Problem converting DOCX to EPUB

[WJSwanepoel]

I have a DOCX that contains numbered items with a frame around it.
I import it into Calibre and then convert to EPUB.
In the Calibre e-book viewer it loses some of the numbers and puts others outside the frame.

However if I view th EPUB in Adobe Digital Editions it does not lose numbers, but still puts some outside the frame.

Attached is:
The DOCX I am trying to covert
The coverted EPUB
A snapshot of how it displays in Adobe.

I would appreciate any assistance.

[WJSwanepoel]

Here is the Calibre log.

[kovidgoyal]

Conversion of lists has limitations as the listing model of docx and HTML are very different, so completely faithful conversion is not possible. I suggest you remove the frames from the source document and then use CSS such as

ol { border: solid 1px currentColor; }

in the extra css under the look & feel section of the conversion dialog, to simulate the frames.

[Quoth]

Also I find it's best to NEVER use a list format, but a suitable paragraph style and manually put the bullets or numbers or letters, because ebooks are more limited in lists than HTML (has to start at 1 and not be interrupted etc).

Using LO Writer and an extra Save As in docx you can have a paragraph style with a box. It isn't a frame, but looks like one and survives even epub to mobi conversion (which surprised me as Mobi has only inline style & HTML3 no CSS). You are limited with how margins work.

Don't use actual frames. Naturally you can't have two boxes side by side, unless they are in table columns, but tables are notoriously poor on smaller screens or user picking larger font. They fail.

[Jaws]

One additional note, due to a really weird quirk in Word (that causes lawyers' blood pressure to rise exponentially when they encounter it):

A "numbered list" and a "multilevel list" look very much the same on the page. The underlying XML code, however, is not — and the "numbered list" (very common in numbered-paragraph legal pleadings) is on the visible first part of the ribbon, but the "multilevel list" (necessary to maintain numbering when copying between documents) is nowhere on the ribbon at all and has to be customized into the ribbon. (Plus Word is actively hostile to the two most-common numbering systems used in legal and government documents, but maybe that's on purpose.)

I've had a lot more success with Word documents (.doc and .docx), and even similar documents originating in LibreOffice, by saving them as HTML in Word/LibreOffice, then running a garbage-stripper on the HTML file before importing into Calibre. It's an extra step that saves lots and lots of time, and preserves the actual numbering scheme.

[WJSwanepoel]

Thank you for all the advice. I simply removed all the borders around my text and everything now appears OK in EPUB.

[Quoth]

Originally Posted by Jaws

A "numbered list" and a "multilevel list" look very much the same on the page.

The reliable solution is to never ever in any program use lists and especially automatic numbered or letter ones. Use regular paragraph styles that look like the required list and manually number/letter the items. Then the docx to epub is 100% perfect.

Only edit in odt in LO Writer and do an extra Save As in docx

Don't use doc.

After importing doc or docx to Word OR LO Writer from somewhere else the styles, links etc all need checked for two reasons:
1) The other user may have been treating a system like a typewriter and using direct or wrong formatting or styles. Anchors need to be only at the start of a paragraph. All tabs, multiple spaces and empty lines need removed and styles used. Footnote links need to be a whole word, not just the number in box [3], otherwise hard to select.

2) MS has multiple doc format and docx format versions. So stuff goes wrong.