the past is the future

1
IEEE TRANSACTIONS ON RELIABILITY, VOL. R-33, NO. 1, APRIL 1984 1 EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL The Past Is The Future After three decades of the IEEE Reliability Society it is useful to reflect on the early days of our formal reliability discipline and ponder the future. The early literature was prepared largely by design engineers and university teachers, not so much by quality control engineers. The result ofthat twist of fate was the typical saying that the reliability of a product reached a peak just when the design documentation was released by the design engineers, and thus the reliability could only be degraded by those who tried to make the product. That most abominable phrase, inherent reliability, was invented and is still with us today, unfortunately. Inherent reliability really is the reliability number ob- tained by charging to it only those failures that an able, agressive, intelligent, and imaginative designer can't blame on someone else. It took a decade or two to acknowledge that reliability needed to grow, both during and after the design. The numbers game is all that many people know or remember of the reliability discipline and effort. It has been with us, probably since the first re- quest for quantitative reliability was undertaken. It is still with us, unfor- tunately. Most reliability engineers and theorists misunderstood the nature of a reliability prediction, and many of them still do. People deprecated prediction because it was well known that a skillful reliability engineer could get any number that was wanted. What they did not, and still do not, understand, is that getting the desired numbers is the anticipated result of the exercise and is trivial in itself. The important thing is the assumptions that had to be made in order to get those numbers. Those assumptions determine the cost and difficul- ty of achieving those numbers — those assumptions are the real meat of a prediction. The main thrust of much early education for reliability engineers and managers was probability theory. That thrust is still with us and few courses or books on reliability can be successful without a similar emphasis. Books labeled as reliability engineering are still largely collections of simple-minded math models — analyzed in all their glory. Most reliability problems for systems are associated with engineering and management — people problems. It was that way in the beginning and will be that way for a long time. Good reliability will always be hard to get, and for the same reasons that it is now, and always has been hard to get: • No one wants to provide the up-front resources (people, calendar time, money) to get it, or to pay for them afterwards. Few even believe that they are necessary. • No one can see it in an off-the-shelf product when they buy it. They must take some untrustworthy person's word for it. • Few people are willing to invest the personal blood, sweat, and tears to get it and keep it. — R.A.E. 0018-9529/84/0400-0001 $01.00 © 1984 IEEE

Upload: dangkhanh

Post on 01-Mar-2017

214 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The past is the future

IEEE TRANSACTIONS ON RELIABILITY, VOL. R-33, NO. 1, APRIL 1984 1

EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL EDITORIAL

The Past Is The Future

After three decades of the IEEE Reliability Society it is useful to reflect on the early days of our formal reliability discipline and ponder the future. The early literature was prepared largely by design engineers and university teachers, not so much by quality control engineers. The result ofthat twist of fate was the typical saying that the reliability of a product reached a peak just when the design documentation was released by the design engineers, and thus the reliability could only be degraded by those who tried to make the product. That most abominable phrase, inherent reliability, was invented and is still with us today, unfortunately. Inherent reliability really is the reliability number ob­tained by charging to it only those failures that an able, agressive, intelligent, and imaginative designer can't blame on someone else. It took a decade or two to acknowledge that reliability needed to grow, both during and after the design.

The numbers game is all that many people know or remember of the reliability discipline and effort. It has been with us, probably since the first re­quest for quantitative reliability was undertaken. It is still with us, unfor­tunately. Most reliability engineers and theorists misunderstood the nature of a reliability prediction, and many of them still do. People deprecated prediction because it was well known that a skillful reliability engineer could get any number that was wanted. What they did not, and still do not, understand, is that getting the desired numbers is the anticipated result of the exercise and is trivial in itself. The important thing is the assumptions that had to be made in order to get those numbers. Those assumptions determine the cost and difficul­ty of achieving those numbers — those assumptions are the real meat of a prediction.

The main thrust of much early education for reliability engineers and managers was probability theory. That thrust is still with us and few courses or books on reliability can be successful without a similar emphasis. Books labeled as reliability engineering are still largely collections of simple-minded math models — analyzed in all their glory. Most reliability problems for systems are associated with engineering and management — people problems. It was that way in the beginning and will be that way for a long time.

Good reliability will always be hard to get, and for the same reasons that it is now, and always has been hard to get:

• No one wants to provide the up-front resources (people, calendar time, money) to get it, or to pay for them afterwards. Few even believe that they are necessary.

• No one can see it in an off-the-shelf product when they buy it. They must take some untrustworthy person's word for it.

• Few people are willing to invest the personal blood, sweat, and tears to get it and keep it.

— R.A.E.

0018-9529/84/0400-0001 $01.00 © 1984 IEEE