This was common jargon in our programming shop in the early 1980s in the same sense that Crescend mentions. Anytime a program was adding things into different totals, it was common to refer to the various totals as “buckets” when talking about the program.
For example, in A/R the buckets might be the customer’s total balance, the amount past due, the amount 30 days old, etc. In G/L the buckets were the period ending balance, the current period amount, the balance forward, etc.
This terminology sometimes carried over into casual conversation about the business as well. But we were a small company started and run by a programmer.
I first came across it as a term other than a pail to carry things in around 1991 or so (though I’ve no doubt it’s older). I was working at the time in the collections dept for a credit card company, and customers were divided into buckets: 30 days delinquent was bucket 1, 60 days delinquent was bucket 2, 90 days was bucket 3, etc. I think it went to bucket 6 IIRC.
From “Addressing for Random-Access Storage”, IBM Research Journal, 1957:
Interestingly enough, this is very, very similar to the last context where I used buckets (a duplicate-merge algorithm for large datasets), though instead of hoping that any identification numbers were uniformly distributed, I used a proper pseudorandom hash function. I’ll try to find an earlier cite - I want to know who put the buckets in computer science.
Like Rube E. Tewesday and Balthisar, I never hear this in a casual context. The only example I can think of is a reference to “raining buckets”. But, given that a bucket is what one would place under a leaky ceiling, I don’t think that’s what the OP means.
