[Doitsu]

@chaot:
Just to add some background information: some older Calibre versions used non-heading tags and special Calibre classes instead of heading tags to identify epub headings, and if you did epub-to-epub conversions with Calibre heading tags were usually replaced with non-heading tags with special Calibre class attributes.
Since Calibre originally didn't use/support heading tags, it made perfect sense to add XPath support and an excellent, fool-proof XPath wizard for users who are not familiar with XPath queries.
Sigil, however, has always used heading tags to identity headings, which, IMHO, made (and makes) perfect sense.

IMHO, headings should be tagged with h1-h6 heading tags, because that's what heading tags are for and moreover some reading apps for visually impaired readers will use heading tags for navigation purposes. While nothing prevents you from formatting headings with non-heading tags and special class attributes, you'll eventually find out that using heading tags will save you a lot of time when you update a book.

I don't really understand why you want to make your life more difficult by using non-heading tags. Why are you so hell-bent on using non-heading tags and XPath???