Ecumenical Field Reliability

In the religion of reliability analytics

Reliability is an organized collection of beliefs, cultural systems, and world views that relate reliability to human behavior, the supernatural, and to spirituality.

0) 501(c)(3) form 1828: "Tax Guide for Churches & Religious Organizations"

To qualify for tax-exempt status, such an organization must meet the following requirements (covered in greater detail throughout this publication):

■ the organization must be organized and operated exclusively for religious, educational, scientific, or other charitable purposes,

■ net earnings may not inure* to the benefit of any private individual or shareholder,

■ no substantial part of its activity may be attempting to influence legislation,

■ the organization may not intervene in political campaigns, and

■ the organization’s purposes and activities may not be illegal or violate fundamental public policy.

*Inure is to cause (someone) to be less affected by something unpleasant : to cause (someone) to be less sensitive to something unpleasant

1) Create a God. One with a catchy name is best. Should be simple and out of the ordinary, but not too far out that people can't remember it.

In our example we will create "The Great God Ulpianus," who made actuarial forecasts that recognize remaining life depended on age (of Roman Legionnaires).

2) Make it in charge of something people already focus on, but don't have a target for that focus.

Consumer bills of rights entitle people to product information, which includes real reliability, not predictions, not test results, and not claims based on anecdote

3) Make it something that people will be reminded of frequently.

Every time something errs, fails, gets sick, or dies: hardware, software, or human.

4) Make it easy for them to "buy into" the worship of your New God.

Whenever you buy a spare or replacement, you are honoring Ulpianus. Whatever you pay for an extended warranty, you are tithing to Ulpianus.

5) Make it ambiguous. Let both sides of an argument claim it as their own.

In this case, Ulpianus was bent on destroying belief in constant failure rate, Weibull reliability, or physics of failure models instead of empirical or nonparametric reliability, survivor, and failure rate functions based on field data.

Why define it when you can let people fight it out themselves. People are funny. They'll fight over anything. Even something you just made up. Enjoy the show and try not to think too much about the Karmic issues you are creating for yourself.

6) Establish some standards by which the God should be referred to, creating an intrinsic reverence right from the start.

In this case, The Great God Ulpianus must always be referred to as "The Great God" Ulpianus. Once people see this a few hundred times, they'll start believing it without even realizing it.

Make sure that the Full Title and Name are Always Capitalized. This is because that everybody knows that something that has Capital Letters Is Much More Important than something that isn't.

This is called "Marketing" and all the Most Holy of The Great God Ulpianus's High Priests study the Dark Arts of Marketing and Acturial Methods, and practice it many times a day in their Most Holy Rites.

7) Make cool symbols. They should be things that people already know and see everywhere. And they should be easy to draw and say.

In our case, The Great God Ulpianus's symbols will be Actuarial Rates. Insurance companies worship them. The US Air Material Command worships them for engines and major service modules. The FDA worships them for clinical trials while killing controls.

8) You need opposing forces. Not necessarily an arch-enemy, but opposing perspectives so that people can pick sides and fight over things.

Ralph Evans said, “Field reliability data is garbage and deserves…”

Reverend Bayes said, “It depends on the prior distribution or beliefs…”

Procurement specifications specify MTBF requirements, not reliability

Ronald Fisher might have said, “Who cares if controls die? They’re only plants.”

Advocates of physics of failure and Weibull reliability claim that constant failure rates are unrealistic, and that some test data fits a double exponential. There are even physics and structural failure models that theoretically should have Weibull reliability.

Nuclear power risk analysts worship HCLPF, High-Confidence, Low-Probability-of-Failure, for managing risk.

Relative risk or proportional hazards seem the basis for clinical trials, even though risks may not be proportional. 

9) You need to confuse everybody. This will make sure that nobody can be really certain WHAT they believe, because it is all so nonsensical to begin with. And when you don't spell it out exactly (or even if you do) you know how those funny humans will all magically just get along, right!

Not a problem. We’re using statistics and mathematics.

10) The Big Reward. You know everything you always wished you had in this life? After you die, you'll get it!

Relief from uncertainty.