Abstract: a sort/merge program incorporating a technique for optimizing the number of strings processed in phase 2, a recording handling technique for minimizing the number of records handled in phase 2, a flexible blocking technique and a flexible merge order technique whereby these parameters are reset to their best values subsequent to the internal sort and where the merge order of phase 2 may differ from the merge order of phase 3, a floating buffer technique to place next blocks of records into core prior to merging, a smart look-ahead technique to place those blocks needed next into the spare buffers, a distribution technique whereby work space on auxiliary storage disk devices can be regarded as a single work space regardless of the physical location of the storage spaces, a track interleave technique whereby each sequence is begun on a separate track, a block address chaining technique and a write-in-place technique for minimizing arm movements.