This is what I was worried about on January 1st, 2000. An unknown single point failure leading to a massive cascading blackout. The good news is that the system upgrades put in place for Y2K seem to have worked and protected the system from eating itself.
Den Beste has a good article on the restart problems. The only thing I'd have to add to it is that there is also a problem with unexpected current flows. I spen a bunch of time in the early nineties working with the provincial power utility (B. C. Hydro) doing a system to manage maintenance issues. One of the things I learned about the power transmission system is that the current doesn't necessarily flow down the transmission lines that you would expect. It follows the path of least resistance (sort of, AC current is more complex than that) and in a sufficiently complex transmission grid you don't know where that is. You can however figure it out using a computer and lots of linear equations.
In the old pre-computer days the grid was designed simply and inefficiently so you did know what would happen, but with modern computer control systems, competitive power sales, and long distance power wheeling they've let it become complex enough to give them suprises. Suprises like all the current going from B.C. to California routing through 1 set of wires, instead of the several sets that do exist, resulting in the protection system taking all of the wires offline in a cascading failure.
As long as the control computers are online it won't matter much in this case, but it's just one more thing you have to pay attention to when doing a grid restart.
Another item of interest is that apparantly several nuc plants tripped offline. I'm not sure how far offline that is, but in the worst case it can take weeks to restart them, operating by the book.
