After changes to HTML in Sigil...

[kovidgoyal]

As a matter of fact the stylesheet that the calibre conversion produces contains exactly one class for every element that has a different style from the other elements. And it is flat, with no complex inheritance rules, which makes it really easy to tweak. For instance if you have 7 distinct paragraph styles used in your word document, and you dont define any per-paragraph overrides, there will be seven CSS classes in the output.

But, whatever, the work speaks for itself, I was just annoyed by your use of the words "chaos and counter intuitive". If you did not intend to attack calibre, then do not use words like that.

[Toxaris]

I will check out the current conversion process again. Again, it is my opinion as I directly stated. What you describe sounds solid, but I wonder what the effect will be on a Word document not so nicely built with styles like for example a doc output from ABBYY. Clean Word documents with correctly used styles are scarce to be honest.

[mrmikel]

I think the problem is identifying in the epub what was the style in Word in the first place. They all come out as calibre something. Is there any way to carry over style names, for whatever use they might be? That would make decision making easier as to keep a given style or not.

[kovidgoyal]

Originally Posted by Toxaris

Again, it is my opinion as I directly stated.

When you express your opinion in offensive terms, be prepared to get it thrown right back in your face.

but I wonder what the effect will be on a Word document not so nicely built with styles like for example a doc output from ABBYY.

You cannot magically generate a "clean" document from garbage. In that case when you assert that you are generating a "clean" document, what you really mean is that you are arbitrarily throwing out almost all the formatting of the original. If you want to that in calibre, convert your word document to txt instead of html.

[kovidgoyal]

@mrmikel: No, there is not not. The cascading models for how styles work in docx and css are completely different. In docx styles are composed at the style level, in CSS at the element level. Therefore, for all but the simplest cases you cannot map one to another directly.

[mrmikel]

They are just about antithetical, aren't they? Sigh. That probably accounts for what confusion exists. We are fortunate you can take a left universe document and produce a right universe book.

[Toxaris]

Originally Posted by kovidgoyal

You cannot magically generate a "clean" document from garbage. In that case when you assert that you are generating a "clean" document, what you really mean is that you are arbitrarily throwing out almost all the formatting of the original. If you want to that in calibre, convert your word document to txt instead of html.

No, you can't indeed. You are correct that I am throwing out formatting of the original, but not all. I throw out things like margins and font-sizes, not styling like centering, italic and bold and more. That was done on purpose, because those are in my opinion more important than margins for most books. The other formatting can then be set in the stylesheets.
The txt format looses of course everything.

Fortunately there are more ways to skin a cat and each can choose the one best suited for his/her own uses.

[kovidgoyal]

Originally Posted by Toxaris

The txt format looses of course everything

No, it doesn't. You can choose to output either markdown or textile formatting in which case inline styling like bold/italic/etc., links and (IIRC) even images will be preserved.

[theducks]

Originally Posted by Education

Toxaris, you mention: "Once the ePUB is created, make any changes there". The issue is that I am not a html pro (I am just writing a book), but once I start editing directly in Sigil, it seems that it will soon require me to have a good working knowledge of html.

Take 10 and stop and look at the HTML side of one of your properly formatted chapters in Sigil.

Complicated? H no! English Grammar is way harder.

Messy code? Probably, if it came from word. (IMHO Messy is unnecessary style divisions: <p class="indented"><span class="none1"> , when the attributes from 'none1' could be combined into 'indented')


CSS is where the real style work gets done (avoid those styling buttons in Sigil) They generate in FILE SGC codes (but valid CSS coding if you need it).
copy (and but RENAME) a SGC# block to your stylesheet as a starter. RENAME the SGC code to something else Easy to remember. both on the stylesheet and on the page (file) it came from. SGC# code style assignments DO NOT span document files . The same code may do something entirely different in the next chapter (file).
Use Sigils 'Clip Editor' to create your own styling menu (and hid those similar buttons in Sigil: view: <remove the tick from the appropriate toolbar>.

Again, for most cases, the CSS you will be using in the books body is very basic (not like a WEB page, which uses the full range of CSS).

Don't know HOW (or have not decided) how to do a styleing, just assign and USE a separate stylesheet class for that block. eg.

.chapter title { }

will show up with the default look for now , you can fill in the STYLESHEET details later. All the places that were used will follow (the beauty of a single place style assignment) .

Your main job is to get the words all in their proper places (lines) on the page.

NB For years, Webscriptions (Baen) shipped e-books with a TOTAL of 40 lines of CSS code . Nothing fancy was ever done, just a Basic, simple book layout.

[JLius]

I didn't know a thing about html and css when I started reading up on it two months ago, with the intention of creating epub out of docx-documents. By now I know enough to clean up the CSS/html of "easy" novels (without to much styling, images etc). I just found out about this forum last week, people are really helpfull here when you're stuck on something - and if you want a novice's help, then I'll try and answer every question you've got :-)