PDF vs PDF/A for Scanned Documents: What to Request

Specify PDF/A when your archive policy or receiving repository calls for an agreed preservation format. Consider ordinary PDF when it meets the intended use and applicable requirements. In either case, separately request searchable text, image quality and delivery checks. “PDF/A” alone does not describe a complete scanning job.

For an Australian business arranging a scanned archive, the useful question is what must remain usable after handover. Who will look after the files? Where will they be stored? Will another organisation receive them? Resolve those questions before asking a supplier to select an export setting.

This guide is a buyer's specification framework. PDF/A production, particular profiles and validation reports must be confirmed with the proposed provider; they are not established Pexatech service inclusions.

What PDF/A changes

PDF/A is a family of PDF standards intended to support preservation of a document's visual appearance over time. It constrains how a PDF is constructed. The aim is to reduce dependence on the original software and environment, rather than promise that a file will look after itself indefinitely. The PDF Association's ISO 19005 overview describes the family and its different parts.

For a concrete example, PDF/A-1 requires necessary fonts to be embedded and disallows file encryption. Its basic level, PDF/A-1b, addresses visual reproduction without requiring the same structural information as level A. It can contain scanned images without searchable text, and its filename still ends in .pdf. These distinctions are documented in the Library of Congress description of PDF/A-1.

For procurement, that means asking for the exact standard and profile, not relying on an extension or an “archive ready” label. A file that opens in your usual viewer has passed an opening check; that alone does not show PDF/A conformance.

Choose the format around the destination

Ask your records owner and destination-system owner to identify any existing specification first. A repository may prescribe a particular profile or delivery package. Your business may already have an archive policy worth following consistently across batches.

On a small screen, scroll the table sideways if needed.

These are discussion routes, not automatic rules based on a document's age. An old record can still need frequent access; a new scan can need long-term care. If the destination or policy is undecided, ask the responsible owner to resolve it before the whole collection is processed.

For NSW public offices, State Records NSW recommends PDF/A with lossless compression for long-term records or State archives, and standard PDF for some short-term modern records. It also discusses masters and access copies. That is scoped public-sector guidance, not a universal PDF/A obligation for private Australian businesses. See its digitisation specifications.

Write down the profile and the separate content requirements

Treat a designation such as PDF/A-1b as a precise proposed target, not shorthand for every desirable archive feature. Ask the provider to state the PDF/A part and conformance level or profile applicable to its proposal. Have the recipient confirm acceptance. Do not select the highest version number simply because it sounds newer.

Keep the preservation-format decision alongside these separate requirements:

• Capture: Which details must survive in the page images, including colour, faint annotations and small print?
• Search: Is an OCR text layer required, and which retrieval tasks will be checked?
• Accessibility: What user needs and accessibility criteria must the delivered documents meet, and how will those be assessed?
• Compression: Which compression settings are acceptable for the intended copy, and what comparison will confirm that essential details remain usable?

PDF/A is not a promise of lossless image capture. Nor does the label establish accurate OCR or accessible reading order. Do not infer those outcomes from a format-validation result. Where staff need extracted fields rather than text search, use the separate searchable PDF versus data-extraction guide to specify that output.

Compare representative files before making storage assumptions. Neither a .pdf extension nor a PDF/A designation tells you the eventual file size. Ask what the proposed settings produce for your material and include any additional access copies in the storage estimate.

Request validation evidence tied to the delivered files

Validation asks whether a file meets the selected format rules. For example, veraPDF's validation documentation describes applying rules through validation profiles. It is a tool example, not a claim that a particular supplier uses it or that one product must be purchased.

Agree the evidence before ordering. A useful handover record identifies:

1. Target: the agreed PDF/A part and profile, with the specification version recorded in the project brief.
2. Files checked: exact filenames and delivery versions, including which files were not checked.
3. Method: validator name, software version, selected profile and validation date.
4. Results: the report or agreed file-level summary, failures and any items the tool could not process.
5. Resolution: replacement files, repeated checks and the person responsible for unresolved exceptions.

Keep the report associated with the exact file bytes that were checked. If your handover process supports it, request a file checksum alongside the filename to help identify that version. A matching checksum helps establish that two copies contain the same bytes; it does not establish that the source pages were captured correctly.

Ask whether every delivered PDF/A file will be validated. If the proposal covers only a sample, record that limitation: a passing sample is not a result for unchecked files. Treat a file that could not be processed as unresolved rather than counting it as a pass.

The report should answer a format question, while your source comparison answers a content question. A standards-conforming file could still omit an attachment, contain a cropped image or carry incorrect recognised text. Use the scanning quality checklist for those delivery checks.

A fictional format decision for a Sydney archive

This scenario is entirely fictional. It describes proposed buyer decisions, not a Pexatech client, tested workflow or promised delivery capability.

A Sydney facilities consultancy wants to scan closed project folders. Staff need to find project references, while the records owner wants a maintained archive that can later be transferred to another system. The current brief says only “searchable PDF”. Nobody has yet agreed an archive profile.

The coordinator divides the unresolved decision into three tasks. The records owner confirms the preservation objective. The system owner checks accepted profiles and whether importing a file changes it. The scanning provider is asked to confirm which profile it can produce and validate, with a representative sample proposed before full processing.

One fictional folder contains typed correspondence and a faint handwritten amendment. The sample needs separate questions: does it pass the agreed format check, is the amendment readable, and can staff find the project reference? A pass on the first question cannot answer the other two.

The consultancy initially requests one retained file per agreed folder. A separate smaller access copy remains optional until staff identify a distribution need. This avoids committing to two versions without assigning responsibility for both.

If the proposed provider cannot deliver the required profile, the coordinator asks the records owner whether an alternative is acceptable or whether a separately scoped conversion step is needed. The team does not quietly change the requirement or assume a later conversion will be straightforward.

Recheck after conversion and destination processing

Make the final delivery the object of the check. If OCR, page replacement or conversion happens after validation, request a new validation result for the changed files. Also compare the affected content and search behaviour; repairing format conformance does not demonstrate that nothing useful changed.

Ask what the receiving system actually retains. Does it preserve the uploaded file, create a preview or replace it with a processed version? Where retained bytes remain unchanged, the original validation evidence still relates to that file. Where a new file is produced, assess and validate that version if it is expected to remain PDF/A.

For an agreed sample, follow the path staff will use: delivery, import, retrieval and export. Identify which version appears at each step, open it and check the intended task. Record conversion settings and replacements so the next archive owner can understand what happened.

Assign responsibility beyond the file format

A preservation format does not provide storage, backups, restoration or protection against unauthorised changes. It is not a read-only guarantee. Name the person responsible for retaining the accepted files, managing access and checking that recovery arrangements work.

Also distinguish encryption of an individual PDF from protection provided by the transfer and storage environment. An archive-format requirement and a confidentiality requirement need a compatible agreed handling arrangement. Do not settle that discussion merely by requesting a password on every PDF.

Keep the decision about paper originals separate. Accepting a PDF/A delivery does not authorise their destruction. The records owner still needs to determine their treatment under the requirements applicable to the organisation and record group.

A format request you can copy

Adapt this short request before comparing proposals. Leave undecided items visible so a supplier can explain feasibility and assumptions.

Purpose and destination: [record group, intended use, receiving system or repository].
Format: [ordinary PDF or exact PDF/A part/profile], subject to written confirmation of support and recipient acceptance.
Content: [file boundaries, image requirements, OCR and separate accessibility criteria].
Copies: [one retained version, or master plus a defined access copy].
Evidence: Confirm validation coverage, validator/version/profile, report and identification of the exact delivered files; describe failures and rechecks.
Destination check: Confirm whether import or export alters files and who checks any resulting version.
Ownership: [delivery reviewer, ongoing archive owner and separate decision-maker for originals].

Keep the agreed request with the scanning quote and scope. The useful handover is a format choice that your organisation can verify, use and maintain after the scanning project ends.