This is the killer. Even if organizations get their code repaired, they will have to test it and debug it. This will take anywhere from 50% to 70% of the total resources, INCLUDING TIME, devoted to the y2k repair. The Securities & Exchange Commission spelled out in detail what needs to be done.
It's easy to spell this out. The hard part is doing it. Doing it in 1999 will be even harder. Where will organizations get the excess mainframe capacity? The extra personel? Only if few organizations reach the testing stage will such resources be available -- another Catch-22.
* * * * * * *
Testing will account for 45-50% of SROs' [self-regulating organizations, e.g., NASD] effort to correct the Year 2000 problem. The financial community is currently working on a definition for "street-wide-testing". Although street-wide testing is ultimately required for all applications, many preliminary tests can be conducted through a test environment that simulates a "live" scenario.
Although most SROs have made adequate progress in developing testing plans, there are certain areas where more progress is needed. For example, testing Year 2000 modified software will require a separate, dedicated test environment on which conditions prior, during, and after the change of century can be simulated, including the leap year condition in the year 2000. In addition to testing for application, systems and utility software compliance, the SROs need adequate capacity to accommodate the needed testing and affects of Year 2000 on its various systems.
Finally, regression tests will need to be performed to identify unpredicted changes to other portions of the systems, i.e., testing that systems not only handle the Year 2000 dates correctly, but that they also correctly perform all the other functions they were designed to accomplish. Regression testing needs to retest even unchanged parts or programs of the system.