On 7/17/06, Tim Culhane <tim.culhane_at_criticalpath.net> wrote:
> I think the real problem with regards the pdf manual is the way in which
> the pdf is generated. As I understand it the pdf is generated using latX,
The problem is that the original document itself is structured (as is
used for the document structure and hyperlinking, for easy accessing
sections etc.), but you'll need the so-called "tagged pdf", which
extends that (pdf) structure further for easy extraction of the text
itself. This currently isn't supported by LaTeX, and thus the pdf
looses all additional "structure" information that was present in the
original source. Generating tagged pdf files from LaTeX is a common
feature request but won't be trivial, so don't expect this to come
soon. Also, this is a LaTeX issue that we can't solve.
> and as a result, the accessibility of the pdf document is very poor when
> read using a screen reader.
This is a known issue.
Unfortunately, as I said before this isn't possible to change easily
as it is a limitation of current LaTeX distributions / releases.
> As you can see, the structure of the document seems mixed up. Also, words
> runtogether without spaces between them.
What I don't understand is why your screen reader fails that badly. I
just copied a section from the manual from kpdf to a text file, and
apart from some broken characters it was formatted well (including a
button table). Speaking the text using festival from kpdf looks also
pretty good (except from some small glitches). As I don't have Adobe's
reader installed I can't test that ATM.
> I'm not an expert on the best way of generating accessible pdf, but it is
> possible (though by way of latX I'm not sure). Firstly the pdf needs to
> be structured. The manual as it currently stands uses unstructured pdf.
You need to distinguish between the various "structured" versions of
pdf. The manual has structures, but not the so-called "tagged"
structures that would be needed in your case.
> More information on how to produce more accessible pdf can be found at
Adobe's website doesn't help at all as we are using LaTeX, not any of
Adobe's products. Find more information on that topic on the internet
(and if you manage to find something that could help please let us
> Hope this helps, and If anybody wants more help on this I'll be happy to
If you want to help you could try building the html version of the
manual. Most likely we will use tex4ht for that task, but it seems
teTeX 3.0 has some problems with that. As I'm pretty busy with my RL
atm I simply don't have the time to look into this issue.
> However, until the html version of the manual comes on stream I think that
> a fully accessible manual is some way off.
As from what I can currently say the html version of the manual will
come before a tagged pdf will happen as creating a tagged pdf just
isn't like adding a switch when generating. Unfortunately.
Received on 2006-07-17