Please check my workflow, thanks!

[NicolaNY]

Hello Everyone,

My goal is to upload 2 or 3 books to amazons ebook catalogue. I did a bunch or reading and after realizing all the complications from different formats, coding, conversions, i decided to try and find a simple streamlined approach.

1) Cut and paste my text into Textwrangler
2) Markup the text using html and created a seperate .css file
3) Import file to Calibre and export to mobi. During this process i add metadata, cover image, and TOC.
4) Preview on my kindle and kindlepreviewer
5) Upload to amazon, uncork champagne

The question is whether my workflow is efficient? Are there any redundancies or am i missing important steps to completing a well structured ebook? Are there better tools than Textwrangle and Calibre? Should i be using a more sophisticated html editor? Also, is there any reason to use KindleGen in my process?

Any advise on my process, or perhaps sharing your own, would be greatly appreciated.

PS I am on a mac running osx

Many thanks
N.

[ATDrake]

Yes, that looks fine.

I'm also on a Mac and I mostly use just TextWrangler/Smultron (haven't quite gotten the hang of TW's particular regexp flavour yet) and KindleGen, plus the Mobi2Mobi tools to fix the metadata afterwards. Although plenty of people set store by Sigil, which technically produces ePubs but can be used to make files suitable for Mobi conversion.

A minor nicety you might want to have if your book is divided into sections is to add an NCX TOC, which will give the reader those flickable chapter marks in the progress bar. Although this might already be handled by the Calibre conversion.

For what it's worth, KindleGen on the -c2 setting tends to have tighter compression than Calibre usually does, provided you use the Kindlestrip script afterwards. If file size is an issue (I've heard Amazon charges the publisher extra for the transfer over certain sizes), then it might make a difference.

My personal workflow was posted some months ago in this thread here, if you're morbidly curious. Mind you, I use it purely to fix existing books I've acquired that have something annoyingly wrong with them, rather than to create distribution-quality works.

[NicolaNY]

Originally Posted by ATDrake

Yes, that looks fine.

I'm also on a Mac and I mostly use just TextWrangler/Smultron (haven't quite gotten the hang of TW's particular regexp flavour yet) and KindleGen, plus the Mobi2Mobi tools to fix the metadata afterwards. Although plenty of people set store by Sigil, which technically produces ePubs but can be used to make files suitable for Mobi conversion.

Glad to hear i'm on the right track. The amount of info out there can be a little daunting. The reason i plan on using Calibre over Mobi2Mobi is because the addition of a cover, metadata, and a functional TOC is very simple in Calibre. Is there any additional functionality from Mobi2Mobi?

I actually did familiarize myself with Sigil, and while it's an a fantastic app, i think it creates an unnecessary step. Going from html/css=>epub=>mobi is an additional step with potential additional distrortion in code etc. By cutting Sigil out i can do html/css=>Mobi . . . i'm guessing its cleaner, but i could be wrong.

Originally Posted by ATDrake

A minor nicety you might want to have if your book is divided into sections is to add an NCX TOC, which will give the reader those flickable chapter marks in the progress bar. Although this might already be handled by the Calibre conversion.

I think it's handled by Calibre, but i will look into it. Thanks for the tip.

Originally Posted by ATDrake

For what it's worth, KindleGen on the -c2 setting tends to have tighter compression than Calibre usually does, provided you use the Kindlestrip script afterwards. If file size is an issue (I've heard Amazon charges the publisher extra for the transfer over certain sizes), then it might make a difference.

I've read great things about kindlegen and Kindlestrip. Again, i prefer not working from the command line, and Calibre is not only user freindly, but also pretty tidy. Kindlestrip is super nifty, but my books will only be text plus one image for cover.

Originally Posted by ATDrake

My personal workflow was posted some months ago in this thread here, if you're morbidly curious. Mind you, I use it purely to fix existing books I've acquired that have something annoyingly wrong with them, rather than to create distribution-quality works.

I did check out your workflow, thanks! Truly appreciate your input.

[ATDrake]

Originally Posted by NicolaNY

Is there any additional functionality from Mobi2Mobi?

It doesn't take forever to load on my mini Mac?

Actually, I think there is, since it's specifically targeted at manipulating Mobi metadata and Calibre seems to have fairly generic standard fields for stuff. But most of it will be for things you'll likely never use.

Originally Posted by NicolaNY

Going from html/css=>epub=>mobi is an additional step with potential additional distrortion in code etc. By cutting Sigil out i can do html/css=>Mobi . . . i'm guessing its cleaner, but i could be wrong.

I don't use Sigil either. It's just simpler for me to handle all the markup directly or via helper scripts and assemble/arrange files by hand. But unless you're doing some really weird stuff, there's not going to be any distortion in the code.

The simple level of HTML that Mobi uses will work just fine within an ePub package, although the same may not be true in reverse, because ePubs allow for much more complex display and greater formatting niceties.

But as long as you keep your code clean and stick to that common denominator in your formatting, you'll have books that work very well as both as assembled ePub and generated Mobi.

Originally Posted by NicolaNY

I've read great things about kindlegen and Kindlestrip.

Actually, KindleGen is really, really buggy. But it's Amazon's officially supplied tool for a format that's not open or well-documented at all and the Calibre and Mobi2Mobi people had to reverse-engineer to get their own conversions working.

So there's a possibility that a Calibre or other conversion will be missing some small significant thing which will cause problems when being processed by Amazon or displaying on a Kindle.

Originally Posted by NicolaNY

Kindlestrip is super nifty, but my books will only be text plus one image for cover.

Even so, KindleGen will put in a redundant backup copy of the source files by default, and possibly people might get the wrong impression that your books are longer than they actually are if the file size looks big. But I suppose you can compensate for that by putting some sort of length indicator in the product descrpiton.

Incidentally, if you want to avoid command-line use, there's a nifty AppleScript wrapper that one of our MR members wrote up for KindleGen. I don't use it myself, but it's supposed to be pretty good, and preset to the most useful command-line options.

You can also drag and drop ePub files onto Kindle Previewer which will auto-generate mobi files. But these will be larger (compressed at the -c1 option) and sometimes it chokes on certain code (to be fair, it chokes on the exact same code that KindleGen does, since it seems to be using KG for the conversion anyway).

Anyway, good luck with it all, and if you have any further questions, please don't hesitate to ask.

[NicolaNY]

ATDrake,

Thanks a millions!!!!

I will certainly have more questions, but i'm first gonna read up on mobi2mobi and those scripts.

Best,
N

[DMSmillie]

Hi NicolaNY

Something you need to be aware of, which was brought to my attention a while ago by Hitch, another (extremely knowledgeable) poster here on the MobileRead forums :

Originally Posted by Hitch

I know that Calibre can and does generate an html toc when you convert from ePUB to mobi; but that first-generation mobi isn't saleable on Amazon, as Calibre's mobi's don't leave enough room in the headers for K4PC's second layer of encryption.

Basically, if you want a DRM'd Kindle ebook to be readable using the Kindle for PC application, you can't use a MOBI file created by Calibre. It's fine if you're just creating a MOBI file for your own use on a Kindle, or (as far as I know) if you intend to publish the ebook on Amazon without DRM, but not really ideal if you intend to use DRM and intend it to be a commercially viable ebook available for sale with Amazon, I'm afraid.

If the above applies, you really need to use Kindlegen to create the final MOBI version that you upload to Amazon.

Note: the quote above dates from Nov 2010, but to the best of my knowledge, the info it contains is still accurate.

[DiapDealer]

I've found that Kindlegen is pretty "hit or miss" when translating the css from my epubs to the standard in-line html formatting needed for mobi's. I constantly experiment to determine exactly what works and what doesn't, but it's pretty hard to nail down.

Calibre, however, usually does a better job of translating what I intended with my css (I'm sure I get too carried away with my styling). But I want the final product to be generated by Amazon's official tool for ensured compatibility.

My solution? Build an ePub. Convert with Calibre (epub->mobi)... unpack that mobi with MobiUnpack->MobiML2Html... manually create the toc.ncx (and twiddle the opf) and then feed that opf (and html) to Kindlegen. Strip the resulting mobi with MobiStrip. And finally, tweak the metadata with Mobi2Mobi.

It's convoluted, but I've found that my mobi's converted from epubs tend to more accurately retain the original formatting I've worked so hard on.

[NicolaNY]

Originally Posted by DMSmillie

Hi NicolaNY

Something you need to be aware of, which was brought to my attention a while ago by Hitch, another (extremely knowledgeable) poster here on the MobileRead forums :

Basically, if you want a DRM'd Kindle ebook to be readable using the Kindle for PC application, you can't use a MOBI file created by Calibre. It's fine if you're just creating a MOBI file for your own use on a Kindle, or (as far as I know) if you intend to publish the ebook on Amazon without DRM, but not really ideal if you intend to use DRM and intend it to be a commercially viable ebook available for sale with Amazon, I'm afraid.

If the above applies, you really need to use Kindlegen to create the final MOBI version that you upload to Amazon.

Note: the quote above dates from Nov 2010, but to the best of my knowledge, the info it contains is still accurate.

Yea, hitch is helping me out in another thread, very smart, helpful member.

Thanks for your advice, i will be testing a bit more with kindlegen/kindlepreviewer. Hopefully i'll figure it out because its important to me that the book is saleable on all kindle platforms. Very much appreciate the advice.

[NicolaNY]

Originally Posted by DiapDealer

I've found that Kindlegen is pretty "hit or miss" when translating the css from my epubs to the standard in-line html formatting needed for mobi's. I constantly experiment to determine exactly what works and what doesn't, but it's pretty hard to nail down.

Calibre, however, usually does a better job of translating what I intended with my css (I'm sure I get too carried away with my styling). But I want the final product to be generated by Amazon's official tool for ensured compatibility.

My solution? Build an ePub. Convert with Calibre (epub->mobi)... unpack that mobi with MobiUnpack->MobiML2Html... manually create the toc.ncx (and twiddle the opf) and then feed that opf (and html) to Kindlegen. Strip the resulting mobi with MobiStrip. And finally, tweak the metadata with Mobi2Mobi.



It's convoluted, but I've found that my mobi's converted from epubs tend to more accurately retain the original formatting I've worked so hard on.

Diapdealer, thanks for your input.

Interestingly, i initially wanted to do html/css->epub->.mobi using Textwrangler->Sigil->Calibre, but i thought the extra step of ePub/Sigil could create coding problems. I guess i was wrong and will have to look into this again, perhaps using kindlegen/previewer instead of Calibre. Very interesting.

You own process does seem complicated, but if it gives you exactly what you want then obviously it's well worth it. Gonna give all of this a try, hope you don't mind some questions later on, i promise to research before i ask.