The thing is that I actually developing desktop software way more than for example writing software for mobile platforms.
The point of this post is really just to hear thoughts from other commercial desktop software developers. :)
For us, copy protection was a problem in itself so we were inclined to look at software as a service. Saas is, at this time, a socially acceptable form of copy protection.
There are lots of points to ponder. I would love to go after say Adobe, or Intuit, but its been tried over and over.
Monopoly is a funny thing. People have thought that they could make an easier version of Quickbooks and break open the market. Perhaps someone will. But one reason no one wants to leave QB is because its hard enough to know accounting and QB. If you start by learning QB then you are scared off of learning another program.
So software opportunities, desktop or otherwise, are about getting new users before they are indoctrinated into hard to use products. So its a long game to earn a monopoly. And if you give them a superior experience they are more likely to be open minded about switching to something else. This is one of the reasons Apple had a difficult time earning PC converts over the years despite consistently low quality software from Microsoft. If Microsoft had ever shipped an excellent product they would have potentially lost customers faster!
When I think about non-tech, but computer-involved relatives, the only one that has purchased software bought Parallels because they needed it for a legacy work program that doesn't play nice with OS X.
The only space I think that it might be worth it would be in games.
SolidWorks, L-Edit and CoventorWare are expensive and require a lot of interaction on the desktop. They are unlikely to be replaced.
Some big software, like Comsol, would be a good candidate for a move to the cloud. Simulations that take a few minutes to hours to run really should be on an elastic cloud server, and should stop being sold as desktop software.
My conclusions are that SAAS and app stores have swallowed the market of general applications and utilities. By "general" I mean, software of mass appeal.
The remaining applications that are still sold as shareware are specialized applications.
I think it would be massive fun to work on the second category of software but it seems like a hard market to break, in general.