PDF is not an eBook format

[cmdahler]

Originally Posted by dmikov

BTW. Few centuries tradition very often means absolutely nothing. Horse cariages and sail boats come to mind.

That's all true, but there's a difference between tradition and science. Typography is a science, not a tradition. It has basic rules that have evolved for good reasons: they increase the readability of the text. For example, long experience has taught that the ideal single line of text on a page should have approximately 66 characters. Much longer than that, and the eye has to move too much to read the text and the reader loses the natural flow of effortless reading. That's why magazines and newspapers are printed with columns. Good typography also pays attention to word spaces throughout a paragraph in an effort to avoid distinct "rivers" of vertical whitespace since that is distracting and interrupts the flow of reading. Kerning and other letter-spacing techniques are designed to provide the reader with an uninterrupted blend of characters to form a word because the eye and brain reads a word as a recognizable single unit, not as a string of characters. More recent developments in typography (over the last century) include, for example, using hanging punctuation marks to create an optical margin - in other words, by letting hyphens and commas and the serif on a capital T extend beyond the margin just slightly, it allows the greater part of the entire character to get closer to the margin, and that produces the optical illusion of having a straighter margin than if you just make everything forced to a rigid margin.

All of this is done to provide the eye with a block of text on a page that is as uniform as possible to keep the eye flowing effortlessly across the page with as few distractions as possible. These aren't just "traditions;" they're proven techniques to make reading a pleasant experience for the reader.

E-readers should strive to produce this while also adding to the experience. Pop-up text blocks, note taking, dictionaries, the ability to resize the font - these are all cool technologies. But since the e-reader is still, at its core, a device that is designed to present readable text, these new technologies should add to, not break, the rules of good typography. For example, if the reader increases the text size, a good layout engine should re-typeset the document according to the rules of typography as much as feasible - software will never be able to do that perfectly, but it ought to at least try. Right now, ePub doesn't even try to do that, and that's a shame, because following some of the more basic rules of typography really isn't that hard to do.

[EowynCarter]

Typesetting a book is a task that even today remains too complex for a machine to do, without some human/manual help. Even LaTeX requires a human eye and human judgment on how to resolve imperfections in the best typesetting it could come up with given its internal restrictions for a given paper size and font size. [...]

True, only the human got get something "perfect".
BUT, do you think publisher are going to proofread that many versions ?They need to begin with doing it properly for one version. (Edit : and i need to proofread myself too!)
So, I definitively prefer a well done ePub than badly done pdf's.

E-reader will never have a "standard" in term of size and the rest. Different peoples, different needs.

For example, if the reader increases the text size, a good layout engine should re-typeset the document according to the rules of typography as much as feasible - software will never be able to do that perfectly, but it ought to at least try. Right now, ePub doesn't even try to do that, and that's a shame, because following some of the more basic rules of typography really isn't that hard to do.

Yes. That is what's is need. Send the ade crap to the hell (DRM and renderer) where it belong and do a proper ePub renderer.
Edit : Oups, got carried away there.

[DawnFalcon]

Originally Posted by ahi

3) PDFs (both mine and other people's) are downloaded like hotcakes--PDF *IS*, despite the general unavailability of eBook-reader-sized PDFs, the most popular eBook format. Hell, this is borne out even by surveys that Mobileread members sporadically try to taint in favour of ePub.

Well yes, it's a relatively new format. And for ereaders, only one format has a rocket under it's takeup, and it isn't PDF. I'm sure PDF will remain as one of the formats used for PC reading, but frankly I prefer Baen's take on HTML - I read on multiple PC screen sizes.

Typography in books is like windows on carriages.

Yep, it's relevant only so long as you're taking the train. When a new transport medium comes along...

Typography is ENTIRELY about enhancing the reading experience by making the text "disappear" and letting the reader focus on the content. If you have decided not to care about that... its nonsensical to talk about ease of reading, which nonetheless seems to be people's mantra about reflow... and is basically nonsense.

This is a very common fallicy seen in designers. Indeed, it's a major hazard in the games industry. Simply because you find something the best way to obtain immersion does not mean that it is the way your audience typically immerses in a given medium.

Any of this getting through to anyone?

That you're unable to see beyond what you like, and insist on applying it to everyone? Yes, you're certainly making that absolutely clear.

[DawnFalcon]

Originally Posted by frabjous

Maybe it's me, but I don't understand what you're talking about. You don't need to use DVI driver with LaTeX--pdfLaTeX exists after all--and the only "program" that interacts with DVI driver that would need to be created already exists.

Again, I don't understand what you're talking about. What's the relevant difference?

Er...pdfLaTeX creates PDF's. Why do you want to generate a PDF on the fly? It's highly computationally expensive. And as far as I can it's x86 only. Also, LaTeX now? That's an additional overhead as well. The only even remotely sensible path for a "TeX renderer" would be device-specific DVI renderer (That's what DVI is for).

Converting between standard TeX and XHTML most certainly is trivial. You're talking to someone who has done so multiple times. The only part of it that isn't trivial is the vast array of TeX code that currently has no counterpart in XHTML, since the latter is more primitive.

XHTML is certainly not "more primative", it's just as high level, and you're willing to release a program which guarantees 100% error-free conversion? Because otherwise it means another round of proof-reading.

Just another advantage of TeX--really all you can do now for graphics in an ePub is embed JPG, GIF or PNG images, none of which are infinitely scaleable. With TeX you could put the TikZ code right in, or use .eps, and then you can zoom in to your heart's delight.

And that support is both bulky and computationally expensive. There's good reason vector support isn't handled on e-ink readers.

Of course if something isn't unqualifiedly public domain, it's not public domain, but it doesn't have to be public domain if it's something in between, but their use of it falls under the usage restrictions of that in-between. And believe me, putting a TeX renderer on a device is certainly allowed by the TeX license, even if the full details of that license aren't entirely legally clear. If companies are scared away by the gray area, so much the worse for them.

It's not what I believe, it's what a company's lawyers will believe. And they are not going to allow it. The problem is that there is no liscence specified, only the conditional (non functional) public domain. Statements by the creator are not legally sound enough to be taken as a lisence when there's a lot of money at risk.

The DRM system isn't going to be "handling TeX". What would that even mean? The DRM system just needs to be involved in en/decrypting the document's code, and I can't think of any reason why precisely the same system couldn't be used that works with HTML code.

Simply wrapping a dumb DRM system arround the text is a good way to murder battery life and slow display processing. You want the DRM system to work in a complimentary way to the format, and that means tailoring it.

Don't put "could" in scare quotes. Dozens of new packages are added to CTAN every month. There already are LaTeX packages for embedding audio and video -- in that sense, it's further along that merely being "discussed".

Quite apart from LaTeX not being the same as TeX, embedding audio and video are not at all the features I was discussing, so I'm really not sure where you came up with them. More, you cannot depend on a "package" system to work correctly across multiple e-readers, and having executable code embedded in books is a major security risk.

I don't understand what you're trying to say. "Doing the things print books can do"--how are print books relevant here? Are you going back to your untrue belief that TeX is currently being used only to produce print books?

Sure, you can use it to make ebooks, but that is not what TeX was designed for, or the intent of it's development.

LaTeX already has a way of marking what language a word is. That's all that is needed.

No, it's not. That is precisely what I mean, you're thinking in very limited terms. How about, for example, one author I know who's going to suggest that language-specific word support be embedded in ePuB's, because he's done some market research, found that his YA-orientated novels sell better with the "correct" spelling in the UK and US markets, but wants to sell a single epub ebook in both.

He's made it pretty clear that he considers TeX in all intents and purposes in the public domain, with one exception: he thinks any significant derivatives should have a different name, which is all he requires.

Conditional public domain == no liscence. Statements are, once more, not good enough for a legal department with any sense whatsoever. And yes, they'd have to have their own version, versus ePuB which is a standard. That's very, very important. We've taken this long to get to a situation where most devices can read ePuB, fracturing it again will only massively hurt the takeup of e-readers.

Other free software elements they use most certainly do have proper liscences, and there is no problem whatsoever in using them. The problem is not the cost, not the ability of other people to use the code or anything else, it is the lack of a useable liscence for the code.

Actually, what I'm suggesting isn't necessarily incompatible with ePub being the standard. It'd probably be easier to modify the TeX renderer to directly render ePub/HTML source (since HTML and TeX source is pretty similar as is) than to produce an entirely new high-quality renderer for ePub.

Then suggest it. I don't think you'll get far, though, especially given the legal issue. (That is a first-look fail, regardless of any technical merits)

[Elfwreck]

Originally Posted by frabjous

The DRM system isn't going to be "handling TeX". What would that even mean? The DRM system just needs to be involved in en/decrypting the document's code, and I can't think of any reason why precisely the same system couldn't be used that works with HTML code.

There is no DRM for HTML code. There is DRM for ePub, because ePub slaps code on top of the HTML. One issue with making ebook devices straight TeX renderers is that someone would need to develop DRM for it, or the major publishers won't release TeX ebooks.

(This is assuming you can get them to release ebooks in a file format they've never heard of to begin with. The vast majority are using Windows, and doing the layout in Quark or InDesign, and converting to PDF with a button that says "convert to PDF.")

Sony right now doesn't use its own code. The OS is a free linux system that probably use precisely because they don't pay for it. The other elements, such as the PDF renderer, they pay for from software designers like Adobe. The attraction to using TeX is that they wouldn't have to pay for it. Yes, Amazon would be free to use the same code, but Sony is really making their money on selling you the hardware and the books from the store as is, and all that would be retained.

Right now, Amazon or Aztak or CoolER or Random New E-Ink Device Maker cannot legally just grab Sony's firmware and flash it to their devices. Cannot make devices that will buy books from the Sony store, nor be registered so it can read .LRX files.

Sony wants to protect that right. (The fact that I think this is stupid, like Amazon's insistence on creating a new Mobi DRM and slapping an .AZW extension on it, is not relevant.) Sony wants to make sure you can only buy their books if you bought their hardware. So does Amazon.

If they're working with open source or "semi public domain" software, they might lose that lock-in ability. Better for consumers... but you have to come up with a reason why it's better for them, or they won't consider the idea.

[Ankh]

Originally Posted by EowynCarter

BUT, do you think publisher are going to proofread that many versions ?The yneed to beging with doing it properly for one version.
So, I definitively prefer a well done ePub than badly done pdf's.

Precisely. What we are seeing at this time is that big names in publishing can't figure out how to do a reflowable (both .mobi and ePub) ebook formats right. Never mind typesetting, the bare text is sometimes not proofread at all, we get the crappy OCR at par with the darknet scans.

Perfection?! You wish.

[DawnFalcon]

Originally Posted by EowynCarter

To explain my point, i attachaed an image. Left is ADE, right is Calibre.

*cough*

You might want to use images of an ebook which might be legal. Just saying.

Also, looks like pressure needs to be put on Adobe to improve ADE. They're not going to go away...


Elfwreck - Uh? Sorry, I'm not seeing that lockin from Sony. Yes, you need to install their software (same for iTunes, I'd note, and you can't even /browse/ the iTunes website without having installed it, as you can the Sony site). But I just downloaded a free ebook off their site with no issues, and can read it on my PC. (I mean, it's a badly formatted google ebook, but still)

They run their readers on the open-source Linux, as do Amazon...using free software has no real relevance to lockins.

[Ankh]

Originally Posted by Elfwreck

Sony wants to protect that right. (The fact that I think this is stupid, like Amazon's insistence on creating a new Mobi DRM and slapping an .AZW extension on it, is not relevant.) Sony wants to make sure you can only buy their books if you bought their hardware. So does Amazon.

Uhm, I thought Sony is converting their store to ePub-only by the end of the year?

[EowynCarter]

Originally Posted by DawnFalcon

*cough*

You might want to use images of an ebook which might be legal. Just saying.

Also, looks like pressure needs to be put on Adobe to improve ADE. They're not going to go away...

Yeap, but it was the only one I have with that problem.
ADE have poor css support, these silly page number. Next on the to fix list being a proper justify. That one point where i miss mobi.

hey're not going to go away...

Why not ? . ePub is open, the only thing where they are needed is the DRM, and we don't really need the drm.

[Elfwreck]

Originally Posted by cmdahler

That's all true, but there's a difference between tradition and science. Typography is a science, not a tradition. It has basic rules that have evolved for good reasons: they increase the readability of the text. For example, long experience has taught that the ideal single line of text on a page should have approximately 66 characters. Much longer than that, and the eye has to move too much to read the text and the reader loses the natural flow of effortless reading.

And yet the posts here often have more than 100 characters in a line, with no kerning or ligatures, and there is no shortage of people reading them and commenting on them.

It's not that typography doesn't work; it's that people's acceptance of less-than-optimal typography is vastly affected by the medium they're reading. Some people's tolerance is higher than others; while I love well-designed pages with crisp edges to the text and no rivers and fonts that vanish when I read them... I'll read letter-sized, single-column pages in 9pt Comic Sans. (I have a teenage daughter who writes. She likes Comic Sans. I'll get her over it, but it'll take a while.)

Everybody cares about not-awful typography. Not everyone cares enough to pay for really good typography.

E-readers should strive to produce this while also adding to the experience. Pop-up text blocks, note taking, dictionaries, the ability to resize the font - these are all cool technologies. But since the e-reader is still, at its core, a device that is designed to present readable text, these new technologies should add to, not break, the rules of good typography.

I'm told Eucalyptus does an excellent job of typesetting the txt files from Project Gutenberg; I've got a friend who can't stand any of the other iPhone e-reading apps. So it does seriously matter to some people. However, the majority of readers seem more willing to tolerate Stanza's layouts, with its much larger range of books to read. (My friend removed Stanza from his iPhone; he couldn't read with it.)

Also, the science of onscreen typography is not yet developed. Margins around the page aren't just to rest the eyes--they're to give the fingers a place to stay that doesn't block text. It's entirely unknown what margin is optimal on an e-reading device where the fingers never hold the screen itself. LCD screens are made of light, and even e-ink screens have DPI issues; serifs are different on screen than on paper. The smaller sized screens, with different ratios, means different layout options, especially where images or charts are concerned.

Perfect typography for ebook readers isn't yet possible; nobody knows what it is. Insisting on a program that would allow it is pointless; first we need to find out what features are necessary. Perfect typography on a computer screen isn't even possible yet; 72-96 dpi limitations drastically changes what looks good & is easy to read, and the scrollbar changes how a "page" works.

[Elfwreck]

Originally Posted by Ankh

Uhm, I thought Sony is converting their store to ePub-only by the end of the year?

They still don't want other companies grabbing their firmware. And the point certainly works the other direction--Amazon doesn't want a new company to be able to grab the Kindle firmware.

I don't think any of the ebook device companies are working with open-source firmware. They're not encouraging end-user development of plugins that might read other ebook formats or allow display changes. Sony's switching to ePub--through its contract with Adobe; they're not letting you read those ePubs with a program that justifies the text, fixes the rivers, and removes the useless page numbers. Can't read PDFs with the Foxit viewer that has zoom. (I understand the 700 has a zoom function; the 505 does not.)

Any suggestion of new software will have to include a way for them to prevent their competitors from offering the same upgrades they have. Right now, that way is copyright law--they can pay for the code or the exclusive right to use it. If the code is open source, their ability to protect their investment and offer code that nobody else has, vanishes.

I think that's both greedy and short-sighted. However, we are talking about the companies who think it's a good idea to force you to install their software to buy their books, or who refuse to sell books to people who don't have their hardware. And the ones who think they'll sell more books if they demand that their DRM software is the only type on any one machine. What I think is bad business decisions, doesn't matter--they insist on the ability to limit their customer's activities, and they have the legal right to do so. They won't consider software that circumvents that right.

[DawnFalcon]

Uhm, what you actually mean is that the rendering engines are propriatory. The base firmware is Linux, and you can get the source used on both the PRS's and the Kindle. Legal requirement, you realise...

[Ankh]

Originally Posted by Elfwreck

I think that's both greedy and short-sighted. However, we are talking about the companies who think it's a good idea to force you to install their software to buy their books, or who refuse to sell books to people who don't have their hardware.

No, we are talking about companies that can get away with it. Buy Hanlin, use Open InkPot, and you are free of their clutches.

It is the market, us, users, that favours their locked-in solutions.

[EowynCarter]

I don't think any of the ebook device companies are working with open-source firmware. They're not encouraging end-user development of plugins that might read other ebook formats or allow display changes.

Yeah, i tought about that. We could have a way bettter reader if everyone worked on it (ie sony, bookeen, Irex, whatever). And why not, let's be crazy, make it open source so we can help too. Everyone would benefit from this no ?

[Kostas]

(Error)