At a glance
PDF type is not one exclusive label. A document can be scanned, have OCR text, contain interactive fields and include tags at the same time. Record observable evidence separately, then review the reader’s task; no one of these properties establishes accessibility or whole-document accuracy.
One packet can require several review strategies.
| Pages | Observed clue | Review task |
|---|---|---|
| 1–3 | Born-digital text | Meaningful structure and sequence |
| 4 | Scanned signed letter | OCR, signature and original retention |
| 5–6 | Interactive fields | Labels, focus and form task |
| Uninspected pages | Unknown | Do not extend a sample label to them |
Scroll sideways to see all columns. Use arrow keys when the table is focused.
A working taxonomy
Use these descriptions as practical terms, not as a normative PDF standard. Inspect the actual file and page evidence, especially in long documents with mixed content.
| Description | Observable clue | What it cannot prove |
|---|---|---|
| Born-digital | Text objects from an authoring/export workflow | Correct reading order or reliable semantics |
| Image-only scan | Page image without usable actual text | Whether other pages have text |
| OCR-layer PDF | Recognized text associated with a scan | Recognition accuracy |
| Mixed PDF | Different content types across pages | One strategy will work everywhere |
| Interactive form | AcroForm or XFA structures | Usable labels or supported completion |
| Printed form | Visible boxes and instructions | Interactive fields exist |
| Tagged PDF | Structure tree present | Correct tagging or conformance |
Scroll sideways to see all columns. Use arrow keys when the table is focused.
Separate an observation from a document-wide label
Record the inspected pages and method. If a local profile inspected six pages in a 200-page packet, it provides clues about those pages. A table late in the packet, a handwritten amendment or a blank scanned appendix may remain unseen.
Selectable text can be generated by OCR or authoring software. Its presence helps identify text availability, but says little about recognition errors or navigation. Likewise, finding no interactive fields does not rule out a printed application form.
- Keep page-level evidence where strategies differ.
- Retain unavailable, uncertain and unassessed results.
- Choose review samples around consequential content, not only convenient pages.
Choose a task-specific next step
For scans, compare recovered text with images before reviewing structure. For tables, compare values and relationships. For forms, test instructions, labels, keyboard focus and the intended submission path. For tagged documents, inspect whether the tree represents the content correctly.
Treat sampled complexity as a routing hint. It cannot set a quote, establish handwriting accuracy or certify a document. The owner’s requirements, authored source and review findings should determine the conversion or remediation path.
Sources and scope
Methods, worksheets and illustrative examples are DocAccessible editorial contributions. These references support the technical or product details described above. W3C techniques are informative guidance; they are not a certification scheme.
- pypdf: text extraction
Why text extraction differs from OCR and cannot establish all visual or semantic meaning.
- W3C PDF7: OCR for scanned documents
Recovering actual text from scanned pages and checking the result.
- W3C PDF12: form field names, roles and values
Programmatic field information required for usable PDF forms.
- DocAccessible PDF inventory
The counting workflow, report options and purchase boundary.
Written by the DocAccessible team under our editorial policy. Suggest a correction if a source or product detail needs updating.