E-mail List Archives
Re: reading order
From: Duff Johnson
Date: Aug 24, 2023 11:09AM
- Next message: glen walker: "Re: Buttons and links: navigating a mixed list"
- Previous message: Nick Bromley: "Re: Buttons and links: navigating a mixed list"
- Next message in Thread: None
- Previous message in Thread: Karen McCall: "Re: reading order"
- View all messages in this Thread
As Karen indicates there are many many cases of documents that cannot be successfully represented to AT users without supporting tagged PDF.
Even if some software can successfully read pages of plain text without headings or columns - which might be ok, sometimes, for student's essays or novels - if it doesn't understand tags it will fail when asked to read, for example, an invoice, bank statement or newsletter.
Understanding these limitations is *critical information* for end users who are investing precious time and energy in learning how to navigate and read using AT.
Bevi's advice is good in terms of remediation. I'm not really a fan of technical terminology that's not rooted in the specification itself… leaving that aside, I do feel compelled to say this much:
First: all PDF software reads PDF's content streams, otherwise it can't even produce a page image.
Second, software that ignores PDF's tag tree cannot be reasonably called âassistive technologyâ for PDF.
Here's just a few of the things that such software cannot do for end users:
Provide navigation based on headings
Distinguish page numbers from the document content
Represent table or list structures
Represent alternative text
Reliably represent slightly more complex content (such as when two articles appear on a single page)
Represent content that flows across pages (lists and tables do this all the time)
Represent documents that include multiple languages
Indeed, some software that doesn't support tagged PDF can simulate these things by (for example) guessing at headings or lists… but these are GUESSES, not deterministic representations of the document's content.
If some tool that doesn't support tagged PDF is otherwise pretty good at guessing, for example, headings on a given document, the user should know the truth: they got lucky. If it works on one document it will likely fail on a different document. Certainly, users with different software will get different results.
Software that does not comprehend PDF's tags tree should not be represented as âassistive technologyâ. If it is, users will spend time and energy learning something that will either mislead them or fall over as soon as slightly more complex content is encountered.
The only solution is to address deficiencies with AT software directly with the vendors themselves. Make some noise!
> Academic institutions discover this every day. Their students come into class with everything under the sun, and their files have to work with those non-compliant technologies. Government, too. They can't tell students and taxpayers which software tools to use!
What they can do is clearly and consistently advise their users of the limitations in their software, and not pretend that software that cannot use tagged PDF is reasonably labelled as âassistive technologyâ.
In many cases governments can require their organizations to be aware of and use proper technology. Likewise, academic institutions - if they purport to serve students with disabilities - must educate their students on what works and what doesn't, and provide respective options.
Duff.
> On Aug 24, 2023, at 08:00, Karen McCall < <EMAIL REMOVED> > wrote:
>
> A caveat to determining the reading order is that if you have a PDF where half of the document is English/for example and the other half is French/for example, and the English is in a column on the right and the French is in a column on the left, organize all of the English content together in the Tags Tree and all of the French content together in the Tags Tree so that someone not familiar with either language will read all of the content in the language they know without having to switch midway through a thought or paragraph and ending up lost in the document.
>
> This goes for tri-fold brochures as well. The reading order for the first "page" should be the title panel of the brochure, followed by the inside panels and then back to the first "page" to read the last two panels in order. This is done in the Tags Tree.
>
> Newsletters with articles broken over pages is another example of ensuring that all of an article is read together which is done in the Tags Tree.
>
> I don't know of a way to do this in the Order Panel/Reading Order Panel or Content Panel. Those tools tend to focus on a single page rather that the logical reading order of content in the document as a whole.
>
> Another remediation in the Tags Tree is to ensure that paragraphs, lists and tables that span multiple pages are "joined" in the Tags Tree.
>
> Cheers, Karen
>
>
- Next message: glen walker: "Re: Buttons and links: navigating a mixed list"
- Previous message: Nick Bromley: "Re: Buttons and links: navigating a mixed list"
- Next message in Thread: None
- Previous message in Thread: Karen McCall: "Re: reading order"
- View all messages in this Thread