Is documented consensus a simple language describing the Maimunkov and Mazier Cadmelia system?

Question

I struggle with the original paper, trying to understand what would seem to be contrary to the paragraph. One example is that in 2.2, the authors state that 160 for the bit space will consist of 160 thousand buckets, and then further say that in fact, buckets are a smaller number spanning wider bit ranges and organized by binary prefix trees. This section 2.4 deals with unbalanced trees, which lead to interpretations as follows: https://stackoverflow.com/a/2126239/220 , where it is not clear whether the answer reflects their intentions MM. Is there a clear consensus that explains how these ambiguities should be interpreted in plain English?

Answer 1

From the homepage of David Mazier [allocated place]:

Peter Maimounkov and David Mazier. Kademlia: A peer-to-peer information system based on the XOR metric. In the materials of the 1st International Workshop on Peer-to-Peer Systems (IPTPS '02), pp. 53-65, March 2002, paper . (Short versions are often given before the trial , but please read the full article.)

Since quoted links point to postscript, here is the full version PDF

The 2.2 sections in the longer 13-page version contain many improvements and refinements that are not part of the original evidence.

Thus, a flat array of 160 blocks can be considered as Kademlia 0.9, which was necessary for the basic proof, while the tree-based version is Kademlia 1.0, which is needed to implement these improvements.

Note that the tree-based and flat approach is almost equivalent if you don't implement any of the things in later sections, such as unbalanced tree processing or segmentation.

Is there a clearly documented consensus explaining how these ambiguities should be interpreted in plain English?

Not that I knew this, but from the above, I would say that the later sections are simply superior to the previous ones.