I've written elsewhere that the single most cited saw in the STS literature is the law of path dependency (with its attendant, self-explanatory paradigmatic example, the QWERTY keyboard.
Because choosing paths requires certain trade-offs, with respect to digital infrastructure and its various applications, it is far too early to accept limitation by fiat. Policies, laws and regulations that begin with "thou shalt not ______" should be treated with some suspicion. Those that enable the proliferation of paths that we can later become overdependent upon are worthy of at least occasional consideration; those that foreclose possibilities (thus limiting those very paths) should be viewed skeptically.
An example: although there are reasons why it is not always preferable to have governments involved in planning, building, and running broadband infrastructure, legislation such as that existing in presently 14 states that precludes government involvement in those activities. True, the "patchwork" of models, providers and networks is creating complexities, but they are not preventing innovation, detracting investment, or limiting the proliferation of platforms, applications and technologies.
So what does this have to do with FISA? Well, I have a position on granting immunity retroactively to telecommunications companies that may have too readily divulged their customers' private information.
But I'll stick the accusation aside (for now), and make the observation that the fallout from FISA-telecom imbroglio points to yet another benefit of the "patchwork" approach. In the era of telecom consolidation, when a single provider releases information to the government, the privacy of millions of customers is potentially compromised. If customers are distributed among thousands of providers, then if a single provider acquiesces to government's pressure for revealing proprietary data, the number of affected parties is potentially drastically reduced.
So when it comes to structural firewalls against government overreaching, a large number of small providers may be more effective than a small number of big ones.
