It's a little bit of both. What really happens is that you're looking at a moving target in terms of both inputs and outputs. Think of this like the early browser wars, where even a standard chunk of HTML could look different.
You can create an EPUB formatter that works to your satisfaction, but you've only coded it to work with that set of books. The next set of books may have different layout requirements, like padding between the text and the images.
Secondly, the epub world is evolving, and the readers that display the epubs all have their own interpretation. EPUB is simply HTML/CSS inside a .zip. e-readers are essentially browsers wrapped in their own chrome.
So, the real trick is to automate as much of the formatting as possible, but then have easy tools for the Author to make the final tweaks.
"It's a little bit of both. What really happens is that you're looking at a moving target in terms of both inputs and outputs."
Can you say more about this, it seems like we have two, perhaps three, targets; Kindle, iPad, and Nook. Together they probably make up over 95% of the e-book 'market' in terms of readers. Also they have a vested interest in making sure that books that look good today, still look good in the future so that publishers won't get stung by having to re-generate books.
Now the original discussion here was about authors who are trying to publish their own e-books. So when you said this:
"Secondly, the epub world is evolving, and the readers that display the epubs all have their own interpretation. EPUB is simply HTML/CSS inside a .zip. e-readers are essentially browsers wrapped in their own chrome."
I sort of tuned it out, while I can certainly see the top three book readers having their own interpretive quirks for EPUB, if, as an author, I get my tool chain which converts my manuscript into an acceptable EPUB on these three platforms, I would expect to be able to re-use that tool chain over and over again. So if I'm writing suspense novels, and my friend Bud gives me the occasional pen and ink illustration, I should be able to create a tool chain which can take my prose, scans of his drawings, and poop out an epub file. Then I sit back and go all Agatha Christy and write 150 detective novels.
Once the pain of getting the tool chain to work, with the caveat that book reader vendors are committed to backwards compatibility, I'm wondering why it would be 'hard' to publish after the first one was done.
Now granted if I went from pulp fiction to coffee table photography books I could see my tool chain being unfit for the task. And I could also see how, as a publisher of many authors it would be a pain in the arse to account for all of their different styles and topics and books and features etc. But if the rules really have changed, and its really 'better' for me to just publish my own books and distribute them through a channel like iTunes or Amazon, once the mechanics of publishing my style of book are over come, why wouldn't they stay over come, at least for me to get several books out using them?
You can create an EPUB formatter that works to your satisfaction, but you've only coded it to work with that set of books. The next set of books may have different layout requirements, like padding between the text and the images.
Secondly, the epub world is evolving, and the readers that display the epubs all have their own interpretation. EPUB is simply HTML/CSS inside a .zip. e-readers are essentially browsers wrapped in their own chrome.
So, the real trick is to automate as much of the formatting as possible, but then have easy tools for the Author to make the final tweaks.