Today's Guardian has an obituary of Geoff Tootill, one of the creators of the "Baby" computer in Manchester in 1948. It claims that the computer took 52 minutes to calculate the highest factor of 2^18. I know computers were slow in those days, but...
Curiously, I remember seeing a similar error on the 50th anniversary of this computer. We used to have a computing department house magazine at work and someone wrote about this computer and said it took some length of time to determine if 2^32-1 was prime. Well, I think it was divisible by 3.
UPDATE
The Guardian contacted me last night to ask if I had a reference as to why I thought 2^18 was wrong. They pointed me to several sites referencing this number:
http://curation.cs.manchester.ac.uk/computer50/www.computer50.org/mark1/new.baby.htmlSo, did they use 2^18 as a test, because it was trivial to find the correct answer - 131072 - and write a brute force program just dividing by larger and larger numbers to see what the largest factor was?
It occurs to me that the first computer I ever programmed had 18 bit words and couldn't handle integers of 2^18 and larger.