Bytes By The Quintillion For Today And Tomorrow

According to estimates offered at the conference, the world churns out new digital information equivalent to the entire collection of the U.S. Library of Congress every 15 minutes. Such a proliferation of information in digital format, occurring almost 100 times a day, adds up to approximately five exabytes (five quintillion bytes or five billion gigabytes) a year. Unlike information stored on paper, however, this digital information can disappear almost instantaneously. Major historical artifacts such as original homepages of breakthrough e-commerce sites are already gone. Photographic records, stored digitally on disks, are in jeopardy of decay in as short a time as five years. At the same time, the rapid pace of technological change, itself, makes it difficult to understand documents preserved in earlier formats.

Participants agreed on the need to build a business case to offer companies in areas such as manufacturing, health care, life sciences, law and defense an incentive to invest in digital archiving. Such a study would demonstrate how access to archived information is critical to trace design rationale in cases of failure, document engineering changes, support product life-cycle use, investigate accidents, defend against patent infringement, compare new works with earlier versions, facilitate mergers and acquisitions. Arguments for archiving everything from engineering discussions, e-mails, and CAD models to design and production logs and manufacturing process plans would be presented. The study would also explore the cost of not archiving such information by estimating avoidable expenses for errors, recreating the data or reverse engineering, retesting, training, education and lost business.

The workshop reviewed current digital archival techniques as well as prospects for future software and standards in the area. The conference participants also discussed the possibility of collaboration on future digital archiving research projects. A report of the workshop is expected in late spring.

Reference: "Long Term Knowledge Retention Workshop", March 15-16, "2006 Interoperability Week at NIST". Gaithersburg, Md.