DARPA Year 2000 Plan of Action (DRAFT of 11/22/96)
Abstract
The Year 2000 Problem (Y2K) is so severe because it is both pervasive and the causes are idiosyncratic. Computer systems rely on operating systems and a variety of support packages, and they may interact with numerous other computer systems. Because a system is only as strong as its weakest link, a Y2K problem in any of the auxiliary systems could cause as much trouble as a direct Y2K problem in the primary system. Dates are stored and calculated in many different formats. (Over 120 date standards exist, and countless ad hoc ones have been devised according to the whims of generations of mostly undisciplined programmers.) Because of the latter issue, the Y2K problem cannot be tackled completely through automated means; however, because of the pervasiveness of the problem, the cost will be out of hand unless technology is brought to bear on the problem. Technology can reduce the cost of addressing the Y2K problem by automating tasks related to: (1) Identifying code that will work incorrectly post year 2000. (2) Date prospecting: Tracing the processing of date related information to determine if problems exist and to identify code or data that needs to be changed. (3) Transforming the necessary code and data. (4) Testing the changes; both in individual programs and across groups of programs (e.g., all programs that access a data file in which dates have been converted from 2 digits to 4 digits). (5) Managing the process; especially how to synchronize the changes in separate programs/systems.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1997
- Accession Number
- ADA332293
Entities
Organizations
- Defense Advanced Research Projects Agency