Originally Posted by JSWolf
The reason to be concerned about the code is that if there are any gotchas—and there can be plenty going from Word to ePub—it's easier to fix.
I've seen some ePub from Smashwords have all kinds of different issues. And they were all Word sourced. Issues like night mode is broken, font size cannot be changed, the ePub in all underline or all italics, and others. So Word > ePub can be a crap shoot.
After the ePub is made from Word, the next thing to do is clean up the code. Don't rely on just seeing how it looks. It could look fine in ADE, but when going to night mode in say Bluefire, it's broken.
Wolfie, my sweet:
You know you are one of my favorite curmudgeons, but on this, you are just wrong. You're not wrong that there are some absolutely appalling ePUBs on SW; but their genesis isn't the problem. The problem is that Smashwords, very much like Amazon or Nook, really, accept almost any old crap and "grind" it into the required format. Smashwords uses "the grinder," which, we all know, is the Calibre API; Nook uses whatever they're using (I suspect the same, but their ePUBs look cleaner than SW's). Amazon takes all sorts of crud in Word (or HTML, or whatever) and mooshes it into mobi format.
Now, TONS of HTML files are sent up to Amazon for creation, and they're every bit as crappy as anything else, no matter the source. The short and long of it is, if you don't clean the damn files, you get crap. I don't care if they're odt, txt, pdf (sigh), or Word.
And, really, I could (won't, obviously, but) send you some odt and html files that we get in here that are utterly the dog's breakfast.
I can take a crappy Word file, clean it in about 5 minutes, (okay--say 30 to be safe) and output some very tasty html. Yes, I'll have to nuke the 1700 lines of cruft that Word sticks in there in the head, but that takes no time at all.
And you can even use Toxaris' macro. But the bottom line is, you simply have to clean up the Word file. I've used BookCreator for this before we had our own in-house macros and clips, etc., and that works a treat for clean-up. It's a nicely-made macro, or whatever one wants to call it. If the person creating the book in the first place just uses Word's built-in styles, (or modifies them and creates their own, whatever), then the output will be perfectly fine.
It's not Word. It gets a really bad rap, unfairly. It's not rocket science to realize that Words styles map directly to CSS, and to html. So, it is like anything else in computers: GIGO. I don't care if the G is from OO, LO, Word, Wordperfect, Scrivener, LiquidStoryBinder or Bob's Big Butt Bookmaking Business, GIGO.
And really, how deranged do you have to be to write in Textile? WHY? I use Textile all day long, in both my PM (Project Management) system and my CSR system (Desk, which used to be Assistly). Trust me, it ain't that much fun.