I too was a little disappointed that the author gave short shrift to her proposed federal index, and I would imagine that maybe it is because she doesn't have a firm grasp of what is actually needed and hasn't actually thought out in detail how such a new and improved system would work. 

Your post  reason  why I tend to think that a federal system as she proposes would likely fail.  The folks who would have the knowledge and experience to help design a new and improved system  will generally have no interest in seeing a federal system succeed, because it would make their hard-won experience gaining knowledge of the local systems worthless.

As far as the idea of "if it ain't broke don't fix it," I'd argue that existing systems really are probably due for some fixing.  In my admittedly rather limited experience working with the Richland County Ohio land records, i found the computer system to be pretty "old school" and lacking in very basic niceties that would have helped me find what I was looking for.  There's no crosslinking of mortgage assignments to mortgages or deeds for instance.  So I think you have a system with the haves and the have nots as far as current local systems.

I continue to believe that there has to be a better way to be upgrading digital land records systems than to be re-inventing the wheel 3000+ times over, once for each location, but I am not prepared to roll out a proposal with any specificity at this moment :-)