Moved from Opening systematic classification.
I agree that a system of categorising corner joseki and middlegame joseki could be very useful. For example, coded references to joseki could be included under the heading "game information", so that one could search a database for games which feature the joseki one is interested in. However, Charles's proposal as it stands looks a little bit unwieldy: his system seems to be more a way of recording sequences in shorthand than one of classifying those sequences. I propose developing a code similar to that used by modern chess opening encyclopedias, in which sequences are named A00, A01 and so on. The letter headings would be used to indicate which type of corner opening or middlegame sequence is to be found beneath. The numbers, and possibly letter sub-headings, would be used to denote specific lines of play. Here are two charts indicating ways in which this system might work:
A 3-3 point openings
B 3-4 point openings
C 4-4 point openings
D 5-3 point openings
E 5-4 point openings
F Middlegame joseki based on the two-space jump on the third line
G Middlegame joseki based on the extra-large knight's move between third and fourth lines
A00-A09 3-3 point. Opponent plays shoulder hit.
A10-19 3-3 point. Variations in which the opponent approaches with a small knight's move.
A20-29 3-3 point. Variation with a large knight's move approach.
A30-39 3-3 point. Variations in which the opponent tenukis.
B00-01 3-4 point. Variations with low one-space approach.
B10-19 3-4 point. Variations with high one-space approach.
-29 3-4 point. More variations beginning with high one-space approach.
C00 AND SO ON!
If headings such as A00, D10-19, F30-39 and such like do not provide adequate space to contain all the important variations of an opening, then the system could be modified quite easily. For example, openings with many variations could use 3 or even 4 numbers, e.g., C100-199 or H1000-1999.
Admittedly, an abstract heading such as F15 does not give any information as to what may be found underneath it, but this does not present any difficulties that would not be encountered with Charles's system. After all, one needs to know his rules in order to decipher a classification like "_c_AZAp" - it is not obvious to the uninitiated what it means. I am sure that once the user becomes familiar with the alphanumeric scheme that I have outlined, then its meaning and usefulness will become quite transparent. Anyway, it's only a suggestion, and personally I'm quite content to refer to joseki longhand (e.g., hoshi point with knight's move approach), but for those who require such a thing this might prove a simple and workable form of shorthand.