[dba-Tech] Questions about 2 Unusual Databases

Arthur,

I believe situation one has been done to the nth degree from a brute-force
approach and to just a somewhat lesser degree through an AI/brute-force
approach.

Situation two however I have contemplated many times on my own, but from a
different perspective.  All marketing blitzes and airwave playtime aside...I
have always pondered why exactly one song can become a hit, and others do
not.  In my pondering I wondered if simple waveforms can explain patterns of
acceptance or non-acceptance.  Taken a bit further, does the individuality
of the human ear come into play?  Is it possible that one group of people
"like" a certain sound because their ears "hear" the sound differently.  So
my thoughts were...if I could take all the billboard hits over the past
years and perform a waveform analysis, would similar patterns emerge?


Mark



-----Original Message-----
From: Arthur Fuller [mailto:artful at rogers.com] 
Sent: Wednesday, October 27, 2004 1:10 PM
To: Discussion of Hardware and Software issues
Subject: [dba-Tech] Questions about 2 Unusual Databases


 From time to time I ponder the following two databases, trying to come 
up with the optimal design in terms of both space and performance. This 
is strictly a question of personal interest, and I have no commercial 
interest in either solution. I simply find them interesting problems, 
and I thought I'd trot them out in search of feedback from my colleagues 
here.

1. A database that records chess games. It strikes me that perhaps the 
most compact way to store a game is by using the modern notation for the 
moves themselves. But in addition to recording the sequence of moves, 
the database would also be expected to record situations and be able to 
compare them. I.e. given two sequences, A and B, that both result in 
exactly the same position of pieces, irrespective of the number of moves 
it took to get there, the database should be able to detect this as 
quickly as possible. For example.... aha! This is exactly the same 
position that Bobby Fischer faced in year 19xx, when playing somebody at 
some tournament, but they got here in 11 moves and the current players 
got here in 13 moves. (The idea behind this requirement is that certain 
positions have known solutions, i.e. paths to checkmate.)

2. A music database that records (let's keep it simple in version 1) 
melodies and single-line compositions (i.e. ignoring instrumentation, 
harmony, counterpoint, etc.). The idea here would be to compare any two 
rows and determine whether they are identical. For example, George 
Harrison v. the Ronnettes, for "My Sweet Lord" and "He's So Fine" 
respectively. Ideally, this database should also be able to see past the 
selected key (in the musical sense), and also the tempo (piece A is 
identical to piece B but played twice as fast). Perhaps version 2 could 
also detect that melody A is identical to B except that it is inverted 
(upside down) or perhaps retrograde (backwards) or even retrograde inverted.

Ok, database designers. There you have the specs. Any brilliant ideas 
out there for solutions?

A.

P.S.
Although these are in fact strictly database issues, I am not going to 
cross-post to the AccessD and SQL lists because they are so obviously 
unrelated to the immediate problems most of us have when posting there.

_______________________________________________
dba-Tech mailing list
dba-Tech at databaseadvisors.com
http://databaseadvisors.com/mailman/listinfo/dba-tech
Website: http://www.databaseadvisors.com