One of the methods of study which has always been favoured by strong players is to go through games. The laborious method of placing 200 stones on a board from a diagram has shown its worth over the years. In the computer era there is a choice.
It is still much to early too say that the machine-assisted study of games brings the same benefits. It may be that it will influence the training of future professionals, in particular outside the Far East. With a database of games some types of analysis become much easier, perhaps even the "line of least resistance". For example the classification of fuseki patterns is likely to be undertaken in the next decade, with results which cannot at this point be predicted, simply because thousands of games will be available in machine-readable form.
Looking instead at amateur study now, the existing game collections, including the "Games of Go on Disk" (GoGod) range under review, give a chance to acquire and go over many thousands of games of the outstanding players of all times. What one makes of this opportunity remains in the realm of personal choice - there is no consensus at all about how to treat this material, in recent years becoming substantial and available at reasonable prices. (If you are able to regard games downloaded from the Internet as "free" then you are one of the fortunate - but their true cost is in pennies.)
Once armed with enough games, you need a program to read them to you. For the GoGod collections, which are in the Ishi or "standard" file format *.go, I use Many Faces for nice graphics, and Yago for sheer speed, but I don't have opinions about what's the best viewer (see another page  for a list of viewers). More to the point is how to use the resource. I have employed two methods extensively: (i) go through taking notes to keep, and (ii) search games for something particular, to make a classified collection. The first of these evolved from the idea that, by the hand method, one can try to understand "what happened" in a game in about 20 minutes. A machine diminishes that time by a factor of ten, if one's quick on the click, so there is time to admire the scenery. The virtue of (i) is that of an inkblot test - it may reveal what it is that interests you personally in go. That might range from complicated openings to endgame tesuji. It seems essential to be honest about this. No one can study everything at once, and if you can extract most of what's in a professional game you certainly don't need my advice.
Which brings us to method (ii). Again it is better to trawl a big collection for material you really find interesting, than to try to classify each game under an excessive number of headings. If you must study complex joseki, try for example the GoGod Shusai games for taisha variations, and make a directory/folder of ones to which you'll return. Since I now spend my time on projects of this sort, I find that comments on the games aren't so important. I enjoy the sprinkling of commented games in these collections in other ways when I come across them (they are easily identified from the file name).
Looking now at the GoGod disks in detail, I have bought most of these (brought to you by T. Mark Hall and John Fairbairn). I like best the career collections of Shusai and Go Seigen, and the 1000 games of Kitani. I also found the Korean Yearbook games a revelation. A little computer skill, which I didn't initially possess, is needed to copy the games from disk and arrange them conveniently. Chronological order is used to index the games. Other aids to searching one has to provide. Given the low cost per game (2p per Kitani, for example) this isn't unreasonable.
Some players don't enjoy professional games. To them I say "Is it that you expect to understand it all? Why not look for those parts you can understand? Not what you cannot, for that you will surely find."
review by Charles Matthews, December 1997