That ABA software certification effort (a couple decades ago) was noble in some respects, but destined to fail in many others. It didn’t purport to reach consumer apps, or delve much into ‘substance.’ And apart from the question of whether the lawyer guild should be in the business of granting seals of approval to competitive products/services, the space of options is HUGE, constantly growing, and frequently changing. Verifying the legal correctness of all possible behaviors of even a subset is roughly analogous (in quantity, not quality) to confirming that all the items in a large grocery store have the right price tags on them. Hard to know without checking.

For reasons developed at greater length in a draft conference paper I’m also skeptical that automated test scripts, code transparency, and better documentation will close the gap, however beneficial they will be. We’re going to need a powerful cocktail of quality-assurance processes and tools.