Comments on: A reward check from Donald Knuth http://www.joaoff.com/2008/02/25/a-reward-check-from-donald-knuth/ Programming, Algorithms, and Mathematics Tue, 06 Jan 2009 15:24:08 +0000 http://wordpress.org/?v=2.6.1 By: jff http://www.joaoff.com/2008/02/25/a-reward-check-from-donald-knuth/#comment-811 jff Sun, 02 Mar 2008 23:11:33 +0000 http://www.joaoff.com/2008/02/25/a-reward-check-from-donald-knuth/#comment-811 Dear Mark, a quick browsing shows that on page 485 of Volume 3, Knuth writes "Deletion from B-trees is only slightly more complicated than insertion (see exercise 6)". The exercise 6 is to design an deletion algorithm for multiway B-trees, and the solution doesn't mention anything about splitting. Anyway, multiway B-trees are unknown to me, but I'll definitely read Knuth's section on them now. If I find any reference to you, I'll let you know! Thanks for your comment, Joao Dear Mark,

a quick browsing shows that on page 485 of Volume 3, Knuth writes “Deletion from B-trees is only slightly more complicated than insertion (see exercise 6)”. The exercise 6 is to design an deletion algorithm for multiway B-trees, and the solution doesn’t mention anything about splitting.

Anyway, multiway B-trees are unknown to me, but I’ll definitely read Knuth’s section on them now. If I find any reference to you, I’ll let you know!

Thanks for your comment,
Joao

]]> By: Mark A. Baldridge http://www.joaoff.com/2008/02/25/a-reward-check-from-donald-knuth/#comment-808 Mark A. Baldridge Fri, 29 Feb 2008 17:55:06 +0000 http://www.joaoff.com/2008/02/25/a-reward-check-from-donald-knuth/#comment-808 That is great! I received on a few years back for providing an example of how a multi-way B-Tree could split during a delete! His response included the "sorry it took so long to get back to you - I have been busy on LaTex" (It was 20 years? Good thing I had not moved in that time frame!) It went on to say that he had taken notes and was returning the original note (along with the check). I never got back to see if I received acknowledgement in a subsequent edition, or even a problem for the issue. Splitting can occur when you get too many entries in an internal node. Also, if you allow variable-length elements, a deletion and rearrangement of elements can cause overflow in the list of elements rather than the number of elements, which forces a node split. That is great! I received on a few years back for providing an example of how a multi-way B-Tree could split during a delete! His response included the “sorry it took so long to get back to you - I have been busy on LaTex” (It was 20 years? Good thing I had not moved in that time frame!) It went on to say that he had taken notes and was returning the original note (along with the check). I never got back to see if I received acknowledgement in a subsequent edition, or even a problem for the issue.
Splitting can occur when you get too many entries in an internal node. Also, if you allow variable-length elements, a deletion and rearrangement of elements can cause overflow in the list of elements rather than the number of elements, which forces a node split.

]]>