Digital publishing · Published and updated
A backlist repair brief needs more than a title and an error count. Before commissioning work, establish which file you hold, whether its editable source is available, and what remains unknown. The useful first deliverable is a catalogue register that assigns the next decision—not a spreadsheet that declares books accessible.
This guide proposes three provisional planning lanes for authors, rights holders and small production teams: repair sampling, specialist assessment, and source investigation. They are Rex Publishing’s editorial synthesis, not official ABELab categories, cost bands or conformance scores.
What the research supports
ABELab’s May 2024 final report recommends collecting file format, other available formats, production information, source availability and editorial characteristics before triaging a backlist. We use that dated collection method as a starting point, not as evidence of current market conditions or tool performance.
Its gap analysis, first published in 2023 and supplemented in 2024, distinguishes content complexity from remediation need and describes sampling and format-comparison limits. Our practical conclusion: neither subject category nor a raw error total should decide a title’s place in the repair queue.
Two other references set boundaries. DAISY’s Ace guidance says automated checks assist a wider human evaluation; they are not a complete accessibility assessment. W3C describes EPUBCheck as checking publications against EPUB specifications. A format-validation result and an accessibility review therefore belong in separate evidence fields, not one universal pass/fail column.
Build the register around the next action
Give each file a stable internal ID and identify the exact copy being considered. Keep title and edition details in your working system; the neutral IDs below make the example easier to follow. If several files belong to one title, give them separate records and link them.
Record the format and layout, available production source, evidence date, report location, and the tool and rules versions actually used. Record the chosen format and accessibility targets separately. A target is what the team intends to assess against; a tool version describes how a particular report was produced. Write unknown or not assessed when evidence is missing. Do not turn an empty report field into zero errors.
Then add the unresolved content question, provisional lane, next action, owner and review point. The review point should describe what evidence will reopen the decision: a sample assessment, a source-file search result, or an editorial answer. A date alone does not explain what must change.
Three fictional records—not tested files
The following inputs are stipulated examples. No EPUB or PDF was processed, no source archive was searched, and no repair outcome was measured. Suggested owners are roles, not people commissioned for real work.
| Register field | File A | File B | File C |
|---|---|---|---|
| Format and layout | Reflowable EPUB | Illustrated fixed-layout EPUB | PDF-only holding; mixed languages |
| Source availability | Editable source available in this fictional brief | Unknown | Source missing from the supplied holding; archive search not performed |
| Evidence date and reports | Fictional brief only; no dated file report | Fictional brief only; no dated file report | Fictional brief only; no dated file report |
| Targets; tool and rules versions | Targets not selected; tools/rules not assessed | Targets not selected; tools/rules not assessed | Targets not selected; tools/rules not assessed |
| Unresolved question | Are suspected metadata/navigation gaps local, or signs of wider structural work? | Does the intended reading sequence remain clear across text and illustrations? | Can an editable source be recovered, and who can resolve multilingual content questions? |
| Provisional lane | Repair sampling—not an approved bounded repair | Specialist manual assessment | Source/reconstruction investigation |
| Next action | Agree targets; inspect a sample and manual-review scope before defining repairs | Commission a sequence/content assessment; establish source availability | Search author, production and archive holdings before estimating reconstruction |
| Suggested owner | Production lead with accessibility reviewer | Accessibility specialist with content editor | Archive/rights coordinator with multilingual editor |
| Review point | Sample findings and agreed repair boundary returned | Sequence findings and source options documented | Source-search result and editorial dependencies documented |
File A is not automatically the easiest job. Its lane makes a smaller initial question explicit: can a sample justify a bounded repair brief? File B needs a content judgment before that boundary can be drawn. File C needs a source search before anyone assumes reconstruction is the right commission. None of those decisions follows from format alone.
Set priority separately from the lane
A lane describes the next kind of work; priority describes why that work should happen now. Keep a separate priority note grounded in reader or distribution need, source readiness and dependencies. Identify who owns that decision.
For example, a documented reader-access problem may justify commissioning File B’s assessment before File A’s sample. An available source may make a task ready to start, but it does not by itself make that title more important. If a dependency is unresolved, record it instead of disguising it as a low score.
When findings arrive, retain the earlier evidence and explain any lane change. Ask the reviewer to name the file examined, the scope covered, what remains uncertain, and the next decision required. An estimate can then refer to an agreed task rather than an assumed condition of the whole catalogue.
Hand off to title-level work
The register stops at commissioning and reassessment. For the subsequent review, see our Ace and SMART manual-review guide and Qualebook workflow overview. Our EAA mapping explainer addresses a different standards-and-delivery question; this register does not determine legal obligations or exemptions.
For format context, our July 2026 EPUB baseline article is dated background, not proof of today’s production target. Confirm the specifications required for your actual commission and record them alongside, not in place of, tool versions.
Start with the files for which you need a decision. A useful row ends with a named role, an answerable question and a review point. It should make the next commission clearer without pretending the assessment has already happened.
Learn about Rex Publishing’s cross-language and format work or visit our contact page to discuss a project.
Primary references checked November 10, 2026. This is an original planning framework informed by dated research and technical documentation, not a report of 2026 remediation results. No cost, timing, tool-version recommendation or compliance verdict is inferred.