Here is how to reproduce an example showing that slidesort is not
an all pairs search algorithm, contrary to the claims of the
original article.
http://bioinformatics.oxfordjournals.org/content/27/4/464.full

1. Make sure that the dataset consists of 50 pairs of sequences at 
a Levenshtein distance of 3 or less.

python compute_Levenshtein_distances.py

The script uses a slow but safe implementation of the Levenshtein
distance and a full comparison of all the sequences in the file.
It should produce 50 pairs of sequences within a Levenshtein distance
of 3.

2. Download slidesort executables.

http://www.cbrc.jp/~shimizu/slidesort/index.php?Download

3. Run slidesort on the input file.

slidesort_v2 -d 3 -i tiny.fasta

The output should specify that slidesort has found 46 pairs.
