Many Cards Versus One Card and a List Field

[Gregory Lypny]

Hello everyone,

Having started this thread, perhaps I can throw in that the question  
for me has always been a practical one.  If I need to work with and  
analyze the data in many records and the characteristics of those  
subsets most of the time (How many in total?, How many are red?, What  
is the median?), then I build something with a list-like or array-like  
structure because an individual record or card isn't really relevant;  
it's the ability to gather them quickly into groups and characterize  
the group.  That's why when I built a program in MetaCard a few years  
ago to index the human genome database (Unigene) and extract subsets  
of the 100,000 plus entries, cards never entered the picture.   
Likewise, when I did the same with 127,000 news releases made by  
Canada NewsWire and associated the words in the headlines with some  
360,000 entries in Webster's Dictionary and the 3,600 firms trading on  
the S&P/TSX.  But when I mostly need to find and study the details of  
an individual record, then the card model is a good one.  However, I'm  
finding that my need to look at individual details is now almost  
always accompanied by a need to aggregate and study groups of records,  
and for databases of, say, 100 records or more, I have to maintain a  
series of lists anyway.

Regards,

	Gregory