In 1998, Jet Propulsion Laboratory created an institution-wide project to ensure that no system depending upon information technology (IT) would suffer an interruption due to date-related processing through 2015. Such an undertaking was unparalleled in the number of affection organizations and assets, the shortness of schedule, and the degree of external interest. In order to meet this challenge, the JPL Y2K Project adopted the following approach: * Provided universal and unambiguous statements of requirements, processes and procedures. * Compiled an unprecedented database about JPL IT assets and plans. * Created real-time, web-based, context-sensitive metrics. * Reviewed progress on the above items, on a weekly basis, with a Project team having representatives from every major JPL organization. These steps ultimately culminated in an uneventful and thus successful transition to 2000.

Bob Miller received his Ph.D. from UCLA in Mathematics in 1977, specializing in algebraic coding theory. He taught mathematics for over 10 years at UCLA and College of the Canyons, and has been working at JPL since 1972 in the following positions: * Communication systems research engineer for the Deep Space Network. * Supervisor of a group conducting algorithm development and implementation of the Corps Battle System combat simulation installed at Army bases around the world. * Deputy Manager of a Section that fielded software-intensive systems. * Year 2000 Project Engineering Manager.

Dr. Miller said that groups at JPL had their own fiefdoms so it was necessary to insure that everyone worked together in obtaining full year 2000 compliance, including operational systems, desktop computers, facilities, business systems, laboratories and others.  In addition such systems as elevators, generators, controllers and many other things needed to be Y2K compliant.  At the beginning there were few believers in the Y2K problem.  Engineers frequently responded that it affected only business systems and that their code was designed to be compliant.  Managers complained that they had no time to address the problem. 

Dr. Miller began by reading what the Government Accounting Office (GAO) had to say about the problem and he found their instructions quite good.  They included a series of steps that should be followed to ensure compliance.  Dr. Miller decided that finding a solution was a management problem rather than a technical one.  He found that it was critical to get executive management support, and he was able to get this by direct contact with the Director of JPL.  JPL uses a matrix organization, which means one group provides the personnel and another the funding.  It was necessary to document a formal compliance process by providing on-line real-time metrics and reports to keep score on people and organizations. 

JPL followed an approach that gave higher priorities to Y2K issues that applied to products or services affecting other organizations and operational systems.  The lowest priorities were given to anything that applied only to one organization or person.  A Y2K compliance flow chart was provided to groups to indicate the necessary steps required and instructions were provided to make it easier to accomplish these steps.  One of the first things done was to check that the latest software was in use because some corrections could be made by shifting to the newer versions. 

People had to make choices on where they were going to make fixes.  The Y2K Project Organization provided input software for on-line inventory that included certifications and waivers, and places for sign off by designated independent witnesses.  They provided code filtering services for developers, generic test work sheets, Commercial Off the Shelf  (COTS) product Y2K status, some software patches, and helped in analysis of the problem.  They provided a desktop computer sticker process to label each Y2K compliant machine.  Each development organization had to supply a description of its program including supporting COTS, certification of plans to retire software before 2000 rather than correct it, and code inspection reports.  The JPL requirement was to certify compliance out to the year 2015. 

The input forms had red sections that were required to be filled out before input would be accepted, and blue sections filled in by incremental reports.  The red sections identified responsible individuals and groups, and quantified the risk as high, medium, or low.  The level of the priority was based on what effect that process being out of service for a period of time would have on operations.  At the beginning the groups had to supply planned dates which were subject to change. When Y2K compliance was reached it would be signed off as an actual date.  There could be only one actual completion date.  Completion was tied to certification of the results with attached code and test reports.  Internal and independent auditors checked the data submitted as to whether they actually met requirements.  Auditors could demote the claimed completion if they found the data submitted were not complete or convincing.  For example, waivers were allowed to some of the requirements.  NASA set a date that came before the Cassini Earth fly by; the project team quite naturally did not want to be working on the software at that time.  They were given a waiver for a later completion date. 

There was a considerable amount of slippage of planned completion dates early in the program and many early submissions did not meet Y2K requirements.  There had to be checks made that software that was scheduled for retirement before 2000 was actually going to be removed, especially if replacement was required.  Things moved faster after the Y2K Project demonstrated they were serious and the JPL Director gave them strong support.  Dr. Miller said that the Y2K corrections required a considerable amount of time and resources, paid for by the development projects.   It would have been impossible to get these resources without high-level executive management support.  He said you have to work top down and have metrics to achieve success and success was achieved at JPL.  The Y2K effort made their top management aware of the value and necessity of their in-house software capability. 

This was a well-done and very interesting presentation by Dr.  Miller.  It was an excellent report on how the Y2K problem was made into a "non-event" at JPL.  This was the eighth meeting of the LA Chapter year and was attended by about 18 persons. 
Mike Walsh, LA ACM Secretary