Writer2xhtml

I used writer2xhtml with Ubuntu 09/10 and OOo 3.2 in order to publish ebooks in EPUB format.
On a whole, it's pretty fast, stable, can process very big and complex files including lots of footnotes, table of contents, etc. I hotly recommend it.

Other comments.

If I understand well what happens, when an OOo file is not totally "clean" (has some overload like italics, graphics and others...), writer2xhtml creates some "sgc" styles to identify them, which in an ideal world, would need to be further processed by the user.

Some overload is pretty necessary. Many times though, it's a user's mistake. Let's begin with these ones. If you informed the user, prior to conversion, that writer2xhtml detected the following sgc styles with a short CSS description of them and a tick box to confirm they are really used, it could save a lot of cleaning work later. If the user dont' understand what it's all about... just click "next"

As for the really used overload styles.
- graphics are "magically" processed. I would advise, when possible though, to avoid this feature which is a little code heavy because it creates a new style for every image... and to process directly the insert of images in html code view.
- italics are a sensitive spot which probably may be refined further. I understand there is no perfect solution, but after conversion I had to add spaces before and after. After this, I had to take out space before point and comas (French typographic rules). The most awkward was dealing with a short Italic sentence including a so-called apostrophe. It stops the sgc style and resumes it immediately after. So you get two of the same sgc to describe one sentence. (ex:Vierge de l'Ascension)
- do you really need a dedicated style to process the ° character ? (like in 36° 12' or like in n° 12)

Thanks you very much for providing a wonderful and efficient tool.