Talk:Fourier division

b values with more than two digits

Experimenting with this method myself, it didn't take long to come up with an example (6601 / 650) where some or all of the b values have more than two digits. I think the article needs to discuss how the final b value is constructed from the partial quotients. --Jay (Histrion) (talk • contribs) 13:47, 11 March 2010 (UTC)[reply]

Had a go at adding the text myself. I'm sure it can be improved, though. --Jay (Histrion) (talk • contribs) 14:25, 11 March 2010 (UTC)[reply]

a₁ shouldn't have a leading zero. I've fixed that. Xanthoxyl ^< 15:22, 11 March 2010 (UTC)[reply]

So you're saying that, using

a=123.45

as an example, the correct pairing would be 12,34,50 and not 01,23,45 ? If so, I might dive in and add text that indicates that. (I know that might seem redundant, but it would in contrast with the pairing paradigm used in the square-root calculation algorithm, which is more widely known - in fact, that's why I assumed the

a_{1}

would have a leading zero.) --Jay (Histrion) (talk • contribs) 15:14, 22 March 2010 (UTC)[reply]

Remainders

In the article example, when I perform the calculation as described by hand, 309/31 becomes 9 r 30. User:Xanthoxyl, you say that the quotient 10 r -1 version is preferable because the remainder has a smaller absolute value; do the partial quotients tend to diverge when the remainder of smaller absolute value isn't chosen? If so, what criterion should be used to choose the remainder - should it have an absolute value ≤ half the divisor? --Jay (Histrion) (talk • contribs) 15:19, 22 March 2010 (UTC)[reply]

External Links broken

Only the double division is available at this point. The other two 404. — Preceding unsigned comment added by 68.187.250.187 (talk) 05:15, 11 December 2012 (UTC)[reply]