Quote:
Originally Posted by ahi
It seems inelegant, until LaTeX and/or PDF supports elegantly storing multiple layouts (without storing the content twice or more times over).
|
LaTeX actually already supports this. Take a look at the 6 different LaTeX-produced PDFs I have
here. These are all produced with the same code.
I'm a bit busy right now, but I'll say more about later, and maybe provide some examples..
Sure, a human touch can get you slightly better output than relying on an algorithm exclusively. There is some human tweaking in my PDFs there (all stored in the same code, but only active at certain sizes.) But I hope no one would use this as an argument
against using an algorithm that *does* know about kerning, hinting, leading, hyphenation, proper whitespace usage in favor of one that doesn't even attempt such things!
LaTeX knows the proper hyphenation of words of nearly all widespread languages, and words that do not occur in its list can be given custom hyphenation rules in the code. Searching the document to discover such words and embedding the proper hyphenation spots for them with the code would be very easy.
It does often have to make "choices" between different desiderata when they conflict, and does so with its badness rankings. Let me be clear: IT does it, not the person making the document. However, its weightings of certain bad features can be tweaked, and even made conditional on page or font size. (E.g., in a small layout like an ebook, I would think widow/orphan avoidance would be less important than avoiding large whitespace areas, but perhaps others would disagree, and perhaps even one could enter one's own preferences right in one's reader...) Maybe it's choices aren't always optimal, but this is a small price to pay for giving our readers the ability to change font and page sizes on the fly.
And I still think the algorithm could be improved, and close to perfected. I don't see what reasons there could be for thinking that the choices made by human typographers when hand-tweaking can't in principle be done by a computer.
OK, so maybe device manufacturers will be scared off by an unclear license. But why on Earth would anyone think that takes them off the hook from producing something at least as good? Even if they did have to reinvent the wheel, surely the fact that there are wheels out there is proof that it can be re-invented.
Maybe hardware specs on readers is quite up to this yet. I'm not even convinced of this, but even if it's true, we're not far off from something that can.
And until that is done, people could and should offer PDFs in multiple sizes. As mentioned, it's not that time consuming, and certainly doesn't require reproofreading the book multiple times.
Quote:
|
As for typesetting multiple times... it's not so time consuming as to be prohibitive. These admittedly less than pristine versions (8 of them) were created with well less than a days work. 2-4 versions of an average length book shouldn't take longer than a day to do tolerably, or longer than a week to do really well.
|
Not nearly so long. It may have taken you that much time your first time, but once you've done this several times (and of course publishers would be doing this all the time), it would get faster and faster.
Typesetting is getting more crappy, but it's not because of ebooks, and certainly needn't be because of ebooks. If anything what's making it crappy is that people are getting used to reading the output of WYSIWYG editors like Word and their crappy output. and websites produced in effect with the same HTML interpreters as currently display ePubs on our readers, on a daily basis, and therefore, this level of quality is becoming acceptable in books too. I've read a number of paper academic books published by real presses lately that I'm fairly sure were typeset in Word, by amateurs, possibly the author him/herself. They definitely bother me. I blame Microsoft. OK, that's unfair, but any excuse to blame MS is OK with me.