[mc09] Problems

[jfine]

Hi

I thought it might help to gather together a list of the problems
we're having with mathematical content.

So I've started this thread.

--
Jonathan

[jfine]

Hi

At the OU, maths and upper-level physics courses are authored and
produced using TeX. This is to produce PDF. By sometimes we want to
produce web pages, and so we need conversion of TeX/LaTeX notation
documents to XML.

I like to divide this problem into two, the words and the math. By
TeX to XML I mean converting the words and leaving the math alone.

--
Jonathan

[jfine]

Tut, tut. You forgot the [mc09]

Jonathan

[Professor James Davenport]

In both contexts, actually, have you look at the work in 'tagged PDF'
coming out of pdfTeX?

James Davenport
Recently: Visiting Full Professor, University of Waterloo
Now back as:
Hebron & Medlock Professor of Information Technology and
Chairman, Powerful Computing WP, University of Bath
OpenMath Content Dictionary Editor and Programme Chair, OpenMath 2009
IMU Committee on Electronic Information and Communication

[Jonathan Fine]

On Tue, Jul 21, 2009 at 11:22 PM, Professor James Davenport <j...@cs.bath.ac.uk> wrote:

On Mon, July 20, 2009 11:04 am, jfine wrote:
>
> Hi
>
> At the OU, maths and upper-level physics courses are authored and
> produced using TeX.  This is to produce PDF.  By sometimes we want to
> produce web pages, and so we need conversion of TeX/LaTeX notation
> documents to XML.
>
> I like to divide this problem into two, the words and the math.  By
> TeX to XML I mean converting the words and leaving the math alone.

In both contexts, actually, have you look at the work in 'tagged PDF'
coming out of pdfTeX?

There is what seems to be quite a good introduction to PDF, which included tagged PDF, at
   http://en.wikipedia.org/wiki/Portable_Document_Format

In LTS (the publishing part of the OU) we prepare print ready PDF by using the Asura PS to PDF distiller, so we don't use pdfTeX.  Instead, we use TeX plus dvips.

However, tagged PDF can be more accessible, and that is important to us.

There are a large number of general questions here, that apply to non-technical content.  Is there anyone on this list interest in accessible mathematics in PDF (say for screen reader use)?  A related topic is copy-and-paste of mathematics from PDF to other applications.

I'll start a new thread on ebooks, in case there's interest in that also.

--
Jonathan

[Professor James Davenport]

On Wed, July 22, 2009 9:48 am, Jonathan Fine wrote:
> On Tue, Jul 21, 2009 at 11:22 PM, Professor James Davenport <
> j...@cs.bath.ac.uk> wrote:
>> On Mon, July 20, 2009 11:04 am, jfine wrote:
>> > I like to divide this problem into two, the words and the math. By
TeX to XML I mean converting the words and leaving the math alone.
>> In both contexts, actually, have you look at the work in 'tagged PDF'
coming out of pdfTeX?
> There is what seems to be quite a good introduction to PDF, which
> included tagged PDF, at
> http://en.wikipedia.org/wiki/Portable_Document_Format

I beg leave to differ on the quality of the description of tagged PDF: I
ceratinly learned a lot more in a ecent lecture (speaker BCCed in case he
wants to join in).

> In LTS (the publishing part of the OU) we prepare print ready PDF by
using
> the Asura PS to PDF distiller, so we don't use pdfTeX. Instead, we use
> TeX plus dvips.

I don't believe this route will get you to accessible PDF.

>
> However, tagged PDF can be more accessible, and that is important to us.
>
> There are a large number of general questions here, that apply to
non-technical content. Is there anyone on this list interest in
accessible
> mathematics in PDF (say for screen reader use)? A related topic is
copy-and-paste of mathematics from PDF to other applications.
>
> I'll start a new thread on ebooks, in case there's interest in that
also.

James Davenport

[jfine]

Hello James

On Jul 22, 10:43 am, "Professor James Davenport" <j...@cs.bath.ac.uk>
wrote:

> > There is what seems to be quite a good introduction to PDF, which
> > included tagged PDF, at
> >    http://en.wikipedia.org/wiki/Portable_Document_Format
>
> I beg leave to differ on the quality of the description of tagged PDF:

I think the article's fine for PDF, but sketchy for tagged PDF.

> > the Asura PS to PDF distiller, so we don't use pdfTeX.  Instead, we use
> > TeX plus dvips.
>
> I don't believe this route will get you to accessible PDF.

That's right. Asura is good at producing print PDF, but not good at
accessibility.

--
Jonathan

[Professor James Davenport]

On Fri, 24 Jul 2009, jfine wrote:
> On Jul 22, 10:43 am, "Professor James Davenport" <j...@cs.bath.ac.uk>
> wrote:
> > > There is what seems to be quite a good introduction to PDF, which
> > > included tagged PDF, at
> > >    http://en.wikipedia.org/wiki/Portable_Document_Format
> > I beg leave to differ on the quality of the description of tagged PDF:
> I think the article's fine for PDF, but sketchy for tagged PDF.

Exactly, so we agree here.

>
> > > the Asura PS to PDF distiller, so we don't use pdfTeX.  Instead, we use
> > > TeX plus dvips.
> >
> > I don't believe this route will get you to accessible PDF.
>
> That's right. Asura is good at producing print PDF, but not good at
> accessibility.

True, but my point was about the entire route. Once you have used
(ordinary) TeX, the information is lost, and cannot be recovered by Asura
or any other program. Hence my comment. You need a route that does not
lose that information, and the onlyone I know is (the being enhanced
version of) pdfTeX.
James

[Ian Wild]

Jonathan,

I am very interested in accessible mathematics, generally. In a
previous life I worked as a software engineer developing screen access
products and I have to say that accessing pdf documents was something
of a nightmare. PDF files can be locked to prevent the contents from
being copied, but that meant effectively locking out a screen reader.

I would need to check again, but the last time I investigated this
issue a locked pdf was rendered in Acrobat as an image (we entertained
OCR'ing them) and there were both legal and technical issues with
accessing the document via MSAA. I would need to investigate how the
document can be made accessible on a Linux system (e.g. Evince +
Orca), as we only developed for Windows.

Ian.

On Jul 22, 9:48 am, Jonathan Fine <jonathan.fi...@googlemail.com>
wrote: