While I agree that you should validate your ideas before spending copious amounts of time programming, I don't think that you should sell people on an imaginary product. Unless you already have an established reputation, it's highly doubtful that people will pre-order something on mere talk alone.
I think that it's a better idea to stick with the MVP (minimum viable product) model and iterate rather than try to figure out revenue and customers right off the bat.
That totally depends on your market. I understand that many programmers are uncomfortable with selling stuff that doesn't exist yet, because to them it feels fraudulent. But there are a lot of ways you can get around that feeling and leave customers feeling happy.
A good example is Kickstarter. People sell imaginary products on there all the time. And this has gone on in the software business for a long time. Microsoft, for example, got their start selling a product that didn't exist:
I think that approach was a little dubious (although it's hard to argue with the results), but I'm perfectly happy with Kickstarter. For me, the difference is that with Kickstarter they're being honest.
There are also things you can do that are in between. For example, the landing page that talks about a "coming soon" product is fine by me. As is a guerrilla user test where you have people look at several things, some of which are products that don't exist yet. Or getting your clients to sign a letter saying, "If you build X, we will pay you Y for it."
And honestly, people do order things on talk alone. A lot of enterprise software starts out that way. It seems crazy from a programmer perspective, but if you're running a business you frequently make small bets on new suppliers to see whether they can deliver. And sometimes large bets if they have something you really want.
There's a difference between a minimum viable product and a minimum viable idea. It sounds like the OP wants to sell the latter before building the former. Or, as the saying goes, put the cart before the horse.
Could be that an MVI is any idea that can be used to secure funding, in which case it would put the concept in league with a rather unsavory brotherhood.
Talk itself (that is, marketing) can be a "minimum viable product". Enough to make sales. But I think that's more valuable in the Kickstarter world of making things you can kick. In software, the standard is usually (not always) higher. As a consumer myself, I'm a lot more willing to pay for the promise of a physical product, than to pay for the promise of software.
I think that it's a better idea to stick with the MVP (minimum viable product) model and iterate rather than try to figure out revenue and customers right off the bat.