Narrow Results

This chapter complements the chapters on technical reviews and software reliability engineering in Vol. 1 of the handbook. It is primarily concerned with the verification of code by means of testing, but an example of an informal proof of a program is also given. A practitioner's view of testing is taken throughout, including an overview of how testing is done at Microsoft.

The following sections are included:

• INTRODUCTION

• TECHNICAL ARGUMENTS FOR TESTING

• NUCLEAR WEAPON SYSTEM SAFETY
  • Enhanced Nuclear Detonation Safety
  • Insensitive High Explosive
  • Fire-Resistant Pits
  • Missile Propellant

• RELIABILITY OF THE NUCLEAR WEAPON STOCKPILE
  • A Nondevice Component
  • The Device
  • Reliability Assessment of the Stockpile

• NUCLEAR WEAPON EFFECTS

• DESIGN COMPETENCE

• VERIFICATION

• POLITICAL ISSUES IN THE COMPREHENSIVE TEST BAN TREATY DEBATE

Testing is one of the most important phases in the software development process, often requiring considerable effort and resources. We propose a novel approach for generating test cases, based on requirements specification. We make use of scenarios used in the requirements specification phase, taking into consideration the various relationships that can exist between them. These relationships are represented as dependency diagrams and they play an important role both in requirements specifications and in test case generation. Using our approach we can ensure that a larger proportion of requirements are actually tested.

There is an odd contradiction about much of the empirical (experimental) literature: The data is analysed using statistical tools which presuppose that there is some noise or randomness in the data, but the source and possible nature of the noise are rarely explicitly discussed. This paper argues that the noise should be brought out into the open, and its nature and implications openly discussed. Whether the statistical analysis involves testing or estimation, the analysis inevitably is built upon some assumed stochastic structure to the noise. Different assumptions justify different analyses, which means that the appropriate type of analysis depends crucially on the stochastic nature of the noise. This paper explores such issues and argues that ignoring the noise can be dangerous.

This paper analyzes the source and the characterization method of white noise in FOCT, researches the statistical properties of white noise in FOCT while the equipment is in a zero-input current environment. Data in time-domain and frequency-domain is analyzed and compared under different primary input current. The result indicates that white noise in FOCT accords with a normal distribution with mean zero, and the noise magnitude is independent of the primary current value. On that basis, we research how the white noise affects the test result of FOCT and provide an effective method to reduce the impact.

Emerging pathogens have no known therapies or vaccines and therefore can only be controlled via traditional methods of contact tracing, quarantine and isolation that require rapid and widespread testing. The most recent outbreak from an emerging pathogen is due to the highly transmissible SARS-CoV-2 virus causing COVID-19 disease, which is associated with no symptoms or mild symptoms in 80–90% of the infected individuals, while in the remainder of the patients it exhibits severe illness that can be lethal or persist for several weeks to months after infection. The first tests to diagnose infection by SARS-CoV-2 were developed soon after the genome of the virus became known, and use probes to measure viral RNA by reverse transcriptase-polymerase chain reaction (RT-PCR). These tests are highly sensitive and specific but can require several days to return results, which makes contact tracing and more generally efforts to control the spread of the infection very difficult. Furthermore, the sensitivity threshold is orders of magnitude below the viral load necessary for transmission; therefore, individuals recovering from the infection may still be have a positive test and be required to isolate unnecessarily while they are no longer infectious. Antigen tests were subsequently developed that use antibodies mostly targeted to the nucleocapsid protein of the virus. These tests are about 100 times less sensitive than RT-PCR, yes they detect viral loads that are about 1/10 that needed for transmission. Furthermore, such tests are potentially much cheaper than RT-PCR and yield results in 15 min or less. Antibody, also known as serological testing, is available and can provide useful information to understand the extent to which a population has been exposed to the virus; however, it is not a good indicator of current infection and not useful for infection control. Viral transmission models that incorporate testing and contact tracing show that infection control is much more readily achieved by increasing testing frequency than by using higher sensitivity testing. For example, compared to no testing at all, testing once every other week has a marginal benefit, while testing weekly can decrease the number of infections to 20–40%, and testing twice weekly or more can bring about a 95%þ reduction in infections. These lessons learned from dealing from the COVID-19 pandemic should guide future planning against potential emerging viruses.