Hello,
I have conducted a series of investigations into lexicon data structures, and smaller really does mean faster, but a graph compression algorithm with instantaneous run time is not an area I decided to research until now. Initially, I was primarily concerned with the functionality and performance of the data-structure AFTER it had already been created. It then struck me that certain search algorithms could benefit from active compression of meticulous records dealing with investigated branches.
The Blitzkrieg Trie Attack DAWG Genesis Algorithm is extremely fast.
Using CRC calculations, node group segmentation, and multilevel stable merge-sorting turn a process that used to take several minutes into one that seems to complete in less time than a monitor's refresh rate.
I documented the algorithm here, and I am interested to know if I have stumbled upon a Google trade secret?
http://www.pathcom.com/~vadco/dawg.html
All the very best,
JohnPaul Adamovsky
PS - Thank you for running the Blitzkrieg Attack program.