Understanding how the typical distribution of failure for a given product must be factored with time is also important when looking at failure patterns(Fig. 3). Early life failures are often associated with fabrication issues, quality-control issues. or initial"shakedown"stresses. while age-related failure rates would increase with time. This is discussed in more detail in the article "Reliability-Centered Maintenance"in this volume failure Wearout failure period Intrinsic failure period Time Fig3 Typical time distribution of failures("bathtub curve Once the concept of a managed life is prudently adopted over a simple failure prevention concept, design and fabrication costs can be reduced and maintenance and other life-prolonging activities can be optimized Diligence in Use of Terminology. Communicating technical information is of paramount importance in all engineering areas, including failure analysis. The choice of technical descriptors, nomenclature, and even what might be considered technical jargon is critical to conveying technical ideas to other engineers, managers, plant personnel, shop personnel maintenance personnel, attorneys, a jury, and so forth. It is instructive in this introductory article to emphasize that a descriptor can mean something very specific to a technical person and mean something very different to a business manager or an attorne For example, the term"flaw is synonymous with defect"in general usage. However, to a fracture mechanics specialist, flaw is a discontinuity such as a crack. Under some circumstances, when the crack is smaller than the critical size (i.e subcritical), the crack is benign and therefore may not be considered a defect. To the quality-control engineer, flaws are characteristics that are managed continuously on the production line, as every engineered product has flaws, or deviations from perfection"(Ref 14). On the manufacturing floor, these flaws are measured, compared with the preestablished limits of acceptability, and dispositioned as acceptable or rejectable. A rejectable characteristic is defined as a defect(Ref 14). To the Six Sigma practioner, a defect is considered anything that inhibits a process or, in a broad sense, any condition that fails to meet a customer expectation(Ref 9). To the attorney, a defect refers to many different types of deficiencies, including improper design, inadequate instructions for use, insufficient warnings, and even inappropriate advertising or marketing(Ref 15) Similar nuances may occur in the basic definitions and interpretations of technical terms used in materials failure analysis Terms such as ductile and brittle, crack and fracture, and stable and unstable crack growth are pervasive in failure analysis. Even these seemingly basic terms are subject to misuse and misinterpretations, as suggested in Ref 16--for example"brittle cleavage, which is a pleonasm that does not explain anything. Another example noted in Ref 16 is the term " overload fracture, which may be misinterpreted by nonanalysts as a failure caused by a load higher anticipated by the materials or mechanical engineers. This limited interpretation of overload failure is incomplete described in the article"Overload Failures"in this volume Judgmental terminology should be used with prudence when communicating analytical protocols, procedures, findings, and conclusions. Communications during the preliminary stages of an investigation should be factual rather than judgmental. It is important to recognize that some of the terminology used in a failure analysis can be judgmental, and onsideration must be given to the implications associated with the use of such terminology. For example, when examining both a failed and an unfailed component returned from service, references to the unfailed sample as"good and the failed sample as"bad"should be avoided. This is because the investigation may reveal both samples to contain the same defect, and therefore both could be considered"bad. Similarly, neither may be bad" if the analysis actually indicates the failed component met all requirements but was subjected to abuse in service. On completion of the failure analysis, judgmental terminology is often appropriate to use if the evidence supports it, such as in the example of asting defect that has been confirmed in the example bolt failure analysis While discussions of the semantics of terminology may seem pedantic, communicating the intended information gleaned from a failure analysis relies heavily on precision in the use of language References cited in this section 4. W.E. Deming, Out of the Crisis, MIT Center for Advanced Engineering Study, 1986Understanding how the typical distribution of failure for a given product must be factored with time is also important when looking at failure patterns (Fig. 3). Early life failures are often associated with fabrication issues, quality-control issues, or initial “shakedown” stresses, while age-related failure rates would increase with time. This is discussed in more detail in the article “Reliability-Centered Maintenance” in this Volume. Fig. 3 Typical time distribution of failures (“bathtub curve”) Once the concept of a managed life is prudently adopted over a simple failure prevention concept, design and fabrication costs can be reduced and maintenance and other life-prolonging activities can be optimized. Diligence in Use of Terminology. Communicating technical information is of paramount importance in all engineering areas, including failure analysis. The choice of technical descriptors, nomenclature, and even what might be considered technical jargon is critical to conveying technical ideas to other engineers, managers, plant personnel, shop personnel, maintenance personnel, attorneys, a jury, and so forth. It is instructive in this introductory article to emphasize that a descriptor can mean something very specific to a technical person and mean something very different to a business manager or an attorney. For example, the term “flaw” is synonymous with “defect” in general usage. However, to a fracture mechanics specialist, a flaw is a discontinuity such as a crack. Under some circumstances, when the crack is smaller than the critical size (i.e., subcritical), the crack is benign and therefore may not be considered a defect. To the quality-control engineer, flaws are characteristics that are managed continuously on the production line, as every engineered product has flaws, or “deviations from perfection” (Ref 14). On the manufacturing floor, these flaws are measured, compared with the preestablished limits of acceptability, and dispositioned as acceptable or rejectable. A rejectable characteristic is defined as a defect (Ref 14). To the Six Sigma practioner, a defect is considered anything that inhibits a process or, in a broad sense, any condition that fails to meet a customer expectation (Ref 9). To the attorney, a defect refers to many different types of deficiencies, including improper design, inadequate instructions for use, insufficient warnings, and even inappropriate advertising or marketing (Ref 15). Similar nuances may occur in the basic definitions and interpretations of technical terms used in materials failure analysis. Terms such as ductile and brittle, crack and fracture, and stable and unstable crack growth are pervasive in failure analysis. Even these seemingly basic terms are subject to misuse and misinterpretations, as suggested in Ref 16—for example “brittle cleavage,” which is a pleonasm that does not explain anything. Another example noted in Ref 16 is the term “overload fracture,” which may be misinterpreted by nonanalysts as a failure caused by a load higher than anticipated by the materials or mechanical engineers. This limited interpretation of overload failure is incomplete, as described in the article “Overload Failures” in this Volume. Judgmental terminology should be used with prudence when communicating analytical protocols, procedures, findings, and conclusions. Communications during the preliminary stages of an investigation should be factual rather than judgmental. It is important to recognize that some of the terminology used in a failure analysis can be judgmental, and consideration must be given to the implications associated with the use of such terminology. For example, when examining both a failed and an unfailed component returned from service, references to the unfailed sample as “good” and the failed sample as “bad” should be avoided. This is because the investigation may reveal both samples to contain the same defect, and therefore both could be considered “bad.” Similarly, neither may be “bad” if the analysis actually indicates the failed component met all requirements but was subjected to abuse in service. On completion of the failure analysis, judgmental terminology is often appropriate to use if the evidence supports it, such as in the example of a casting defect that has been confirmed in the example bolt failure analysis. While discussions of the semantics of terminology may seem pedantic, communicating the intended information gleaned from a failure analysis relies heavily on precision in the use of language. References cited in this section 4. W.E. Deming, Out of the Crisis, MIT Center for Advanced Engineering Study, 1986