Posts

Showing posts from June, 2025

A consistent maximum size for data compression

In February, I wrote about an experiment in data compression. The resulting algorithm is lossless and has an unanticipated virtue: it yields a compression that is consistent across all data types, as it indiscriminately reduces it to the type integer as a primary step. This results in a theoretical max value for storing bulk data. While some data types may respond better to specialized compression strategies, ALL data reduces by the SAME ratio when considered by [modified] "Golumb" compression.  This can apply to storage estimates.