You should read the whole post, but I wanted to pull one quote out and think it over.
On the one hand, JA is right here. There will be fewer books printed. That will result in fewer books sold through bookstores.
However while that may well result in fewer bookstores the surviving stores will do better.
Follow the logic through:
1) Ebooks claim a greater share of book sales
2) Print runs drop (for most books) to accomodate this
3) Gross physical book sales drop
4) Marginal bookstores close
5) Marginal sales drift
a) away for ever
b) to ebooks
c) to other bookstores
6) Surviving stores will win sales and market share for print
7) Surviving (well run) stores will be more profitable even in declining print markets.
Fewer books printed means fewer sold in bookstores, who will no longer be able to stay open. Without bookstore orders, publishers will print even fewer books. And so on.
via A Newbie’s Guide to Publishing: The Changing Face of Publishing.
Eoin Purcell's Blog 
Except that the Expresso book machine will allow any bookshop to hold an enormous range of books and make them available to customers in the time that it takes you to drink a cup of coffee. So while there will be more ebooks, and probably fewer ‘ordinary’bookshops, specialist bookshops wih Expresso machnes will serve specific markets. And publishers (who adapt) will survive because, contrary to what some self-publishers believe, publshers do add value – an independent editorial eye, which gives readers some assurance that the book will be readable (and grammatically correct), being the most overlooked benefit.
Brian,
I remain somewhat unsure about the business case for Expresso while it remains as expensive as it is.
I reckon AT most (working 14 hrs a day, 7 days a week, 50 weeks of the year) you could see 29,000 books printed through an Espresso. That’s allowing 10 minutes per unit (select, print, finish, sell, next customer). But it would require the machine to be going all the time.
On an asp of €10.70 (the average selling price of books in Irish bookstores in 2009) that’s just over €300K revenue. How likely is that?
Presumably when you count maintenance, power and other costs running the machine is hardly cheap to run though we know from the On Demand site* that materials cost about 1c a page so what’s the likely margin on that and how long will it take to earn back the up front fee.
Eoin
PS They suggest a target of 10k books a year