I have been teaching workshops at my university on creating accessible PDF forms from MS Word. I am very familiar with PDF accessibility. For the most part, the process of field recognition, marking up tables, writing tooltips, manipulating the tag tree, etc., all goes smoothly. I have noticed something that is creating a lot more work for our people, however.
After adding forms to the tagged PDF using recognition and the Tools > Forms > Add New Field tool, we adjust the field tab order and add tooltips. Once all of that is in place and we are ready to fine tune, we click the Tools > Accessibility > Add Form Fields to Tags button. Adding Form Fields to Tags does not work consistently if you edit the tag structure manually while in the Forms creation process. You can do some tag remediation prior to entering the Forms tool, but if you edit the tag structure while in Forms, the Adding Form Fields to Tags often does nothing -- this is certainly a bug in Acrobat X but it is not hard to work around.
The larger problem is that after we click Add Form Fields to Tags button, the tag structure reveals itself to be very much distorted and extreme amounts of tag structure remediation is required. Here are three screen shots that illustrate the point.
This first image shows the on-screen appearance.
Below is the expanded tags tree, prior to adding fields to the tag structure:
The tag structure is clean and elegant. If I were to publish this, no remediation would be necessary.
And here is the tag structure after adding form fields:
What has happened is Acrobat has grabbed recognized form fields and other elements in the document and has jammed them into the tag structure. Other things are affected as well. For example, headings in the page have extra, redundant textruns in them.
This has radically negatively affected the tag structure, requiring the author to do lots of remediation.
Any recommendations? solutions? similar experiences?
It seems obvious to me that this is a flaw in Acrobat that needs to be addressed, if accessibility is a goal. Or perhaps I'm doing something wrong??
Nope. No answer and nothing from Adobe.
I know something is going really really wrong, too: We have a product called CommonLook. It's an enterprise-level PDF accessibility remediation tool. It works well in fixing most structural problems in PDF through a relatively easy to use interface--you just drag textruns and elements around. It also has some tools that automate cleanup for accessibility.
I have no problem using this tool 99% of the time, but when I use it on forms I have made in Acrobat, the CommonLook plugin hangs when I try to save the PDF. It identifies a "structural error" in the PDF and crashes, leaving the PDF unchanged.
You can change the PDF in Acrobat by spending time moving around and creating new elements and deleting extra textruns and such, but, wow, what a royal pain in the *** that is!
Since no one is from Adobe is responding, I'm going to file a bug report.
Europe, Middle East and Africa