View Single Post
Old 07-05-2012, 01:55 PM   #98
ProDigit
Karmaniac
ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.ProDigit ought to be getting tired of karma fortunes by now.
 
ProDigit's Avatar
 
Posts: 2,157
Karma: 9023682
Join Date: Oct 2008
Location: Miami FL
Device: PRS-505, Jetbook, Jetbook Mini, Jetbook Color, Astak Ez Reader Pro
I have several tools to make HTML.
If you use MS OFFICE or OOO OpenOffice, they add too much crap to a HTML.

Create your document in MS word (or something). Save it as an HTML, open it in a browser, and select and copy all the information on your screen. Then paste it in VMU or KOMPOZER. You can further trim down with these programs if you wish. Once you finish with that, you'd have about 10-30% of your HTML cleaned up.

What I personally do, is after that, do further downtrimming in Notepad++, because of it's superior 'search and replace' feats. It makes it very easy to trim an HTML between 30 upto 50% of it's size, by only keeping the very basics of the HTML.

I don't really care about font, as one can always change it in the reader.

For that reason, what makes less of an importance on e-ink devices is fonts and classes; thus it's one of the first to go on an html.
Get rid of classes, and font changes (size/type) and advanced commands and stick to basic commands. Just keep basic Headings (H1, H2,H3), Perhaps define font type once in the body, text formatting (Bold, Italic, Underlined), and that's it.
text that is Heading automatically will be centered on some e-readers, and you can also get rid of strange numbers, or font colors.

keep pagebreaks (HR; any heading will automatically have a page break before it, so no need to insert it after a chapter), in your case tables (TR), etc.

There are many codes. Keep a list like this close to you:
http://www.htmlcodetutorial.com/quicklist.html

There are also sites out there that allow you to test your code in a browser, but even if it works on a computer, there's no guarantee how well, or even if, it displays on the e-ink device.


Not all HTML codes can be converted to epub, and some auto-converters, like calibre, might substitute codes for other (or similar) codes that may not have the effect you wished to achieve.
It's with trial and error that you'll figure out which codes are transported into Epub and which are not; and which programs transport the most codes. For that reason I prefer a tool I found somewhere on this site, that basically zips the HTML into an epub zip file, after you manually created all other docs (like TOC and so forth).

To have the best (perfect) results, best is to do it manually.
Create a good and clean HTML, trimmed down to the basic necessities of displaying the info, then convert it to Epub. You'd often find that apart from some specific font, or text alignment, applying this principle could save you lots of book space, and reduces auto-conversion errors (which are not solvable).
Doing things manually takes up a lot more time, but gives you control of even the smallest details. In automatic conversion, you basically rely on the convertor to understand or interpret your creation correctly (and you'll not have that annoying BookDesigner/Calibre sticker pasted at the end of your book).

Last edited by ProDigit; 07-05-2012 at 02:07 PM.
ProDigit is offline   Reply With Quote