pdfLaTeX to get tag support?

[frabjous]

I just saw the program for the TeX User Group (TUG) conference this year:

http://tug.org/tug2009/program.html

Two talks struck my eye:

Ross Moore: "Towards tagged PDF"
Kaveh Bazargan: "TeX as an ebook reader"

The first suggests that they're working to make it so that pdfLaTeX handles inserting tags, for properly reflow-able, cut-and-paste-able PDF output from TeX source.

The second suggests that packages, or perhaps even new output profiles, are being worked on for .tex source to create ebooks... (A TeX>ePub profile even?) ... although it's hard to tell from the title.

Either would be a huge step forward, but makes me wonder if in fact, PDF does still have a place as an ebook format for the future. It also makes me wonder if anyone will be using BookDesigner or similar software if the wonderfully powerful resources of TeX become available to ebook creators!

[Moejoe]

Originally Posted by frabjous

I just saw the program for the TeX User Group (TUG) conference this year:

http://tug.org/tug2009/program.html

Two talks struck my eye:

Ross Moore: "Towards tagged PDF"
Kaveh Bazargan: "TeX as an ebook reader"

The first suggests that they're working to make it so that pdfLaTeX handles inserting tags, for properly reflow-able, cut-and-paste-able PDF output from TeX source.

The second suggests that packages, or perhaps even new output profiles, are being worked on for .tex source to create ebooks... (A TeX>ePub profile even?) ... although it's hard to tell from the title.

Either would be a huge step forward, but makes me wonder if in fact, PDF does still have a place as an ebook format for the future. It also makes me wonder if anyone will be using BookDesigner or similar software if the wonderfully powerful resources of TeX become available to ebook creators!

I don't know enough about TeX, other than what I've just been reading on the Wiki page. It's, and correct me if I'm wrong, a layout language for publishing as far as I can tell. So could we theoritaclly see PDF like layout but in an ePub format? Is that what all this means? If so, I'm very excited and can't wait to here more.

Thanks for letting us know

[mklynds]

I am with Moejoe on this one. Sounds very very interesting. One can only dream that with all the advancements in ebooks lately where will we be in 6 months to a year!! Can't wait.

Thanks again for sharing this bit of News

[Abecedary]

From the linked site:

{Kaveh Bazargan}
{\TeX\ as an ebook reader}
{An important advantage of ebook readers is their ability to modify
text size and page orientation, for the most comfortable reading
configuration. The ebook reader has to reformat the text on the fly
and with minimum delay. Current ebook readers (e.g.\ Stanza on the
iPhone) can do this reformatting, but cannot deal with complicated
text such as mathematics. We have been experimenting with using \TeX\ as
the formatting engine. Of course it can handle complex mathematics,
but it also creates the best line breaks of any ebook reader, e.g.
Stanza. We will report our experiments with using \TeX\ as an ebook
reader on the iPhone.}

Sounds interesting, but it's also fairly obviously in the early stages of anything, assuming their results are good (which they likely are).

[Jaime_Astorga]

Originally Posted by Moejoe

I don't know enough about TeX, other than what I've just been reading on the Wiki page. It's, and correct me if I'm wrong, a layout language for publishing as far as I can tell.

TeX is a typesetting language designed to allow people to format good quality books with reasonable effort. It's very popular in academia as LaTeX, as us folks studying or working in the quantitative fields (mathematics, engineering, science, and economics) tend to have trouble producing complex equations that are readable by other methods. It's kind of like the html/css deal in that content is separated from presentation. One just writes the content, gives it the correct logical meaning with the appropriate commands (for example, labeling subsections as subsections, or choosing whether the document will be an article or book), and LaTex plays the part of a book designer by figuring out the best way to present the book based on the commands specified by the user. If you are interested in learning how to use it, I recommend The Not So Short Introduction to LaTex 2e, which is freely available here under the terms of the GNU General Public License.

[Jellby]

Originally Posted by frabjous

Ross Moore: "Towards tagged PDF"
Kaveh Bazargan: "TeX as an ebook reader"

The first suggests that they're working to make it so that pdfLaTeX handles inserting tags, for properly reflow-able, cut-and-paste-able PDF output from TeX source.

It rather means "someone has some ideas", more than "they're working"

I'm not sure, however, which PDF readers (if any) properly us tagged PDFs.

The second suggests that packages, or perhaps even new output profiles, are being worked on for .tex source to create ebooks... (A TeX>ePub profile even?) ... although it's hard to tell from the title.

I'd say (and what Abecedary says seems to confirm it) that it's almost the opposite. It's not using TeX to create ebooks, but to display them, which would involve some kind of ebook->TeX conversion, even if it's hidden. One of first (and easier, I think) things ebook readers need is to improve the paragraph breaking algorithm, and using TeX's for that is an obvious way.

[zerospinboson]

Originally Posted by frabjous

I just saw the program for the TeX User Group (TUG) conference this year:

http://tug.org/tug2009/program.html

Two talks struck my eye:

Ross Moore: "Towards tagged PDF"
Kaveh Bazargan: "TeX as an ebook reader"

The first suggests that they're working to make it so that pdfLaTeX handles inserting tags, for properly reflow-able, cut-and-paste-able PDF output from TeX source.

The second suggests that packages, or perhaps even new output profiles, are being worked on for .tex source to create ebooks... (A TeX>ePub profile even?) ... although it's hard to tell from the title.

This indeed would be marvellous

Either would be a huge step forward, but makes me wonder if in fact, PDF does still have a place as an ebook format for the future.

But as long as ePub has no footnote support (footnotes that pop up at the bottom of a reflowable page, that is), and there is no standardized way to reference page numbers, PDF will surely stay around.

[Jellby]

Originally Posted by zerospinboson

But as long as ePub has no footnote support (footnotes that pop up at the bottom of a reflowable page, that is)

It sort of has. There is the oeb-page-foot property, which no reading software supports, as far as I know, but the possibility is there.

[zerospinboson]

Originally Posted by Jellby

It sort of has. There is the oeb-page-foot property, which no reading software supports, as far as I know, but the possibility is there.

Well, I'm mostly interested in working implementations

[frabjous]

Originally Posted by Jellby

It rather means "someone has some ideas", more than "they're working"

I'm not sure, however, which PDF readers (if any) properly us tagged PDFs.

Adobe Reader, which is of course the most common, does (at least with accessibility plugins, depending on version). And more importantly for my purposes, the PDF reader on the Sony's, right?

Some tidbits here:
http://www.tug.org/pipermail/pdftex/...il/007638.html

I'd say (and what Abecedary says seems to confirm it) that it's almost the opposite. It's not using TeX to create ebooks, but to display them, which would involve some kind of ebook->TeX conversion, even if it's hidden. One of first (and easier, I think) things ebook readers need is to improve the paragraph breaking algorithm, and using TeX's for that is an obvious way.

Yeah, I misread that, but Abecedary's posting of the abstract cleared it up. In a way it's an intriguing idea, though it would mean overriding the source's page size and margin settings, etc. Also, I think it would probably have to be a fairly minimal TeX system unless the memory and processing powers of the readers were beefed up a lot. A full texlive, with support for tikz/pgf images or pstricks or you name it, seems unlikely.

Yeah, I know these are early results, and it'll be awhile before we see anything, but I'm just glad people are working on it.

[Jellby]

Originally Posted by frabjous

Adobe Reader, which is of course the most common, does (at least with accessibility plugins, depending on version). And more importantly for my purposes, the PDF reader on the Sony's, right?

Where did I read that the Sony does not actually use tagging but just does the reflow "the hard way"?

EDIT: OK, here it is:

Originally Posted by llasram

the current Adobe Mobile SDK doesn't actually use the reflow tagging! (I wish I could find the blog/forum post, but one of the Adobe developers mentioned that they had better luck at the moment just ignoring the tags and reflowing as they could.)

[llasram]

Originally Posted by Jellby

It sort of has. There is the oeb-page-foot property, which no reading software supports, as far as I know, but the possibility is there.

I haven't tried this myself, but Adobe DE supports using a page template to define distinct page "regions" and assigning content to them via CSS selectors (the Adobe "EPUB Best Practices" documents gives the details). DE doesn't yet grok the oeb-page-* 'display' property values, but when it does, it should be prepared to sanely render content with books which use both a page template and the oeb-page-* display. So what you should be able to do is provide a page template for DE which places the footnotes in a footer, while also using 'display: oeb-page-foot' for future standards-compliant renderers.

Of course, as I said, I haven't actually tried this, so YMMV.

[kovidgoyal]

Hmm if they're looking at TeX as a rendering system (presumably rendering to PDF/DVI) then there will need to be an HTML->TeX conversion layer, which wont be lossless and then a TeX to PDF conversion layer, which will be slow. And all this for only better paragraph breaking?

[Jellby]

Originally Posted by kovidgoyal

Hmm if they're looking at TeX as a rendering system (presumably rendering to PDF/DVI) then there will need to be an HTML->TeX conversion layer, which wont be lossless and then a TeX to PDF conversion layer, which will be slow. And all this for only better paragraph breaking?

They might implement only the TeX algorithm for paragraph breaking and "full" TeX for certain elements, like equations (as does Auctex with preview-latex: http://en.wikipedia.org/wiki/File:Auctex.png)

[kovidgoyal]

Yes, but how will the equations be represented in the source ebook? MathML? In which case there are already ways to transform MathML to TeX