PDF is not an eBook format

[Elfwreck]

Originally Posted by Sonist

You are like a guy riling against the widescreen movie format, because your cheap B&W TV watch can't display it properly....

The lowest common denominator may not be good enough for everyone.

The screen size that's comfortable for me to carry in my purse is not "lowest common denominator." Small size is more important to me than good typography; that's why I read ebooks on a Clié with eReader program for a couple of years before I got my PRS-505. I switched because of battery life issues--not eye strain, not format preferences, not because I wanted a larger screen.

The 6" e-ink reader dominates the market; obviously, a lot of people like it. It's also clear that a lot of people would like a letter-sized reader--but that doesn't mean the 5" or 6" readers are "cheap early prototypes;" they're a very comfortable and useful size.

They're not a comfortable size for *printed books.* If you want to convince people that PDFs are the best ebook format, start by convincing publishers to release their PDF ebooks formatted for 6" reader screens. After people have read dozens of novels that worked well on the hardware they have, and compare those to epub & mobi books of the same content, they'll be more convinced that typography is important.

Right now, the choice is mostly between "good typography at a size that involves eyestrain" and "default computer typography at a size I can read."

(While you're at it, convince ebook reader manufacturers that 300 dpi really would be much better. Because a lot of the features of good typography are lost when the serifs are blurry.)

The tech to support PDF as the best ebook format is almost here. (It's waiting on better DPI support.) However, the market support for it is nonexistant... publishers often don't put title & author metadata in their PDFs; I don't think you can effectively argue that they'll format them to fit different screens. Or to fit ANY screens... right now, no commercial PDFs fit properly on any of the available ebook screens.

[Moejoe]

Originally Posted by Sonist

You are like a guy riling against the widescreen movie format, because your cheap B&W TV watch can't display it properly....

The lowest common denominator may not be good enough for everyone.

PDF is the B&W 4:3 TV, fixed, looks stupid when you play a 16:9 movie upon it. The Widescreen on the other hand is ePub, able to flawlessly scale from 4:3 to 16:9 and accomadate both with only minimal loss of quality.

[cmdahler]

Originally Posted by pdurrant

No need to use TeX for input. What is needed is for the layout engines for the ebooks readers to be improved to do proper paragraph and page level optimisations, including justification, hyphenation and widow/orphan control.

That's the whole point to TeX. There's no need for mobile reader manufacturers to have to develop their own layout engine that in the end does nothing more (and probably does a great deal less) than TeX. That's kind of like reinventing the wheel. No wait, it is EXACTLY like reinventing the wheel. The layout engines that are currently used in any mobile reader you can find are like chipping flint arrowheads in the stone age compared to what a TeX implementation could do. They've got a LONG way to go to produce an engine that delivers pleasing layout.

[Elfwreck]

Originally Posted by cmdahler

That's the whole point to TeX. There's no need for mobile reader manufacturers to have to develop their own layout engine that in the end does nothing more (and probably does a great deal less) than TeX. That's kind of like reinventing the wheel. No wait, it is EXACTLY like reinventing the wheel. The layout engines that are currently used in any mobile reader you can find are like chipping flint arrowheads in the stone age compared to what a TeX implementation could do. They've got a LONG way to go to produce an engine that delivers pleasing layout.

I have no idea what TeX is. (Something in my brain says it's a Unix thing?)

Is it something I could easily format my own ebooks for? Part of why I use PDFs is that I'm very comfortable formatting Word docs for conversion to PDF; I don't have any HTML-editing software that's nearly as easy to work with, so I don't format for ePub & mobi. (Or, possibly, I just haven't been trained on the software I have.)

Do Word docs convert fluently to TeX docs, if the interpreting device has the right software?

[EowynCarter]

Originally Posted by Elfwreck

They're not a comfortable size for *printed books.* If you want to convince people that PDFs are the best ebook format, start by convincing publishers to release their PDF ebooks formatted for 6" reader screens. After people have read dozens of novels that worked well on the hardware they have, and compare those to epub & mobi books of the same content, they'll be more convinced that typography is important.

Yeah. And even, that still won't solve the big problem pdf have. It's the fact it can't be edited, or converted into something else properly. Mobi can be annoying for that, but pdf is worse.

Pdf might sounds nice because it's probably the ones used for printed edition. But if the publishers wants to make pdf for 6", they're most likley to be badly formated (And there, you're screwed, see point one).

[DawnFalcon]

Originally Posted by pdurrant

No need to use TeX for input. What is needed is for the layout engines for the ebooks readers to be improved to do proper paragraph and page level optimisations, including justification, hyphenation and widow/orphan control.

I'd prefer that an ereader /dosn't/ do that. It's expensive on the battery life. No, what I'd prefer is that the software on my PC looks at the screen size and each font size, produces the appropriate set of page markup and copys it to the ereader aloneside the book.

And like PDF, TeX is designed to produce pages which look good on a single size of page. It's popular in exactly the same fields where PDF is popular - academic and technical publications. It is not a deacent way of handling situations like...oh...ereaders where reflow is used.

[WillAdams]

W/ tagged .pdf, one gets the option of having re-flowability along w/ a perfectly formatted representation.

W/ epub and other ebook formats which don't have a concept of size and position and page, one only gets reflowability, which means that one never gets a perfectly formatted representation, but always has to make do w/ how the page happens to be rendered by the viewer.

For me, ``good enough'' isn't. I don't accept anything less than perfect, and there are a number of printed books which I've purchased which I've never been able to finish because the typesetting is so bad --- similarly, I've found a number of ebooks to set so badly that I've been unable to continue reading them.

PDF is the only format (w/ matching viewer technology) which offers the level of control which is necessary to prevent stacks, widows, orphans &c. --- EPUB and the like desperately need improvements in their viewers to improve the quality of representation (at a minimum providing user options for justification and the like would be welcome)

William

[EowynCarter]

EPUB and the like desperately need improvements in their viewers to improve the quality of representation (at a minimum providing user options for justification and the like would be welcome)

Yeap.
And that's the problem. You're all saying ePub is crap. Not true. Adobe's implementation of epUB is.

[WillAdams]

I didn't say that epub was worthless --- it's a useful format, just one w/ marked limitations in the currently available viewing programs, and w/ other limitations which no viewing program will be able to resolve for the foreseeable future (there is no extant H&J algorithm which will preclude stacks, and research into page building at the chapter level or higher is still in its infancy --- list of recently published papers here: http://groups.google.com/group/comp....9?dmode=source )

While the vast majority of books may be handled adequately by epub (if one is willing to accept typographic infelicities such as stacks) there are many books which can't be and if those books are to exist in versions readable as ebooks, the only format suitable for them at this time is .pdf

William

[Jellby]

To be fair, you should compare Adobe-rendered ePUBs not with carefully crafted PDFs, but with default letter-size, Arial-type, MSWord-created PDFs...

[Elfwreck]

Originally Posted by WillAdams

W/ tagged .pdf, one gets the option of having re-flowability along w/ a perfectly formatted representation.

Only with well-tagged PDFs. Some programs don't set tags when they convert to PDF, and Acrobat's auto-tagging, like its OCR, is problematic: it works well with some originals (in the case of tagging, simple novel formats) and much less well with others. It has a tendency to tag multiple short lines of dialogue or poetry as part of a single paragraph, for example.

Setting tags properly is an exercise in masochistic nit-picky work. I've done it, and I'd do it again... for a book I dearly loved, or if someone paid me. A lot. I wouldn't do it for book conversions for the general public.

W/ epub and other ebook formats which don't have a concept of size and position and page, one only gets reflowability, which means that one never gets a perfectly formatted representation, but always has to make do w/ how the page happens to be rendered by the viewer.

They can get formatting & positioning, although much of the details are based on viewer capabilities, just like some web browsers won't view some page layouts.

For me, "good enough'' isn't. I don't accept anything less than perfect, and there are a number of printed books which I've purchased which I've never been able to finish because the typesetting is so bad --- similarly, I've found a number of ebooks to set so badly that I've been unable to continue reading them.

Do you think that an industry-wide move to PDF instead of other formats will fix this? That publishers will produce PDFs arranged to match your reading requirements?

PDF is the only format (w/ matching viewer technology) which offers the level of control which is necessary to prevent stacks, widows, orphans &c.

No, it's not. Don't mistake the viewer for the format. Plenty of the PDFs I read on my PRS-505 have orphans... because I can't read them at the letter size they're made in, and have to reflow. PDF certainly has plenty of orphan words & lines after reflow. And that's for the good ones, with tags. Untagged PDFs are hideous when reflowed.

Some PDFs have orphans & widows before reflow, because the decision to prevent them has to be made by the creator of the PDF.

I have Acrobat Pro and lock-breaking software; I add tags to my PDFs before reading. And fix the metadata, so they don't show up on my reader with names like "0689865384_INT," which is one I downloaded this week. And I add bookmarks so I'll have a TOC.

Any argument that I shouldn't reflow, presumes that someone is going to provide PDFs that fit in my screen. Since publishers can't be bothered to set the metadata and add a table of contents, I'm not holding my breath on that one.

Feel free to convince the publishers to make PDFs that fit in mobile ebook devices; there'd be a lot less arguments against PDFs if they were offered in several sizes. (And, of course, if the buyer were allowed enough page sizes to fit each of her screens.)

[WillAdams]

I didn't say all .pdfs were perfect, merely that .pdf is the only format which allows a level of control which allows one to achieve perfection.

I'm not seeing anywhere in epub a control which will allow me to detect and prevent stacks --- where are the controls in epub which allow for contextual information on how text formatted at a given size is at such and such a position?

William

[Lemurion]

While PDF at its best is much better than other formats - such as EPUB - I would say that in excess of 95% of the PDFs available are much worse. I'm not going to recommend a format that could be better but is almost guaranteed to be worse.

I see no need to spend as much time as it takes to read a novel in preparing it for my reader. Perfection is a worthy goal - but rarely attainable - and perfectly typeset but unreadable on my device is worthless to me.

[cmdahler]

Originally Posted by Elfwreck

I have no idea what TeX is. (Something in my brain says it's a Unix thing?)

Is it something I could easily format my own ebooks for? Part of why I use PDFs is that I'm very comfortable formatting Word docs for conversion to PDF; I don't have any HTML-editing software that's nearly as easy to work with, so I don't format for ePub & mobi. (Or, possibly, I just haven't been trained on the software I have.)

Do Word docs convert fluently to TeX docs, if the interpreting device has the right software?

TeX started out as a Unix typesetting command line tool. It pretty much runs on any major platform now. It accepts marked up plain text (similar to HTML), and outputs a typeset .dvi, .ps, or .pdf file.

Advantages to Tex: it's free, and it produces high-quality professional typesetting suitable for commercial publication. Many major academic journals require article submissions in TeX. It's second-to-none at quality equation typesetting for math journals, which was, unsurprisingly, why TeX was initially written.

Disadvantages: Steep, steep, steep learning curve. You can produce a pretty high-quality typeset page within a few minutes, but to move beyond the basics essentially requires reading a book about it. No WYSIWYG editor, just marked-up plain text.

As to whether your Word docs would translate to TeX, the answer would be not without a conversion tool. There may be one out there in academia-land; I wouldn't be surprised if someone has written something like that. I've never had the need for such a tool, however, so I've never looked for one.

[WillAdams]

There's a commercial Word -> TeX conversion tool which is quite good, Chikri Software's TeX2Word. The freeware rtf2latex works reasonably well though.

WYSIWYG is the wrong approach to take w/ TeX --- the LyX project's ``What You See Is What You Mean'' is much more appropriate:

http://www.lyx.org

People here may be interested in this though:

http://www.tei-c.org/Tools/Stylesheets/

(specifically the TEI to LaTeX stylesheet)

Sample documents made using (La)TeX:

http://www.tug.org/texshowcase/

(ob. discl. I have a few documents in there, most notably _The Book of Tea_)

William