Why do economists decry software patents?

Economists seem to be universally against the concept of software patents.  An iEconomy article in the New York Times in 2012 explained with a specific case study how the present system stifles innovation (albeit it’s a sample of one). The article argued such patents are overly broad and vague.

As reported in IP Value Wire, Mark Lemley of Stamford offered a solution to this:

“…by applying the rules of means-plus-function claims to software, we could begin to get a handle on the software patent issue. Indeed, ending functional claiming may be the only way out of the software patent morass. As long as patentees can claim to own the problem itself – not just the solution – defining better boundaries and invalidating obvious patents won’t do much to make the patent mess go away.”

Why haven’t economists lined up in support of Lemley’s solution? Eric Maskin, a professor of economics at Harvard, suggests that if the problem were simply the way the claims are written, Lemley’s solution would work. Rather, in order for innovation to work the way it should in software development, the patent process should not apply at all. Maskin’s letter to the New York Times explains why.

With software, innovation “is highly sequential: progress is typically made through a large number of small steps, each building on the previous ones.” Running into a patent barrier at any step requires license negotiations, additional fees, and/or litigation plus additional time. In short, a patent holder can delay progress and make it prohibitively expensive to move to the next step.

Perhaps there is a step between no patents, with inherently less incentives, and where we are today. As an example, for the most part (with a few notable exceptions), there appears to be goodwill amongst software innovators in the world of standard essential patents, with its nominal licensing, lack of litigation, and air of respect. Is there a model there to consider?