4 ms·

Some in this thread have already mentioned platform lock-in/DRM, which is a huge issue, the elephant in the room really. But the other huge issue I encounter v

by acabal 12y ago

Some in this thread have already mentioned platform lock-in/DRM, which is a huge issue, the elephant in the room really. But the other huge issue I encounter very often is the absolute garbage typography and formatting of even commercial-grade ebooks. For example, of the ebooks I've read, 80% of them that were sourced from books that were in print before the ebook era are clearly slap-dash OCR jobs that never even got a proofread from a frazzled intern. Even modern books, those published this year or recently, are subject to this sort of mistreatment, though at a lower rate. King's Dark Tower series is a great example of this: when I bought the ebooks from Amazon a few years ago, they were filled with spelling errors, mis-curled quotation marks, errant or missing hyphens, broken paragraphs, and so on. And I paid nearly the same price as a print version! Even if an ebook isn't an awful OCR hackjob, publishers seem to insist on dragging our collective decade of experience writing web pages and the HTML5-based epub3 spec through the mud. An ebook is, at its core, nothing more than a handful of HTML pages and some metadata. Ebook readers, at their core, are web browsers. Yet commercial ebook developers insist on using outrageously bad markup that confuses a device's simple mind, awful custom styling that stomps a device's sane defaults, and markup that's only compatible with one or two platforms. (Amazon is partly to blame for this, because their stubborn and idiotic insistence on the god-awful mobi format instead of the open and sane epub format means many producers have to trust automated tools to convert between the two.) Free ebooks aren't excused from this either; books from Project Gutenberg are also hit-or-miss in quality, mostly miss. (Though that may be because many of their ebooks were produced before ereaders and modern standards were even a thing!) All of that together makes ereading a distinctly sub-par experience compared to the beauty that centuries of typographical knowledge produces in print books. I'm working on a public-domain project, standardebooks.org (not online yet), to help nudge the industry in the right direction. The idea is to promote a rigorous typography and coding style guide to public-domain books transcribed at Gutenberg. Anyone can use these guides, our toolchain, and the ebooks produced from them as a sort of "best practice" for creating high-quality epubs that look great on any modern ereader. If you're interested in helping, drop me a line.

jpatokal 12y ago
Otherwise +1, but characterizing EPUB as "sane" is rather charitable of you. EPUB 2 is an incompatible fork of XHTML c. 2001 with crude CSS and proprietary cruft, and EPUB 3 is still largely unusable in practice.
- acabal 12y ago
  I disagree on epub3; Standard Ebooks is using it as a base format. While it certainly is a sometimes-strange mix of XHTML and HTML5, and there are things I'd personally have done differently, as a means of presenting static documents it rivals the power that HTML5 on desktop browsers has. (Because it's basically just HTML5!) Whatever formatting limitations there are are generally are on the client rendering side--for example, Kindles are notoriously bad at rendering tables, but that's because the Kindle renderer sucks, not because any of the competing standards don't have good <table> tag support. Plus, since it's basically HTML in a zip file, anybody with basic web page production knowledge can jump in to producing an epub book. Sure there's some epub-specific cruft here or there but the core HTML markup has been well-understood for a decade. Additionally, the power of HTML5 and epub's semantic inflection standard lets us mark books up in fascinating new ways that aren't easily possible in simpler formats. Whether or not that ends up going anywhere we have yet to see; another thing Standard Ebooks is trying to do is add semantic inflection to the books we produce. Yes, it's not great for more complex stuff like interactive ebooks or javascript and so on, but no format is yet, and ereader technology isn't really there yet either.
  - jpatokal 12y ago
    Sorry, wasn't clear there: EPUB3 as a standard is theoretically OK, but real-world support of it on ebook readers just isn't there yet. And the update cycle is a lot slower than browsers or even mobile phones. Also, do you know what is a great platform for "javascript and interactive stuff"? The web browser!
wtallis 12y ago
Have you or anyone you know of done a feasibility study for TeX-quality layout being done in realtime on current mobile hardware? [La]TeX itself obviously isn't well suited to the task even if you make the obvious optimizations like rendering each chapter on-demand as a separate file, but since it is the gold standard for automated layout with good typography it would be interesting to know where the bottlenecks are.
- acabal 12y ago
  No, but since ebooks are just HTML documents the layout the code generates isn't particularly bad. It depends on the renderer of course. Renderers like iBooks even support fancy things like automatic ligatures and so on. CSS is growing to include more typographic details like the ability to specify lowercase numbers, etc., and when the epub spec and ereader renderers catch up to that we'll be doing pretty well for ourselves. Maybe not TeX-well, but pretty damn well.
  - shabble 12y ago
    By 'lowercase numbers', do you mean https://en.wikipedia.org/wiki/Text_figures https://en.wikipedia.org/wiki/Text_figures? I've never heard the term before, and initially wondered if it was some sort of joke, but it would seem not. Interesting stuff, typography.
  - _delirium 12y ago
    It's definitely possible to match TeX output on the HTML+CSS stack. PrinceXML is roughly there already, rivaling TeX for an output-to-PDF print publishing workflow. Unfortunately it's proprietary and expensive. Of the open-source renderers, the browser engines are probably the best bet to build on.
jay_kyburz 12y ago
I think the other side of the DRM elephant is piracy. Books are so small compared to movies, a few hundred meg can provide years of reading material. It can be hard to find a pirate copy of a single book, but it's fairly easy to find a huge package of hundreds of fantasy or science fiction books. I think there are some interesting opportunities for startups to explore new publishing models. I think authors should be looking at what is happening with free to play games or crowdsourcing. It might be hard to convince investors there is serious money in fiction, but people like to be entertained and I think there are ways you can get people excited about, and spending money on, written stories. Actually, if anybody wants to toss ideas around I'd love to chat some more.
aolagers 12y ago
There seems to be something wrong on site. All the book links currently return 404. But the project itself is really interesting. Definately bookmarked. Formatting on some Project Gutenberg books has indeed been awful.
- acabal 12y ago
  The site is still totally broken, and not ready for release. But I hope to have it complete in the coming months.

by acabal 12y ago

jpatokal 12y ago

Otherwise +1, but characterizing EPUB as "sane" is rather charitable of you. EPUB 2 is an incompatible fork of XHTML c. 2001 with crude CSS and proprietary cruft, and EPUB 3 is still largely unusable in practice.

acabal 12y ago
I disagree on epub3; Standard Ebooks is using it as a base format. While it certainly is a sometimes-strange mix of XHTML and HTML5, and there are things I'd personally have done differently, as a means of presenting static documents it rivals the power that HTML5 on desktop browsers has. (Because it's basically just HTML5!) Whatever formatting limitations there are are generally are on the client rendering side--for example, Kindles are notoriously bad at rendering tables, but that's because the Kindle renderer sucks, not because any of the competing standards don't have good <table> tag support. Plus, since it's basically HTML in a zip file, anybody with basic web page production knowledge can jump in to producing an epub book. Sure there's some epub-specific cruft here or there but the core HTML markup has been well-understood for a decade. Additionally, the power of HTML5 and epub's semantic inflection standard lets us mark books up in fascinating new ways that aren't easily possible in simpler formats. Whether or not that ends up going anywhere we have yet to see; another thing Standard Ebooks is trying to do is add semantic inflection to the books we produce. Yes, it's not great for more complex stuff like interactive ebooks or javascript and so on, but no format is yet, and ereader technology isn't really there yet either.
- jpatokal 12y ago
  Sorry, wasn't clear there: EPUB3 as a standard is theoretically OK, but real-world support of it on ebook readers just isn't there yet. And the update cycle is a lot slower than browsers or even mobile phones. Also, do you know what is a great platform for "javascript and interactive stuff"? The web browser!

wtallis 12y ago

Have you or anyone you know of done a feasibility study for TeX-quality layout being done in realtime on current mobile hardware? [La]TeX itself obviously isn't well suited to the task even if you make the obvious optimizations like rendering each chapter on-demand as a separate file, but since it is the gold standard for automated layout with good typography it would be interesting to know where the bottlenecks are.

acabal 12y ago
No, but since ebooks are just HTML documents the layout the code generates isn't particularly bad. It depends on the renderer of course. Renderers like iBooks even support fancy things like automatic ligatures and so on. CSS is growing to include more typographic details like the ability to specify lowercase numbers, etc., and when the epub spec and ereader renderers catch up to that we'll be doing pretty well for ourselves. Maybe not TeX-well, but pretty damn well.
- shabble 12y ago
  By 'lowercase numbers', do you mean https://en.wikipedia.org/wiki/Text_figures https://en.wikipedia.org/wiki/Text_figures? I've never heard the term before, and initially wondered if it was some sort of joke, but it would seem not. Interesting stuff, typography.
- _delirium 12y ago
  It's definitely possible to match TeX output on the HTML+CSS stack. PrinceXML is roughly there already, rivaling TeX for an output-to-PDF print publishing workflow. Unfortunately it's proprietary and expensive. Of the open-source renderers, the browser engines are probably the best bet to build on.

jay_kyburz 12y ago

I think the other side of the DRM elephant is piracy. Books are so small compared to movies, a few hundred meg can provide years of reading material. It can be hard to find a pirate copy of a single book, but it's fairly easy to find a huge package of hundreds of fantasy or science fiction books. I think there are some interesting opportunities for startups to explore new publishing models. I think authors should be looking at what is happening with free to play games or crowdsourcing. It might be hard to convince investors there is serious money in fiction, but people like to be entertained and I think there are ways you can get people excited about, and spending money on, written stories. Actually, if anybody wants to toss ideas around I'd love to chat some more.

aolagers 12y ago

There seems to be something wrong on site. All the book links currently return 404. But the project itself is really interesting. Definately bookmarked. Formatting on some Project Gutenberg books has indeed been awful.

acabal 12y ago
The site is still totally broken, and not ready for release. But I hope to have it complete in the coming months.