Unfortunately the paper is riddled with errors, both in the concepts and in the examples. Just a few to mention: - The segment search algorithm in listing 5 is wrong, since it mistakes the usable range for the currently used range - The binary search algorithm in listing 6 (something very basic) is totally wrong - The delete method is incompatible with the binary search. Take as an example the data set in Fig. 3(a), where a search for number '6' would miserably fail, due to deletion and the 'VOID' marker used. - Some running times are badly estimated, even if it was to work (without any deletion on data, which breaks it)

Nice try, maybe could be reworked to be usable (which is not).