Hmmm - the potential is great, but I dunno....
Columbia has one of THE premier collection of shorts from the 30's and early 40's. So far, they've clearly transfered a significant number of Stooge shorts from film negative to moderately well restored digital masters. They've done the same with the Buster Keaton shorts from the 30's.
The two open issues here are:
1) How many of the shorts in their library are either available in a form that lends itself to moderately good digital transfer - or have already been transfered. My sense is they will largely be putting up for On-Demand sales those items they've already transfered, though the fact the release mentions "Classics and older TV shows never before released on DVD" indicates they either have existing digital transfers they've not released yet (this certainly is true to some extent) and/or they are reviewing the entire existing library of negative or positive prints that are in moderate good shape to decide which can turn the greatest profit the most quickly.
2) What price premium will they add to these On-Demand products to monetize the clear interest a customer exhibits by coming to the service in search of something they want and can't easily obtain. We all know from the arguments over inflated CD costs that once the digital master exists, the cost per copy to the manufacturer is minimal. However, when coming to such a service, a customer is announcing their special interest, and you can be certain Sony will want to make $$$ from that interest. I can remember when I used to shop (and sell) at flea markets ages ago, there was this style of buying that could avoid announcing your interest, so as to be able to barter down the price. If you were over-enthusiastic, the seller would immediately recognize they could hold out and expect to get a good price for the item from you. My sense is the price/short for this On-Demand service won't be anywhere near the price we all paid by getting the premium pre-sale price from Amazon (< $1/short). My low-ball guess this HP On-Demand service which needs to slice profit out for HP too will end up charging ~$10/hr of content or ~$3/short. For the remaining 166 shorts, that's would mean paying nearly $500 total, as opposed to ~$150. A pretty significant difference - and that's assuming they'll even make the effort to do the digital transfers for all the remaining shorts that have not yet been mastered in digital form.