To: John Mansfield who wrote (746 ) 12/21/1997 7:43:00 AM From: John Mansfield Respond to of 9818
Y2K-TESTING - Famous GM Test Guide Found on the SIM site; thanks to SIM and David Hall. You may have to log in to the site before being able to access this URL; so Go to Topic 11 after logging in.year2000.unt.edu David Hall of SIM: 'Here is a Word Document containing the current version of the General Motors Year 2000 Test Procedures. It is one of the better test procedures documents that I have read. ' JM: It presents in a clear way the different levels of testing (component .. cluster testing). Also, comprehensive lists of sifferent types of tests are presented. '2.4. Combined Component Testing ... The RTC is tested for rollover and reboot operation.' : No mentioning ot the 'Time dilation effect' discussed on C.S.Y2K; possibly in a newer edition of this document that wil be done. This is a very good document if you want to have a real idea about embedded software testing and remediation, IMHO. John ------------------year2000.unt.edu '1. Executive Summary The objective of the Year 2000 Test Procedures document is to: Introduce the process of testing software Publish a set of procedures and test cases for Year 2000 testing Formalize the vocabulary of the testing teams Formalize the process for reporting test results Publish testing references and test tool recommendations The manufacturing & facilities core Year 2000 test team has developed this base document from several sources and pilot experiences. The test cases included are based on industry wide information and actual Year 2000 problems encountered during the pilot process. These test procedures can be used to test commercial off the shelf components and systems assembled from those components. ...' 'Test Planning This section of the document provides guidelines for writing specific test plans. Example test plans for manufacturing plants and manufacturing systems are included in the appendix. The resources required to conduct the tests are also described. Testing shall be conducted on large systems using the divide and conquer method. Unit testing addresses the smallest division for which tests are reasonable, the number of test cases is smaller and the likelihood of finding all the errors is increased. Once a unit has been tested, integration testing begins by assembling larger systems from tested units. At this level the number of test cases may increase and the likelihood of finding all the Year 2000 errors decreases. Balance and judgment are required of the test designer to select the system boundaries and appropriate test cases. Manufacturing Systems which are unique may share many common components and similar architectures. Two systems using identical hardware in a similar architecture may have different applications and use date related functions in completely different ways requiring separate testing. Testing Levels: Component testing, testing a single component controller or application in isolation Combined Component testing, testing an assembly of standard components and custom application programs or configurations System testing, testing a collection of Combined Components and Components which are assembled into a complete manufacturing or facilities control system. An example is a welding cell with PLC's, Robots, HMI's, PC's, and weld controllers. Cluster testing, testing a collection of Systems and Common Systems and / or Business Systems that constitute a complete manufacturing process. An example is a test of the paint shop systems taken as a whole including the interfaces to any common systems. Cross Cluster testing, the highest level of integration testing where the test cases are organized around specific functions, or across functional boundaries. An example is the entire process of Automobile Order through Automobile delivery. ...'