That's an interesting take on the reason to bill hourly, and I know a lot of developers who do that. Personally I've switched to mainly doing fixed-bid contracts.

Rather than selling my time, a fixed-bid contract allows me to sell my expertise -- and that often is more lucrative. For example, a client may need a solution to putting hundreds of documents online and making them searchable. That may only be 40 hours of work, but it's a problem that has been costing the client tens of thousands of dollars in lost productivity each year. As an expert, I can sell them a solution to their problem in a way that saves them money and is still profitable for me.

Beyond that, a well-written fixed bid contract lets everyone know exactly what the costs and expectations are, and what the schedule is. No more having projects drag on and conflicting with other projects I need to be focusing on when we all know that there's a set deadline for the delivery of a specific solution.

Of course, doing this successfully requires knowing what questions to ask to get a detailed spec, understanding how long it'll take you to really do something, and anticipating the gray areas.

I also like to build in "flex time" into fixed-bid contracts. For instance, I might specify that the contract includes "20 hours of revisions" to the requirements once the client has had a chance to review the prototype. I'll price this into the estimate. This way the client knows they'll have a chance to make changes, but they also know there's a limit after which the project cost increases.

I have also experienced the problematic client you describe. The client had neverending revisions. He could never present me with a final list, even after all items in the contract were complete. He then refused to pay, unless I added additional features not included in the contract. Even after adding some of these, he was never satisfied. He made a long list of false claims against me and threatened to sue unless I added more features. After this, I immediately halted work and (successfully) pursued the rest of the money in court.

My conclusion is that fixed-price bids are bad only if there is a lot of risk -- in this case, the bid amount was far lower than it should have been, the client was underfunded, etc.

Nowadays, I recommend making fixed-price bids that are large overestimates. I am taking a risk in making a fixed bid, and in return the client gets the surety of a fixed amount. If my fixed bid is supposed to literally be my average expected time times my hourly rate, then there is a 50/50 chance (or likely more depending on the client) I will lose money on the deal!