IDL Routine: COMPLEXITY_2

If you send me <[email protected]> your email address, I'll let you know of any major bugs that are found in this library.

13 Mar 2002: Some hints regarding use of the routine (all discussion is based on runs on a Sun workstation):

• Memory requirements: A partial run for word length of 16 was done, which indicated the maximum needed for that case is approximately 600 Mb.
• As a practical matter, analysis of sequences using word lengths longer than 14-15 take too long on a workstation to be useful.

12 Mar 2002: Discovered last version (11 Mar 2002) is also too slow. Yet another new version made, to try and speed up computation while conserving memory.

11 Mar 2002: Discovered that versions after 2 Feb 2002 and prior to today utilize an inordinate amount of memory for word lengths larger than n=13. This is corrected, and routine runs approximately 46% the speed as the 2 Feb 2002 version (at longer word lengths). New FIND_WORDS is also utilized.

5 Mar 2002: More changes were made to increase calculation speed by another 30% or so. A bug was also fixed that affected the previous version (RCS complexity_2.pro,v 1.5 2002/02/28 22:37:04) which prevented the routine from executing if no forbidden sub-words exist. Note that this bug did not affect the original release (dated 2 Feb 2002) of the routine.

28 Feb 2002: A few changes were made to increase calculation speed by approximately 40%.