintain signed business associate agreements with vendors who touch that data. When in doubt, redact before publishing and keep an unredacted master under restricted access.
Capture at the Right Resolution and Format
The hardware depends on what you’re scanning. Flatbeds work for loose flat documents. Overhead scanners or high-resolution digital cameras handle fragile items, tightly bound volumes, and oversized materials like maps or architectural drawings, because the original stays flat and doesn’t press against glass.
Getting resolution right the first time avoids rescanning. Federal digitization guidelines from the Federal Agencies Digital Guidelines Initiative (FADGI) offer a widely adopted framework:
- Text-based documents: 300 DPI minimum. Enough for clean optical character recognition and readable reproduction.
- Photographs and detailed graphics: 600 DPI or higher, to preserve tonal gradations and fine detail.
- Oversized items like maps: often 400 DPI, since 600 DPI can produce unmanageably large files at large formats.
Save archival masters in a lossless format like TIFF, which keeps everything the scanner captured. Then create access copies from those masters: JPEG for images, PDF for searchable text documents. Users interact with the access copies; the TIFFs are what you store for the long term. Never compress the masters. You can always make a JPEG from a TIFF; you can’t recover what compression discards.
Name Files and Add Metadata Consistently
A naming convention seems tedious until you’re looking at your 40,000th file called scan_final_v2_revised.tiff. Make file names machine-readable and self-describing. A common pattern combines collection name, series identifier, container number, and a sequential item number, like CollectionName_Series01_Box03_Item001. Avoid spaces, special characters, and anything that shifts over time.
Metadata is the descriptive information embedded in or attached to each file. The Dublin Core Metadata Element Set provides a widely used baseline of 15 standard fields, including title, creator, date, subject, description, format, and rights.9Dublin Core Metadata Initiative. Dublin Core Metadata Element Set, Version 1.1 The rights field matters especially for digitized archives: it should record the copyright status of the original, any access restrictions from donor agreements, and whether the digital copy is publicly available. Good metadata keeps a file understandable even when separated from the rest of the collection.
Store With Redundancy and Check Fixity
Creating the files is the easy part. Keeping them intact and accessible for decades is where projects quietly fail.
The 3-2-1 Rule
Keep three copies of every file, on two different types of storage media, with one copy off-site. Local servers give you immediate access. A second copy on a different medium, such as LTO tape, protects against the failure mode specific to your primary storage; LTO-9 tapes hold 18 terabytes each and have an expected shelf life over 30 years. The off-site copy guards against fire, flood, and other disasters that would take out everything in one location.
Fixity Checks
Digital files can degrade silently. A single flipped bit can corrupt an image, and nobody knows until someone opens the file years later. Fixity checking generates a checksum (a digital fingerprint) for each file at creation, then periodically recalculates it. If the values don’t match, the file has changed and needs to be replaced from a known good copy.
Frequency depends on the medium. Tape-based systems are commonly checked annually; hard-drive systems benefit from checks every six months. More frequent checks catch problems sooner but put more load on the storage. At minimum, verify checksums when files are first ingested, and generate fixity information for any files that arrive without it.
Format Migration
Storage media and file formats become obsolete. Anyone who has tried to read a Zip disk or open a WordPerfect file understands the problem. Preservation means periodically migrating files to current storage and, when necessary, converting them to updated formats. TIFF has been stable for decades, which is why it’s the preferred archival format, but the media holding those TIFFs will need refreshing every five to ten years.
Decide What to Do With the Originals
Once material is digitized, the question follows: can you discard the paper? Federal agencies have specific authority under National Archives General Records Schedule 4.5. Source records can be destroyed after validating that the digitization meets NARA’s standards for digital quality and completeness. Several exclusions apply:
- Source records created before January 1, 1950, that are scheduled as permanent or are unscheduled cannot be destroyed under this authority.
- Records with intrinsic value, where the original physical medium carries historical or evidentiary weight beyond the information content (a signed treaty, a document with wax seals), must be retained.
- If the digital version doesn’t meet the quality requirements in federal regulations, the originals cannot be destroyed.
Agencies must also consult legal counsel before disposal to confirm there are no litigation holds, appeal rights, or other legal constraints requiring retention.10National Archives. GRS 4.5 – Digitizing Records
Non-federal institutions don’t fall under GRS 4.5, but the same principles work as best practice. Keep originals when they have intrinsic value, when the digital copies haven’t been independently verified for quality, or when donor agreements require physical retention. The default should be to keep them unless there’s a clear, documented reason to deaccession.
Make the Archive Findable and Accessible
A digitized archive nobody can find hasn’t accomplished much. Choose a platform that supports full-text searching through OCR-processed documents and browsing by the metadata fields you’ve built. If your institution receives federal funding, Section 508 of the Rehabilitation Act requires electronic content to be accessible to people with disabilities, which means PDFs need embedded text layers rather than image-only scans, and image-heavy collections need descriptive alternative text.
For copyrighted materials made available under fair use or limited permissions, practical access controls help demonstrate good faith. Lower-resolution access copies or visible watermarks signal you’re not trying to replace the commercial market for the work. For public domain material, offer high-resolution downloads without restrictions. The point of digitizing a public-domain collection is to get it into as many hands as possible.