Respectfully, I have to say that your "rational strategy" is a terrible suggestion. This strategy only makes sense if your desired goal is to reduce treble damage risk. Moreover it completely ignores the risk of devoting a ton of resources to developing a product that is already protected by a patent.
They're all protected by patents. All of them. I can pretty much guarantee you that any of IBM, Microsoft, Google, Oracle, Intellectual Ventures, and many others have sufficient patent coverage to be able to sue you for anything you could possibly build that involves software in any way.
Lawsuits aren't always about winning. Sometimes they're about bullying and showing dominance. Sometimes they're about stalling a competing product while you finish your own. In the case of a big company or troll going after a small company, drawn out legal battles are used to bleed the company dry until they settle or go bankrupt.
> "Moreover it completely ignores the risk of devoting a ton of resources to developing a product that is already protected by a patent."
That's what the legal advisers (who recommend this strategy) are for. You bring them a concept or prototype, they do clearance/patentability checks and make a recommendation -- not only in whether to proceed, but potentially including advice on how to proceed in the most legally advantageous manner.
(Your specific invention may be patentable even when generalized to cover additional uses -- indeed much of the value of good patent counsel comes from their ability to get you the most-broad patent possible.)
