Ugly formatting

[Lexicon]

Originally Posted by GregS

The problem is, as has already been seen in HTML, CSS1 & 2, is that different implementations interpret differently, so unless there is a fundamental agreement on using the same code base everywhere, it is condemned to being an unreliable ideal.

You seem to be saying that the problem here is that different devices display the same CSS differently. I don't believe that is the real problem here, the real problem is that designers expect the same kind of control of layout and presentation that they have traditionally wielded over printed materials.

I think designers need to loosen the reigns a bit, when it comes to reflowable text displayed on a vast range of devices (each with displays of differing sizes and aspect ratios) it's simply not possible to have fine grained control of appearance. Nor is it necessarily right that they should expect it - as the reader should I not have a say in how the information is presented to me? That wasn't even an option with print books so it was never an issue for the consumer, times change though.

Most of the decisions on how to render content should be made by the device instead of the book creator, and devices should succeed or fail according to how well they make those decisions. The presentation abilities of an eReading device should be as much a feature as how many books it can hold or the physical size of the screen. As consumers we should be looking at which device renders our content best - a weak hyphenation algorithm should draw as much criticism as poor reliability or bad ergonomics.

[RWood]

Project Gutenberg (for some books) puts in the page numbers of the original books that they used as reference. An example would be "then Tom{p 23} said, "

While this removes the variation between devices, it requires the reference to a specific edition of the original book. It also makes reading fiction a lot less interesting.

[GregS]

Originally Posted by Lexicon

You seem to be saying that the problem here is that different devices display the same CSS differently. I don't believe that is the real problem here, the real problem is that designers expect the same kind of control of layout and presentation that they have traditionally wielded over printed materials...

Lexicon I am not suggesting that the vast majority of what is read should behave exactly as you propose. Indeed I am thrilled with epub as a much needed standard that acts exactly in this way. A few tools would deal with problems such as en-dashes, curly quotes and the rest. Widow and orphan control is no biggy and can be safely be left in the hands of display device software as it develops.

The real problem arises with non-fiction works where typographical elements are sometimes critical. It is a matter of horses for courses. Some works are simply better designed on a fixed page basis. PDF being based on a fixed media positioning language is proven in application.

It is early days, but I believe there is a way to generate; at the users end, from XML literature, mix and match stylesheets that produce PDF that fits the device being used, or indeed can produce epub type reflowed books.

This approach would be total overkill for most ebooks, epub and similar approaches work well, do their job, and are more than adequate. Plus they are relatively simple to create, which is important.

Think of the problem of producing a text, that requires a special type face (Ancient Sumerian), and has translated lines under each glyph that have to be kept in strict order regardless of the device being used. Add to this marginal notes, footnotes (which may be important to display on the same page at the time of reading and might occupy most of the page) and the simple CSS layout crumples. Even if CSS could do it, it would be a nightmare to create, and perhaps very slow to flow.

Reflowing text is a good thing, in general, but not always. I am proposing non-dynamic reflow that produced a fixed page ebook for different device display sizes. I am suggesting that it has a place, but not as a single universal standard, but as a potential special purpose standard.

Of course creating the tools and the stylesheet language capable of doing this is a long way off. As far as I can work out it is doable and the tools for constructing it are already developed, of course this is merely technical, design is critical if it were to be useful to book designers and users. This is for literature best marked up in something as complex as TEI, I emphasize "best marked up - not everything.

[HarryT]

Originally Posted by RWood

Project Gutenberg (for some books) puts in the page numbers of the original books that they used as reference. An example would be "then Tom{p 23} said, "

While this removes the variation between devices, it requires the reference to a specific edition of the original book. It also makes reading fiction a lot less interesting.

I always strip out those page numbers when I create books, because I find that they interfere with the reading experience. Very easy to get rid of them in BD with RegEx search and replace.

[Patricia]

Originally Posted by HarryT

I always strip out those page numbers when I create books, because I find that they interfere with the reading experience. Very easy to get rid of them in BD with RegEx search and replace.

Harry, is there any chance of your adding instructions on how to do this to your excellent BD tutorials?

[HarryT]

For a page number of the form:

[Pg xxx]

set the search string to "\[Pg [0-9]*\]" - that is, the character "[", then "Pg ", then a sequence of characters in the range 0-9, then another character "]".

Set the replace string as emply, and check the "RegEx" box on the search dialog.

[GregS]

Originally Posted by HarryT

For a page number of the form:

[Pg xxx]

set the search string to "\[Pg [0-9]*\]" - that is, the character "[", then "Pg ", then a sequence of characters in the range 0-9, then another character "]".

Set the replace string as emply, and check the "RegEx" box on the search dialog.

Harry T, forgive my suggestion if it does not suit the format being discussed, but would not it be better to mark-up page numbers either as attributes (<pg number = "xxx"/>) or elements (<pg>xxx</pg>) that could be hidden by a style sheet?

[JSWolf]

Originally Posted by DaleDe

Can you actually tell the text size? Other than s,m,l which varies from book to book?

Dale

I'm not sure really.. But I'm guessing it goes up by 2 point sizes. But the problem is I don't have a reference size most of the time. So if small was 10 medium would be 12 and large would be 14.

[JSWolf]

Originally Posted by HarryT

But it's NOT so easy on devices such as the CyBook where, not only do you have 12 available text sizes, but the user can install any font (family) they wish on the machine. Different fonts will produce different pagination.

Very true. Different fonts will have different results.

[jeffw]

Hi gang,

I'm helping a friend with his eBook. Silly me I had no idea what I have agreed to do.

I really thought the eBook production process would be more standardized these days vs. a few years ago.

Silly me, I was hoping for a one fit all solution. But I have discovered (by reading this thread) that there are many paths possible and not one of them is even close to being easy, clean or simple

“I use this program to do this and then I use this program to do that and then I go and manually clean it up.” ... Ouch --- “This font face and font size for this device and this font face and font size for that ...” double ouch.

Though I'm a newbie to eBook publishing, I'm an old fart in PC cyberland. I go back to before VisiCalc. The problems or better said 'challenges' I see here …are all to common in our cyberland world.

I've seen it in Operating Systems, Apple DOS, IBM DOS, MS DOS, DR DOS, -- I've seen it in Word Processors, Spreadsheets, and Image Graphics. Sadly, its part of the product development process. Each company believing that their approach is the best and they are all correct -- but only to a certain extent. Sure they might do task A really great, but they fail at Task B.

But there is one main difference I see with eBooks vs. all the other devices and interrupters.

eBooks as a media, is still very much a child. People need the flexibility to bring their eBooks with them into the next generation of eBook electronic device readers.. The eBook you own and formatted for a Palm, should be cleanly readable on a Kindle, when you move over to the newest toy. Or even XYZ electronic device that comes out 3 years from now. I don't see this as a VHS to DVD issue. This is data, not storage media. Data is data.

For this reason I see the problem with the reader devices not the formatting software.

I would love to see a product ad like this.

The new STARR eBook reader from Borne comes with the following drivers/engines and can read the following eBook formats:

1) ePub
2) Mobipocket
3) Kindle
4) PDF A4
5) LIT
6) Sony Reader
7) iLiad
8) MS Word
9) HTML ver 1 to 6
10) Custom PDF
11) MS PP
12) Text to Speech (EN, FR, SP, IT....)

This would allow an open door for new Writer formats and 3 party (drivers) software. Drivers/engines could be tweaked by manufactures to work on their devices. The eBook device integration would come with choices of drivers/engines/ sort of like printer drivers. But instead of changing the driver to fit your equipment --- you would change the driver to the eBook format you own. Or maybe even a automated discover the format and auto change it to the proper one. (requiring very little knowledge or understating)

Whew, well there is my wish list and just in time for Santa to see it.

I’m just not sure if I my behavior was good enough this year for Santa to stop in here with gifts.

Thank you everyone for these posts. I learned more here in an hour than the six hours I spent reading the web last night.

[kovidgoyal]

This thread is just an exercise in making mountains out of molehills. Just tell your friend to write directly in HTML or use editing software that outputs HTML and for all practical purposes he'll be just fine.

[jeffw]

ROFL - WOW

thank you kovidgoyal I like (love) simple and flexible apporachs

[Hadrien]

Originally Posted by Tanzaku

Yes, you are right, it does. In fact, that's the exact pattern for GREP or basic TEXT replacements. The draft document we posted has already had the single word "I[space]" replaced with "I[nonbreaking space]" to illustrate the point. We simply haven't yet run the rest of the bad-boy offenders through their paces yet.

As to the serif versus non-serif issue, again you are right. The common wisdom is to use non-serif fonts for screen displays and serif fonts for the text of printed displays. So, which category does the eBook fall into? Printed or screen? Or somewhere in between -- in which case, it is debatable which camp this technology falls into. We do, indeed, live in interesting times.

BTW, I see that Hadrien has responded with a post in this thread. I am most impressed with Feedbooks approach to this problem and their results are very good for the current state of the art. I have no doubt that as time goes on their algorithms will get better and better. At least I hope so! I love their approach to customizing the PDF creation with personalized fonts, margins, and headers/footers. Very user friendly. I can see a time when they will have customizable leading, hyphenation, paragraph styles, etc that will easily approach the kind of professional layout possible with InDesign -- but customizable on the fly. They are actually not that far from it now and clearly demonstrate a keen knowledge of the PDF architecture. So, if you are not familiar with their site, may I give them a plug?

Cusomizable leading, hyphenation and paragraph style could be possible too.

I'm wondering about something: currently, you're working on custom reg-exp in order to remove common typesetting mistakes. Are you planning on distributing these patterns ? It would be very interesting for everyone to create a database with all the useful patterns to fix common problems.
These could be used in any new software, and improve the overall look of every book available.

Maybe we should add a special wiki page for this ?

As for the Feedbooks approach: I believe that it's important to have choice and flexibility. For both the Cybook and iLiad, Feedbooks already offer 2 choice: Mobipocket for something flexible, yet not as nice from a typesetting point of view, or PDF files for better typesetting/presentation but without the ability to change the font on the device. As soon as Sony add Digital Editions support, it'll be the case too: epub for flexible formatting, PDF for fixed layout.

[ath]

Originally Posted by Tanzaku

Yes, you are right, it does. In fact, that's the exact pattern for GREP or basic TEXT replacements. The draft document we posted has already had the single word "I[space]" replaced with "I[nonbreaking space]" to illustrate the point. We simply haven't yet run the rest of the bad-boy offenders through their paces yet.

As long as the non-breaking space stretches as shrinks in the same way as an ordinary spaces does, fine. If it doesn't, the text will look weird.

In english typography, I can't ever remember seeing a rule against leaving a single-character word (I or a) at the end of the line. Where does that come from?

[pruss]

Originally Posted by JSWolf

Just add in the other fonts needed to make a true font family. That problem will then be solved. Full justified would not be much of an isue if we have hyphenation support. And it would also fix the spacing with larger text sizes as well.

Hyphenation looks better, I guess. But a bit of searching with Google and Google Scholar suggests that there is evidence that it somewhat retards reading speed. I am guessing that this is going to be particularly an issue on small screens.

(One might think that hyphenation will reduce the number of page turns, and thus improve reading speed. But I did a little experiment in fbreader, and there doesn't seem to be a significant difference in the number of page turns.)

Nicer italics is nice. I also suspect that slanted text, while uglier, is read faster, but I don't have data.

I suspect curly quotes do improve readability and hence are a plus.