PDF/A conversion is a preparation step, not an archival verdict

An archive may ask for a specific PDF/A profile and a validation report. Creating a file with a PDF/A export option is useful, but the file still needs to satisfy that profile and the receiving archive’s own rules.

Ask for the exact requirement

“Send a PDF/A” is an incomplete specification. Different PDF/A parts and conformance levels have different requirements and capabilities. Ask which profile is accepted, whether searchable text or accessibility is required, and whether attachments are allowed. Also ask whether the archive expects a particular validator or report format.

SandboxPDF’s PDF to PDF/A offers PDF/A-1b, PDF/A-2b and PDF/A-3b export through a local LibreOffice engine. Those options do not mean every possible input can be converted perfectly or that the receiving archive will accept the result. A successful conversion is the beginning of an output review, not the end of the submission process.

Keep the original file and record the selected profile. If the source contains a digital signature, interactive features or attached material, determine how the archive wants those handled. Converting can alter the document and its evidence, even if the pages continue to look familiar.

Understand the preservation objective

PDF/A is intended to support long-term preservation by constraining aspects of ordinary PDF that can depend on external resources or ambiguous rendering conditions. It is not a guarantee that the document’s facts are correct, that the scan is readable or that the file will be stored safely for decades. Preservation also depends on organisational processes beyond the file format.

A compliant file can contain a poor scan. A readable file can fail technical conformance. A technically valid file can still be the wrong version of a report. These are separate acceptance questions, so a good workflow checks format, appearance, content and submission requirements independently.

Do not use PDF/A as a generic synonym for “secure PDF.” Archival conformance is different from encryption, redaction or certificate signing. If you need to withhold information, prepare the authorised delivery content first. If the archive requires signed originals, follow its procedure for preserving that evidence rather than assuming conversion is harmless.

Fonts are a common source of visible change

Fonts affect character shapes, line lengths and page flow. A conversion engine may substitute a font when the original is unavailable or cannot be handled as expected. The result can still look broadly similar while a table wraps differently or a line moves to the next page. Small typography changes can therefore create meaningful layout differences.

Review pages with dense text, unusual symbols, mathematical notation and multilingual content. Check that accented letters and non-Latin scripts remain readable. A missing glyph can appear as an empty box or a substitute character. Compare actual words, not just the overall block of text.

If you control the source document, exporting from that source with appropriate font settings is often preferable to converting an already flattened derivative. Preserve the editable source where your retention policy calls for it. A later PDF/A conversion cannot recover structure that was removed earlier in the workflow.

Interactive and embedded features need decisions

Ordinary PDFs can contain forms, actions, attachments and other features whose handling depends on the chosen archival profile and converter. Do not assume that every interactive behaviour will survive or should survive. An archive may want a static record of completed form values rather than a form that can still change.

Inspect the input before conversion. If it includes attachments, determine whether they are part of the record. If it includes comments, decide whether those comments should be visible or retained separately. If it includes links, distinguish their visible labels from the ability to retrieve external content years later.

The appropriate choice depends on the record’s purpose. A visual page copy can preserve what a person saw at one moment, while an original package may contain supporting data that the page does not display. Do not discard that context without checking the archive’s instructions.

Run a small representative conversion first

For a large collection, start with one document that contains the difficult features you expect: tables, embedded fonts, images, rotated pages or completed fields. Open the local conversion tool, select the requested profile and allow the engine to initialize. The first download can be substantial, so use a suitable connection and device.

Save the result under a distinct name. Compare its page count and dimensions with the source, then inspect representative pages at a readable scale. Check the beginning and end of sections where pagination changes would be noticeable. If the conversion alters an important part of the document, investigate before processing the rest of the collection.

A pilot conversion also reveals performance constraints. Browser-based Office processing can take time and memory, especially on a phone. Running several heavy jobs simultaneously can reduce reliability. Process a manageable batch and retain clear records of which source produced each result.

Validation checks the file against rules

A PDF/A validator examines the actual file for conformance to the chosen profile. It is different from a viewer simply opening the document without error. The veraPDF validation documentation explains its machine-checkable validation approach for PDF/A and related standards. Use the validator or process required by the receiving organisation.

Read the report rather than recording only a green or red indicator. Confirm which profile was assessed and whether it matches the requested one. If validation fails, identify the reported rule and decide whether the source or export settings need correction. Renaming the file or changing a metadata label is not a substitute for fixing the underlying issue.

Keep the validation report with the delivered version when required. If you edit, compress or sanitize the file after validation, the report no longer describes those new bytes. Validate the final output again after any transformation that could affect conformance.

Visual review remains necessary

A technical validator is not a reader who understands your document. It does not tell you whether a chart legend became illegible, whether the wrong appendix was included or whether a source page was already missing. Compare the converted file with the intended record and check the material details.

For a scanned document, inspect the smallest text, faint marks and page edges. For a digital report, inspect equations, tables, headers and footers. For a multilingual document, examine script shaping and punctuation. If a form was flattened, check every important field value and selected option.

When the submission includes many files, use a repeatable checklist and sample strategy appropriate to the risk. A page count comparison can catch omissions quickly, but it cannot establish semantic correctness. High-value or regulated records may need more extensive review than an informal personal archive.

Accessibility is not implied by appearance preservation

An archival requirement and an accessibility requirement can coexist, but they should be assessed separately. A visually faithful page image may be difficult to use with assistive technology. Searchable text alone does not establish meaningful headings, table relationships or a correct reading order.

If the archive asks for accessible documents, clarify the relevant standard and validation procedure. Prefer an appropriately authored source document where possible. Converting a visually complex PDF into an archival profile does not automatically repair missing semantics or produce useful alternative text for figures.

Record which checks were actually performed. Saying “converted to PDF/A-2b and validated with the required tool” is more informative than claiming the document is “fully compliant” without specifying the requirement. Avoid extending one successful format check into a guarantee about every aspect of the file.

Example: preserving a project report

Imagine a report with fifty pages, a few scanned appendices and a completed approval form. The archive requests PDF/A-2b plus a validation report. Start by confirming that the report is the final version and that the scanned appendices are readable. Determine whether the approval form’s signature must be retained as separate evidence before conversion.

Create the archival candidate and compare all page boundaries around the appendices and form. Check that fonts and symbols survived. Run the required validator on the downloaded candidate, save the report and resolve any failures. Do not then perform a metadata purge that removes structures needed for the validated profile.

Package the final PDF and its report according to the archive’s naming and delivery rules. Keep the original source and any separately required signed record. The process is successful when the correct content, valid profile and required evidence reach the archive together—not merely when an export button finishes.

Maintain a repeatable handoff

For recurring submissions, record the profile, conversion method, validator version, source identifier and final filename. Keep that record concise enough to use consistently. It helps distinguish an actual content change from a different export setting when someone later asks why two versions differ.

Use Compare PDFs as one supporting check for changed text or appearance, while recognising its comparison limits. Consult the certificate-signing guide when signed records are involved. Treat archival preparation as a documented workflow with independent checks, rather than a format label that replaces judgment.