Y2K Countdown: Fear of Frying
You’re done with your Year 2000 efforts. You have tested and are about ready to move from the development environment to your production system. Is there anything else that can be done to reduce the risks and improve the correctness of the project? Has anything been missed?
The fearless say that the Year 2000 project is completely done. There are no outstanding mistakes. Nothing is going to go wrong. No need to run away and hide, this will be “just another day at the office.” No big deal.
Are you sure “all” is 100 percent complete, no hidden mistakes? Be objective and professional in your reply. A simple YES or NO – just as attorneys demand in court. Do you have some qualifying remarks you’d like to mention? More you would like to cover? Are you apprehensive?
Did you answer YES or are you feeling a little fidgety. OK then, no one minds an independent audit tomorrow, right?
Audits
The very thought of audits is usually distasteful to almost everyone. It sounds like a new, unexpected dimension has been added to your immediate matrix, blowing away your very tight schedule. But they are very helpful and save time and money too.
Audits and independent verification and validation (IV&V) are consistently finding high error rates in allegedly Y2K compliant code. Year 2000-introduced error rates of close to 10 defects per 10,000 Lines of Code are commonplace. Add to that the fact that IV&V can find these defects at a much lower cost per defect found and fixed while testing than later in the game. IV&V frequently finds more errors than testing.
The idea is to reduce the anxieties that go through your mind as you think of the entry of Year 2000 transactions into the company business and your data processing system. Electing to bypass an audit opens the possibility for more highly active triage teams. As time draws closer, more and more dates into and beyond 2000 must enter into your applications. Small incidents become larger, more critical and need immediate repair. No one wants this added work. It fiercely complicates your tasks and is radically difficult to undo.
New date transactions will enter from controlled and uncontrolled sources. We are accustomed to demanding scheduled and last minute changes to meet one or more new business process changes. This Year 2000 area is a new experience – we have never been here before – at least not like this. Why not minimize your risks?
Day to Day Example
As an airline passenger, you have no control beyond booking a flight from Airport “A” to Airport “B” at determined times. Think of your Y2K journey like this. Once you have passed through airport traffic, arrival and security checks you check in at the counter.
Onboard, the cabin crew greetings are followed by the “fasten your seat belts” message (there they go again – more precautionary messages). Taxiing down the runway, your captain reassures you about the flight (a calming message). Then the cabin crew presents the safety rules on emergency exits, emergency oxygen, life jackets and emergency lighting (the ultimate graphical caution presentation). Most people read, chat or sleep through this message.
Moments after we race down the runway and into the Wild Blue Yonder, our fate is in the hands of a coordinated mix of air traffic control centers and the pilot’s navigational and flying skills. You are the Y2K pilot maintaining all actions for your company’s survival.
Similarly, the audit or IV&V is there to do the same thing – help you through the journey. Some of the catches I have seen relate to missing or duplicate source members, missing source, or file objects. Even if only .006 percent of the programs are found to have potential errors, the system-wide impact can be much greater. The error could introduce erroneous data into key files or business decisions/processes. If there is a logic error in a called standard routine, the spread of the error throughout your system is possibly very severe.
Limiting the discussion to your systems only, an audit might look at source code from several views. Another concern is the database and the date field validity today, during your test phase, and after your migration plans to Year 2000 compliant systems. Inspections might also extend to the internal and external interfaces to ensure that they properly provide for receiving and sending your data.
Audits and IV&V might offer a more prioritized list of key elements that feed contingency planning. In the Year 2000 world, audits generally are performed before live implementation, with time to react and change if necessary. Contingency plans, while developed before implementation, are there to cover the “just in case it happens” anyway scenarios in an organized manner.
You can have audits in many forms. Be sure you get the level of protection that is correct for your situation.
Conclusions:
Both audits and contingency plans are worthwhile contributions to your success, regardless of how hard we try we make mistakes. There are standard error rates per “n” line of code corrected. Because of the large size of Year 2000 projects, there are few, if any, that will get through without mistakes.
The prime mission is to discover, minimize and eliminate these areas as soon as possible, then be ready for the contingency areas. No one in the flight crew expects to crash or get lost in the skies, but they are prepared for optimum survival if need be.
Good Luck!
Glenn Ericson is president and founder of Phoenix Consulting LLC, in East Elmhurst, N.Y. He specializes in Year 2000 and risk management issues. Glenn-Ericson@Att.Net