The idea that a single definition of "long-term" in digital preservation simplifies everything is appealing. Retention schedules align, fixity checking intervals follow, and funding conversations get easier. Cross-institutional standards seem within reach.
Survey data from the EOSC EDEN project appears to support this. In its 2025 survey of 250 responses from 31 states and nations, 76% of archives reported having an agreed definition of "long-term" in digital preservation. Only 26% of libraries and 7% of museums and galleries could say the same. This gap looks like a maturity gap, suggesting that everyone should adopt the archival standard.
Why We Disagree
The conventional view equates a stated definition with a functioning program. The same survey suggests otherwise.
Definitions vary widely. For those with a definition, "long-term" ranges from 1-5 years to 100 years. Of these, 32.4% name 100 years as the duration, while 21.2% name 10 years. These aren't minor differences. A ten-year commitment and a hundred-year commitment require different decisions about file format normalization, storage migration, and staffing. A shared word can hide incompatible programs.
Consider the most rigorous cohort in the data. EDEN identified organizations that define "long-term," perform quality checks at ingest, and reassess digital objects later. This group made up 18.8% of respondents. Surprisingly, 35.7% were museums or galleries, 31% were libraries, and only 16.4% were archives.
Museums and galleries, the sector least likely to have a formal definition, are most represented in rigorous work. Archives, with the highest rate of agreed definitions, are the smallest in this cohort.
A definition on paper isn't the same as a reassessment cycle in practice. Treating the first as evidence of the second can lead to programs passing policy reviews but failing recovery tests.
The Evidence
Three points from the data deserve more attention than headline percentages.
Definitions Don't Predict Verification. Archives report the highest rate of agreed definitions and more quality checking at ingest, especially for provenance and completeness. However, few GLAM organizations regularly reassess preserved objects. A definition set at accessioning is a promise; reassessment shows the promise is kept.
Definitions Reflect Mission Differences. A research library supporting active computational use of collections has different obligations than an archive holding permanent records. Forcing one duration across both would produce a meaningless number.
Shared Concerns Despite Different Definitions. All organization types raised issues like sensitive data, data volume, and access provision, which threaten FAIR principles over time. This convergence suggests practical work is more aligned than definitions imply. It also highlights the importance of TRUST and CARE principles, which address accountability and rights over data, regardless of promised duration.
What to Do Instead
Stop asking "what's our number?" and start asking "what do we commit to verify, and how often?"
Define Commitments by Collection Tier. Assign each material body a preservation tier with a review horizon. A tier can be reviewed in a decade or a century; the point is that the tier is tied to a decision, not a slogan.
Schedule Reassessment as a Control. Build periodic fixity checks and quality reassessment into your Records Control Schedule equivalents for digital collections. If a definition has no review cycle, it's just a statement of intent.
Anchor the Process to ISO 14721. The Open Archival Information System (OAIS) reference model provides a shared vocabulary for Submission Information Packages, Archival Information Packages, and Dissemination Information Packages without dictating duration. It standardizes functions, where interoperability actually lives.
Adopt Process Standards Over Duration Standards. EDEN's Core Preservation Processes, 30 specific actions a trustworthy digital archive should take, apply regardless of content or domain. Standardize what you do, and let each institution state how long it commits to do it.
Document the Rationale. An auditor or funder will accept a 10-year horizon with a reasoned basis and a review date more readily than an unexamined 100-year default.
When the Conventional Wisdom is Right
Definitions aren't optional. The archival sector's 76% rate reflects real discipline, and the contrast with the 7% of museums and galleries is concerning. An institution with no working definition can't scope storage, justify budgets, or explain to a depositor what it has promised. Resourcing, metadata quality, and growing data volume are urgent pressures, and an undefined commitment makes them harder to defend.
Where institutions share infrastructure, like a consortium repository or national service, a common baseline is necessary. Shared storage can't be planned against five incompatible horizons.
The position isn't "skip the definition." It's that the definition is the cheapest part of the work and the weakest evidence of it. Write it down, attach a review cycle, test it against reassessment, and be wary of any program that can recite its number but can't show you the last time it checked.
