The BeechickSort Algorithm better than quicksort?

We know Quicksort is a efficient sorting Algorithm, now here they say this:

BeechickSort (patent 5,218,700) has these characteristics:

• Sorts two to three times faster than the quicksort algorithm, depending on the list.
• Unlike quicksort algorithms, it provides stable sorting of duplicate keys.
• Whether the list is previously sorted or shuffled makes no difference.
• Uses no compares.
• Uses no swaps.
• Uses no pivot point.
• Works equall开发者_开发知识库y well with short or long lists.
• Is economical with memory.
• The first sorted results are available for other processes almost immediately, while the rest of the list is still being sorted.

Do you know the implementation, or we have to wait until the realese?

---

It appears to be basically a radix sort: that is, classify items by their "most significant part" (leading bits/digits for an integer, first character(s) for a string), then recursively by "less significant" parts. You can do this by, e.g., setting up an array with one entry per possible most-significant part, then doing a single pass over all the items and assigning each to the appropriate element.

Most versions of radix sort actually process the least-significant part first; this turns out to make things easier. "Beechick sort" apparently involves processing the most-significant part first; apparently the inventor has, or claims to have, a novel way of doing this that doesn't incur enough overhead to outweigh the advantage of not needing to process parts of the data that aren't needed to establish the ordering.

You can read the whole thing at http://www.freepatentsonline.com/5218700.pdf if you want to figure out exactly what contribution this patent allegedly makes beyond plain ol' radix sort (which has been pretty well known for ages) don't mind wading through a load of patentese. Alternatively, there's some explanation at http://www.beechick-sort.bizhosting.com/abcsort.html. The latter includes C code for a simple version of the algorithm.