Y2K Testing and AWIPS: Lessons Learned

AWIPS is the final element in the $4.5B NWS Modernization Program. Y2K testing was a critical step in the approval of the build 4.2 software. This paper describes how PRC and the Government tested the AWIPS system

There were several objectives to Y2K testing. The first was to test Government Furnished Software (GFS) using special data sets developed to test for Y2K compliance. The second objective was to verify hardware and Commercial off-the-shelf (COTS) software compliance using date changing on the platforms in correlation with Y2K modified data. In addition to system integration testing, PRC participated in National Weather Service end-to-end testing along with other Government agencies.

Several types of errors were discovered. The first type of error was what is commonly associated with Y2K - problems in correctly recognizing a date because the year is truncated. The second type of error was a leap year error with programs that didn't correctly recognize the year 2000 as a leap year. The third type of problem was also a leap year problem, but would have arisen in any leap year, not just the year 2000.

There were essentially no Y2K problems with the hardware and COTS software. Some data decoders had problems because they could recognize the century in the date-time field, and there were data storage and usage problems encountered with various applications. However, most of the Y2K errors were centered on the two types of leap year problems.

All AWIPS software and hardware were tested for Y2K compliance. Several mild surprises and some critical issues emerged during the course of testing, but all significant problems were overcome and resolved by the end of testing.