How it should be
The tag structure is the layer that says what a heading is, what a list is and in which order the text is read. Assistive technology reads that layer, not the page.
Without a tag structure, someone with a screen reader hears loose pieces of text in the order in which they are stored in the file, and that is usually not the order in which you read them on the page. All other topics depend on this one: if there is no tag structure, the rest of the check cannot be done.
What often goes wrong
- The document has no tag structure at all. In our audits this occurred 10 times in 43 documents.
- Part of the document has no tags. That part does not exist for assistive technology, and you cannot see this on the page.
- A page is tagged, but a piece of text on it is outside the tag structure and is not marked as an artifact either (content that a screen reader skips on purpose). That text is not available to assistive technology. If it is a page number or a running header, it should be marked as an artifact.
- Information is only in the header or the footer, for example a version number or a date. Those lines are rightly artifacts, but as a result someone with a screen reader does not hear that information anywhere. Put it in the text as well, for example on the title page. A program cannot detect this; a person has to check it.
- There is a tag structure, but the document is not marked as tagged. PDF readers that go by that setting ignore the structure that is there.
- The document is a scan: it contains no text, only images of text. Adding tags does not solve that, because there is nothing to put a tag around. Text recognition (OCR) turns it into text, but that text can contain errors and is not tagged yet. So first ask for the original file, or put the content on a web page. If neither is possible, text recognition is the first step and not the last.
Where you fix it
The tags come from the source file. Export from Word for Windows with Save As PDF and not with Print to PDF, and under Options tick the box for document structure tags for accessibility. Even then, Word usually does not produce a PDF that is entirely right, so always check the result. In InDesign you turn on the option for a tagged PDF. A file that was exported without tags means manual work in Acrobat afterwards.