[Katsunami]

For this once, I'm agreeing with JSWolf.

If plain text is an e-book format, you can count every document containing text to be an e-book. Plain Text, doc, docx, pdf, rtf... everything.

In my view, an e-book format is a format specifically designed to be able to:

1. Allow reflowable text
2. Mimic a real book, including markup, images, front cover, and metadata such as author, title, blurb...
3. Be read on devices with varying screen sizes, mainly due to feature 1, and/or by repositioning images.

PDF: Very often missing feature 1, and thus feature 3.
Plain Text: Missing feature 2 entirely.
RTF: Missing parts of feature 2 (no images, can have comments in the text, but not all implementations support metadata AFAIK)
doc/docx: Like PDF, often misses feature 1 and/or 3.

It's true you can use files such as TXT, PDF, or RTF as e-books (and I think RTF would actually be the best here), but they were never designed with that goal in mind, and thus to me, they're not e-book formats. Using those files as e-books is the same as using Excel as a database: possible, but not as it was intended.