notes from claus from PDF session at F2F

Notes from the group discussing pdf issues at F2f Friday morning.

Participants.
Tom, William, Sally, Katie, Loretta, Claus.

The discussion took a different form than was planned, because the issues 
about access in PDF is not well understood. Therefore we made up a list of 
issues that needs to be addressed, either in PDF itself in the automated 
tool that converts other formats to pdf or in the ua that renders pdf 
whatever that ua will be in the future. The ua uncertainty is caused by the 
fact that we currently do not know how the different screenreaders will 
incorporate the MSAA based access work that Adobe is carrying on at the moment.

Issues that must be addressed

Constraints in the way pdf is used
In a pdf file characters  codes are mapping  in the font.
One problem we cannot always go from character code to unicode 
representation. The conversion tool need to perform all conversions correct 
to produce the pdf source that contains the access features needed for the 
uas. things correct.
2. if characters are put into a file as images there need to be a 
explanation of which characters you use. Provide the character code that is 
represented with a picture.
3. If you have an image it needs to have a description of the image
4. You can attach multimedia contents as annotations into a pdf 
file.  There is no native format so maybe smil could be used. What we need 
to say is the attachments must be accessible. You have the freedom to 
choose multimedia format and player this is outside the pdf format and viewer.
5. Provide logical structure for the document maybe preserve structure.
6. The protection method must be set so that accessibility information can 
be preserved. The problem is that the protection technology as it  is 
implemented currently when you copy protect a document you also deny the 
accessibility aids access to the content needed to render pdf in an 
accessible way.
7. Use bookmarks and links within a document to provide navigation aids.
8. Mark tables appropriately.
9. When tables are created table summary table title row title column 
titles must be provided.
10. Provide expansion attributes for abbreviations and acronyms.
11. There is a lang attribute that defines the language of the text.
12. PDf has a lot of marks that is considered as being artifacts. The 
decision on which is artifact and which information should be provided as 
information regarding the contents is very important. One example of an 
obvious artifact is the lines that is used to divide cells in a table. 
avoid them.
13. How do we decide which tags should be considered artifacts and which 
tags are part of the content. If the decision is based on the needs of the 
user viewing pdf on a screen there will be tags that is going to be 
considered artifacts that will degrade the quality of the contents that 
will be provided to the accessibility tools.
14. There is an issue in PDF to tell where a wordbreak occurs and if a 
hyphen means - or a real hyphen. PDF must be able to distinguish between 
hard and soft hyphens.
15. when searching should alt text be included in the text being searched. 
Normally the alt tag is not shown in the pdf file but it will also depend 
on what ua that the new accessibility features that currently is being 
developed will use either the Acrobat reader a browser both or a special 
viewing utility provided by the screen reader manufacturers.

16. Should alt text applied to intermediate alt notes? This has something 
to do with rendering a tree structure where you can expand and collapse the 
notes. This seems to be very close to an issue Charles is working with.

Claus Thoegersen





--
wendy a chisholm
world wide web consortium
web accessibility initiative
madison, wi usa
tel: +1 608 663 6346
/--

Received on Wednesday, 25 October 2000 13:13:37 UTC