You’re ignoring the wish of the customer. A copy of Windows is 100$~300$. Given a copy of the software for 5000$ with guarantees, which one would the average home customer select?

Customers with specific up-time needs negotatiate for more than what the EULA asks. For example, 1 M$ per hour of unexpected down service assuming a 99.9… availability. This works out nicely, it forces the software vendor to ship a higher ‘quality’ solution however it also allows the software vendor to limit the use of the product by the customer. E.g. only specific hardware, drivers, software pieces.

This wouldn’t work for the general market. The white box PC with unknown components is the best example of looking for trouble.

If there was no such thing as product liability and failure was the accepted norm, would people be complaining about the high costs of liability? Is Caveat Emptor the ideal doctrine? I think there should be norms regarding what is acceptable and what constitutes negligence or fault on the part of a software manufacturer or individual programmer.

It’s a matter of policy really. Yes there is an economic cost that will be passed on to the end user. But I’m wondering if were not approaching a point were that initial lower cost to the buyer/licensee is offset by the subsequent high costs to society in general.
I found some stats for 2001. http://www.toplayer.com/content/cm/news199.jsp and it seems there is a fairly wide margin to play in.

The issue of Open Source came up: wouldn’t this preclude non-commercial development? Maybe, but I could imagine a doctrine under which, since the end user has way more control and information on a product, having access to the code and all, that the burden of judging the fitness of the product might be partially shifted to the end user.