Talk:B-tree and UB-tree
Review of the article "B-tree and UB-tree" authored by Rudolf Bayer
This is an excellent article: very competent, pedagogically well structured, clearly flowing, and very well written. Needless to say that the topic is of extreme importance. This article will likely become a classic reference for the community.
I have only minor comments regarding details of the presentation. The article should be accepted, the minor comments are trivial to accomodate.
As evidence for the extremely wide use of the B-tree family, you cite Comer 1979. It would be good to add another, more recent, source, for example, one of the database system textbooks by Ullman/Widom or Ramakrishnan/Gehrke (in addition to Comer - which is an important reference).
Occasionally, some technical terms are used in variations, that is, in a not perfectly consistent manner. An example is "HD-blocks" vs. "disk blocks". I would prefer if the article used only one term consistently. My personal preference would be disk block. The same holds for "secondary storage" vs. "peripheral storage"; I prefer secondary storage (or disk storage).
At the end of the section on Prefix B-trees, it would be nice to have a reference to the original paper that introduced prefix b-trees.
When mentioning caching (in the Parallel Processing section), I would mention DBS in addition to OS caching.
November 2008 Gerhard Weikum