I must be missing something!
Why would they go through all the trouble of scanning a paperback when the original was probably submitted in some digital form, ie. MS doc, etc (this applies to more recently published books)? I would think that this file would go through the editting process and saved and if they are anything like me, archived as well). It is this file that would then be formatted into whatever ebook format you want. Granted it's been a while since my programming days, but I would think that with the ease of use of the programming languages today, that software could be written to do any additional formatting needing for e-publishing. It would be great if someone would then physically edit the e-file for chapter breaks, etc., but not totally necessary as software can be written to do that as well. I understand that the process to create a paper book would be different, but I've written and in-house published fairly complex user's guides in both paper and digital formats and the process just wasn't that complicated - and that was 25 years ago.
As a side not - if I were a book publisher, I would want to make sure that any books coming out of my publishing house were availble in all formats so they anyone who wanted to buy it could. Why limit availablity to only those who read ereader, Mobi, etc, or paper for that matter!!