Ross Mayfield's comments in the Paying for Software thread have come closest for me.

I buy software when I know it will get better, rather than worse, over time.

There are some pieces of software that I depend on. That software needs to improve. It needs bug fixes, it needs to move to new versions of operating systems, it needs new features, etc. I've been disappointed by software that I've depended on in the past. Symantec took More and stopped development on it. You could argue that this action didn't affect the value of the copy of More that I already had, but I certainly didn't feel that way. I invested a lot of money and effort into a program called Arrange, by a nice group of people in Palo Alto. They didn't make it, and all my data ended up locked up in Arrange. Some years later I switched from the Macintosh to Windows. I used a program called Ecco. Netmanage stopped development on Ecco. There was nothing wrong with the copy of Ecco that I still use. But it doesn't grow, it can't talk to my PocketPC, and its web integration is lacking.

Traditionally, we've assigned the value in software to the software itself. But without the people who wrote the software, the software has markedly less value. I'm more than happy to pay people to keep working on the software (if you like, you can think of this as a service), but my experience (and hard-earned money) shows that paying for the software alone is a dubious value proposition. So to me, the question is: Where is the value, and how do we make sure that when we pay (because as Mark Bernstein points out, we do pay), that we are paying to get the value that we want? I don't have an answer for this yet, but our industry is going to have to come up with one.