For those who might be interested, I'll make quick mention of an inessential but quite pleasant piece of software I've used a lot over the years: Powerbook. It essentially comprises four files, two of which I never use and two I use very regularly. It comes with two databases: Powerbook and Strong, and with two opening trees with the same names, because they're built from those databases. The Powerbook database has about 1.5 million games, including only players rated 2100 or above (games from the pre-Elo period have been back-rated, so potentially relevant games by good-to-great historical figures are included), while the Strong database is much more stringent, taking only games with players 2500 and above from the past century.
If you've already got a "complete" database like Mega2011, you won't have any need for the database, but it's useful to use one or the other of the trees for your computer's opening book. The advantage of using this rather than a tree based on some enormous, unrestricted database is at least twofold. First, you'll see what main lines and main sidelines stronger players gravitate to (and with the 2100 cutoff it's far from the "Black to move and draw" realm we sometimes see at the top level), which saves time when you're starting a do-it-yourself opening repertoire. Second, the stats will be more reliable as to the "truth" in a given line, as they won't be affected by hundreds or thousands of games from events like the girls' under-three championship of Antarctica.
One can of course create a similar database on one's own, but it will only go back to around 1970 or so, when the Elo ratings started. (True, there are variations where what happened before 1970 is nearly meaningless anyway, but plenty more where older games are relevant.) So it is a useful tool, because it helps reach that mean between too much and too little information.