I am saying that the real book is the data, the text, but in some cases, other kinds of information are also in the container. It may be a book that is primarily or exclusively images.
I'm saying that we may prefer one container over another. And that one person may prefer one container for one purpose over another container.
The book is Austen's Emma. It may be Austen's text in an 18 century octavio with buckram binding, a 21st century limited edition archive quality hard cover, or an inexpensive paperback, a book club edition, or an ebook in any one of several file formats.
To me, the text is more important than the container.
I am even more interested in the provenance of the text than I am the container. A poorly produced ebook or cheap paperback that uses Austen's text is more valued by me than a beautifully bound hardcover that's an abridged version of Austen's text.
I'll take the Riverside Shakespeare with dog-eared pages, or the Arden digital Shakespeare's Works over The Globe Illustrated Shakespeaer which is a piece of poorly made crap as a book, and as an edition of Shakespeare, doesn't have much to do with what Shakespeare wrote. But if you read reviews, you'll see people talk about what a handsome book The Illustrated Globe Shakespeare is, and how pretty the "leatherette" binding is and how nice it looks on their coffee table.
But you'll have to look a bit to find readers noting that The Globe Illustrated Shakespeare was edited by poltroons, that Cordelia lives at the end of King Lear, or that the text is barely readable because it's set so poorly.
I love books as containers, but the text (or sometimes the image) is more important to me than the container, almost all the time. The public domain Emily Dickinson is Emily Dickinson edited, not the poems as Emily wrote them. Punctuation, meter, and even lines were changed by her editor at will.
Coleridge's edition of Shakespeare is not what Shakespeare wrote, and what Coleridge did to Donne is a shanda.