The Metaprocess Puzzle
September 29, 2012
BeyeNetwork suggests one reason metadata is not implemented comprehensively or well: “Lack of Metaprocess Information Impedes Ability to Collect Metadata.” Writer and database management expert Bill Inmon pins the lack of enterprise-wide metadata primarily on a lack of metaprocess information. Metaprocess covers high-level descriptive details about a process, like its name, the technology that houses it, its input and output, and algorithmic variables. It is pointless, Inmon insists, to attempt to understand a large organization’s information flow without this information.
Why is metaprocess information so hard to come by? The article explains:
“It resides in the old legacy code. In COBOL. In assembler. In AS/400 modules. In PL/1. In technology that has not seen the light of day in decades. Once there were technicians that could be hired to read and go through the old code. Today those technicians have retired or have been promoted to management positions. In another generation, it won’t even be possible to find anyone who understands these older technologies. And by that time, SQL and C++ will be the old legacy technologies of the day.”
How does one solve a metaprocess problem? What is the meta-metaprocess? Inmon doesn’t really have an answer to that. He does suggest that, since legacy code is a form of text, someone may someday find a way to coax this information from a text editor. Anyone up for the challenge?
Cynthia Murrell, September 29, 2012