Atlantis Word Processor - ePub creation tool

[abeonis]

Congrats awp. Very nice piece of sotware. I'll try to make it pseudo-batch with AHK_L.

[awp]

By the way, there is the French version of Atlantis Word Processor. You can download it from here:

http://www.atlantiswordprocessor.com/fr

[abeonis]

I know awp. I was surprised to find the french version but not the spanish one. Anyway, I prefer to use the english version, it facilitates the communication in Forums.

[JSWolf]

You cannot use FlightCrew as your only error check. It misses things. You have to use epubcheck 3.0.1 to be able to say your ePub is free of structural errors.

[abeonis]

I agree JSWolf, thanks for the warning. Actually it is even worst than that, passing epubcheck doesn´t mean that it will work on my kindle.

I did only a first and brutal 5 minutes run to check whether my PC started issuing some smoke or refused to work or if I needed to learn another gaz-factory software. There was no surprise, the formatting was not 100% good, it would have been a miracle. But the results were very promising.

After a brief check out, watching some of their online videos, If you start an eBook directly with the tool and respect the rules, it should work. This is not my case, I want to reuse the word files I did in the past, these files were all formatted/styled in the same way. If I am able to create my frst mobi mannually, I will be able to reproduce it, then systemize it.

OK, stop dreaming. A more structured evaluation starts now and I am prepared to face some problems. I have to understand how the tool works, specially the CSS stuff. From there, evaluate how much do I need to clean up my word files, how much do I have to regex, etc. The usual fun.

[DaleDe]

Originally Posted by JSWolf

You cannot use FlightCrew as your only error check. It misses things. You have to use epubcheck 3.0.1 to be able to say your ePub is free of structural errors.

I have used epubcheck 3.0.1 and it passes fine, no errors however that is no guarantee that KindleGen will work properly on the files, but so far it works well on files I have prepared properly.

Dale

[abeonis]

Originally Posted by DaleDe

I have used epubcheck 3.0.1 and it passes fine, no errors however that is no guarantee that KindleGen will work properly on the files, but so far it works well on files I have prepared properly.

Dale

Cool. Thanks for the motivation.

[abeonis]

Originally Posted by awp

By the way, through the latest post to the Atlantis blog, you can download a free utility "tweak_epub":
Atlantis Word Processor Blog. Tweaking EPUBs.
(you will find a download link plus instructions there)

It can be used to make minor changes to existing EPUBs.

I am not surprised you felt the need to create this utility. l'll prefer Sigil.

[Toxaris]

Originally Posted by abeonis

I am not surprised you felt the need to create this utility. l'll prefer Sigil.

I prefer Sigil too, but I still have this nifty tool lying there. Sometimes, only sometimes, I need to change things I cannot do in Sigil. Then this tool becomes handy.

[Jellby]

Originally Posted by abeonis

I agree JSWolf, thanks for the warning. Actually it is even worst than that, passing epubcheck doesn´t mean that it will work on my kindle.

Doh! Of course, the Kindle doesn't read ePubs (unless you use a 3rd party app in the Fire)

[abeonis]

Originally Posted by Jellby

Doh! Of course, the Kindle doesn't read ePubs (unless you use a 3rd party app in the Fire)

Thanks to remind me this. I am a total idiot.

[abeonis]

Originally Posted by DaleDe

It is quickly becoming my ePub generation tool. Much less work that Sigil.

Another reason why I still use Sigil is when, for instance KDP, do not want me to use HTML element A but use HTML element B to achieve the same formatting (for instance how to implement anchors).

I have no doubt that awp tracks all these spec related suff and implement them in its epub to mobi module. As I am the ultimate responsible of the quality of my eBook, it makes sense to check whether it is really implemented or not. If it is not, I can use a sequence of regex to implement them myself.

It can also happen to discover a new bug between 2 different versions of Kindlegen. Unless awp is in a very reactive mode and implement the workaround immediately, I will be happy with my regex in Sigil

[abeonis]

Good morning awp,

From KDP Guidelines:
3.10.2 HTML Guideline #2: Anchors Must Be Added Before Formatting Tags
Correct: <a name=”Chapter1”/><h1>Chapter 1</h1>
Incorrect: <h1><a name=”Chapter1”/>Chapter 1</h1>

(By name= I suppose they wanted to say id=)

When atlantis generates the epub, it does the following:
<h2 class="h21"><a id="ncx1"/>Hyperlinks</h2>

If you use this same HTML to be used for the mobi, it violates this rule.
Now, it may work as it is but I want KDP to be happy with me, so I wisely follow their rules.

Now, let's assume that some of us use Sigil after Atlantis and for any reason we have to recreate the TOC, Sigil will add its own id:

<h2 class="h21" id="sigil_toc_id_1"><a id="ncx1"/>Hyperlinks</h2>

So ... unless Sigil causes urticaria to Atlantis, I would suggest that it creates:

<h2 class="h21" id="ncx1"><Hyperlinks</h2>

And everybody will be happy.

[abeonis]

I am not sure the good strategy is to try to exactly reproduce in the mobi what our book looks like in MS-Word or Atlantis. I personnally do not work that way.

I have a master CSS file that I always use because I know it works well and is KDP compliant, so the first thing I have to do is to replace the styles.css by mine and clean the code generated by Atlantis, for instance:

<h3 class="h31" id="ncx3">Internal</h3>

becomes:
<h3 id="ncx3">Internal</h3>

And I´ll do the same for <h2> to <h9>, <ul>, <ol>, <li>

It is a little bit more complicated for <p> where I'll have different classes ... and Atlantis do not use the same style name that MS-Word. (See your answer to Chang).

I have an idea to workaround this I have to check.

In conclusion, and taking into account my last post, the flow that I foresee should work for me is:

MS-Word > .docx > Atlantis > .epub > Script1 > .epub > Sigil > .epub > Script2 w/ Kindlegen > .mobi

And if I do the right things right in my script (in case of urticaria):

MS-Word > .docx > Atlantis > .epub > Script w/ Kindlegen > .mobi

[awp]

clean the code generated by Atlantis, for instance:

<h3 class="h31" id="ncx3">Internal</h3>

becomes:
<h3 id="ncx3">Internal</h3>

I do not think that what you are doing can be called "cleaning". If you have "h3" elements associated with different CSS classes, there are differently formatted "Heading 3" paragraphs in your source document. You are not "cleaning the code". You are removing formatting present in the source document.