"Handbook of Reliability Engineering and Management", Second Edition, Edited by W. Grant Ireson, Clyde F. Coombs, Jr. and Richard Y. Moss, McGraw Hill, 1996.This is the second edition of the Handbook, which first appeared in 1988. There are individual contributing authors for each of the 27 chapters; the resultant is already out of date. The current "Limbo" status of military documents, such as Military Standard 105 and Military Handbook 217F, makes this present edition of the Handbook out of date in 2000. Will Military Handbook 217F ever be revised to provide up-to-date models for the failure rates of modern electronic components?
This Handbook would benefit from better editing. For example, as a Handbook, it is expected that the index in the back of the physical book would contain page references to pertinent reliability terms and techniques. Today's technology emphasizes Environmental Stress Screening, ESS. The term, ESS, has been around for at least 15 years (or so: I recall doing an IEEE paper on ESS in 1988). It is expected that ESS and related terms such as HALT (Highly Accelerated Life Testing) and HASS, would be readily found in the Handbook's index. They are not.
Same thing holds true for Software Reliability, where the chapter's author uses the cute term, "SMERFS Model" (p. 22.15), but this term will not be found in the index. (SMERFS = Statistical Modeling and estimation of reliability functions for software.) By the way, from a technical point of view, this chapter's half page on the various models for software reliability is very skimpy. For example, Dr. Michael Elbert has written an entire IEEE paper on the selection of the proper model for software reliability. It would be expected that a Handbook would be more complete than just a small portion of a page. In Chapter 22, on Software Reliability, the author references a paper on the Rayleigh curve by "Gaffney" (p. 22.13). The reference at the back of Chapter 22 gives the reference's name as "John Gafney". This is a discrepancy that should have been caught by the editors. I suspect that the correct spelling is "Gaffney", but both can not be correct.
There are other editorial lapses: on page 16.24, Bellcore failure rates are compared to MIL-HDBK-217, and it is state that Bellcore "... provides generally better failure rates than does MIL-HDBK-217F, which is supposedly based on field experience in communications equipment." The way this sentence is presented implies that MIL-HDBK-217F is based upon experience in communications equipment, when, as most reliability practitioners know, MIL-HDBK-217F failure rates are based on environments from Ground, Fixed, to Ground, Benign, to Naval Sheltered and Airborne, etc, and on equipment from radios to radars to sonars, air data computers and fire control units. This inconsistency in the use of the language ought to have been caught by the editors.
On page 16.6, there is a nice comparison of FITs to failures per million hours to %failure per 1000 hours. The Editors should make the reader aware that a billion in American usage (1 with nine zeroes after it) is not the same as European usage. Chapter 6, on FMEAs is totally into the Risk Priority Number method where "gray beards" sit around and use the "Delphi" approach to ascertain the probability of an event, the severity of that failure and whether or not the failure can be detected. This is a very subjective method, and, in my humble opinion, RPN is being superceded by more objective, quantitative methods (see IEC 61508).
In summary, prudent purchasers should await the next edition of this Handbook, if that edition considers all the changes taking place in the availability of military standards and the new leadership role of international standards in the reliability arena. For example, IEC 300, on "Dependability" is not even mentioned.
John Peter Rooney, ASQ Certified Reliability Engineer #2425.
