WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: Z-Order and Tag Order Need to Match?

for

From: Duff Johnson
Date: May 28, 2020 11:57AM


Hi Lisa,

> Yes, it is complex. However, what I would say is that in Word, we don't
> need to deal with layers...So what I am saying is why is PDF different?

PDF is different because PDF can handle many layouts and types of content that Word cannot, for example, as generated by InDesign, or CAD software, or scanners, etc, etc. PDF can even accommodate documents created from multiple source technologies (Word + Excel + scan, etc).

This flexibility is unique to PDF and key to the technology's value proposition.

The original design for PDF ensured accurate reproduction of the author's visual intent irrespective of platform or operating system. The feature of PDF that supports accessibility - Tagged PDF - was added seven years later, in 2000.

> Now
> is Word perfect? Nope...but why are we having to spend endless hours trying
> to fix PDFs?

In my view, three principal reasons for that:

- Most authors don't use best practices in authoring documents (i.e., with a view towards accessibility).
- The PDF creation software users choose doesn't always create tagged PDF
- Word-processing and other creation software doesn't do much to help users do a good job of authoring for accessibility

> This is my question, and I am sorry if I went off the original
> answer into the weeds, but how many years we will have the same discussion
> about PDFs?

My view:

- We have to help authors use styles, relevent structures, create reasonable tables, use headings properly, etc, etc.
- We have to insist on software that supports tagged PDF, and is (ideally) PDF/UA-aware

> I wish EPUB had awesome readers, it would change the ballgame
> slowly and surely (I know, another story for another thread!).

It would be great. Sadly EPUB, although fantastic technology, doesn't approach the practical utility or flexibility of PDF.

Duff.



> On Thu, May 28, 2020 at 9:56 AM Duff Johnson < <EMAIL REMOVED> > wrote:
>
>> Hi Lisa,
>>
>> There are no "multiple layers" in the PDF in this sense of the term. There
>> is simply... content and tags.
>> Content is text, graphics, form fields, annotations, etc.
>> Tags are semantic structures (H2, P, Table, etc). These are applied to
>> content and are used to describe "info and relationships' as per WCAG SC
>> 1.3.1. Tags are the only means of providing these structures in PDF.
>> Otherwise I fully agree with you, and am very encouraged to hear that you
>> are working to convince other software developers to improve their support
>> for Tagged PDF.
>>
>> Duff.
>>
>>> On May 28, 2020, at 07:53, L Snider < <EMAIL REMOVED> > wrote:
>>>
>>> Hi Duff,
>>>
>>> Unfortunately I know all about these layers, but there is still no
>> reason.
>>> There is no reason to have multiple layers, what you see should be what
>> you
>>> get...whether you print, tag or whatever...This is partially why, in my
>>> personal opinion, PDFs are still inaccessible. 90% of the ones I have
>> seen
>>> since 2000 are a problem in some way, and most in major ways. Plus tags
>> are
>>> just one aspect of PDFs, people will all sorts of disabilities use PDFs
>> as
>>> you know well.
>>>
>>> I gave up on Adobe, they moved forward for a long, long time and the last
>>> 10 years they moved backward. Not in InDesign, InDesign is amazing and
>>> kudos to whoever worked on the accessibility aspect of it, now if they
>>> could move back those people to PDF we might get back to where PDF
>>> was...Microsoft has been kicking butt in terms of accessibility the last
>> 5
>>> years, so maybe one day Adobe will follow suit.
>>>
>>> I am now trying to find other PDF creation programs and convince them to
>> do
>>> better, may have a shot!
>>>
>>> Cheers
>>>
>>> Lisa
>>>
>>> On Wed, May 27, 2020 at 7:48 PM Duff Johnson < <EMAIL REMOVED> >
>> wrote:
>>>
>>>> Hi Lisa,
>>>>
>>>> Content order and logical reading order are "separate" in PDF because
>> the
>>>> technology is obliged to represent content for different purposes.
>>>> Rendering (e.g., printing) is a different purpose than accessibility,
>> for
>>>> example, and often requires ordering of content for processing purposes
>>>> that differs from logical reading order. PDF was originally designed for
>>>> maximum fidelity in print. Accessibility considerations were only
>> addressed
>>>> in 2000 with the addition of the Tagged PDF feature to the format.
>>>>
>>>> In 2020 the difficulty is not in the format but in the software.
>>>> Unfortunately it remains the case that...
>>>> PDF viewer developers don't do a great job of supporting tagged PDF in
>>>> many cases. This is simply a business decision.
>>>> Users continue to use software that doesn't understand tagged PDF
>> instead
>>>> of seeking out and demanding better.
>>>> Authors continue to author content without consideration for semantics
>>>> (e.g., use tab stops instead of table structures, etc.)
>>>> I'll beat my usual drum once again: if you want better PDF support,
>>>> complain to those who make your software. Demand better support for
>> Tagged
>>>> PDF. In 2020 there's simply no excuse.
>>>>
>>>> Duff.
>>>>
>>>>> On May 27, 2020, at 18:30, Paul Rayius < <EMAIL REMOVED> > wrote:
>>>>>
>>>>> As per ISO 32000, one of the intents of PDF is that the "accessible
>>>> layer" and the visual layer are independent of each other.
>>>>>
>>>>> Paul Rayius
>>>>> Director of Training
>>>>> CommonLook
>>>>>
>>>>>