Accessible Documents, Content, and Media
Documents
Accessible authoring starts upstream: build the source file correctly rather than retrofitting. In Word, PowerPoint, and InDesign, that means real heading styles (not bolded paragraphs), lists, meaningful hyperlink text, alt text, column-header table markup, document language, and a file name used as the title. Tagged PDF preserves this structure; PDF/UA (ISO 14289) is the conformance standard. A scanned page is an image - it needs OCR plus tags, not just an accessible-looking render.
Alt text craft
Describe purpose in context. Decorative images get a null alt (alt="") or are marked as artifacts. Avoid "image of…" and never paste a file name. A complex chart needs a short alt plus a longer description or data table.
Media
Four distinct artifacts, often confused:
- Captions - synchronized text of dialogue and meaningful sounds.
- Transcript - a text version; supports deafblind and search.
- Audio description - narration of visual information.
- Media alternative - a text rendition of the whole presentation.
Auto-captions are a starting draft, never a deliverable: they mis-hear names, jargon, and speakers, and they lack punctuation and sound cues.
Content
Plain language, short sentences, no color-only meaning, text that survives zoom and text-spacing overrides, and consistent navigation reduce cognitive load.
Traps: "It's a PDF, so it's inaccessible" (a well-tagged PDF is fine); "alt text should describe every visual detail"; "captions are optional if a transcript exists" (a transcript alone does not satisfy the captions criterion for pre-recorded video with audio); "the accessibility checker passed, so it's done" - checkers miss reading order and sense.
Rule of thumb
If you cannot navigate it with a keyboard and read it with a screen reader in a sensible order, fix the source file first, then re-export.