Leonard Rosenthol <[email protected]> wrote: >-=-=-=-=-=- > >Jason - do you mean support for creation of tagged PDF or >consumption/viewing or something else?
To reply to this and associated questions in the thread, I have in mind both creation of tagged PDF and its reading/consumption. As I understand this project, the purpose is to write a PDF library that supports both consuming and producing applications. Tagged PDF and accessibility are strongly interrelated, since the preservation of logical structure is invaluable to both braille and speech-based renderings, and the specification indicates that it can also be used to assist in visual presentation under certain circumstances which, I suspect, would include screen enlargement. As another example, the Orca screen reader, which relies on the ATK/AT-SPI accessibility infrastructure, makes extensive use of logical document structures in such applications as Firefox 3 to enable structural navigation and reading of documents. I am also aware of Adobe's proprietary implementation of both the reading and writing of tagged PDF, and their support for accessibility APIs, but the present concern is with free software implementations and with what happens on the free desktop. Naturally, I would appreciate details of any free/open implementations of tagged PDF and associated features, as well as further discussion of what might be possible in the context of the GNU PDF project. Obviously, a free (as in freedom) accessible PDF reader would be desirable, as would more support for tagged PDF in PDF writing applications. A less ambitious option, such as a conversion tool that can extract the text and logical structure of a tagged PDF document to an XML file, would also be highly useful. The pdftotext utility in XPDF can't do this due to the underlying lack of support for tagged PDF in XPDF itself. For untagged documents, pdftotext is nevertheless very useful, and it's my tool of coice for reading PDF files under circumstances in which I can't obtain the document in its original file format.
