Some y2k experts say that switching to a new, compliant hardware/software system is the only way out. It's too late to fix legacy systems. (Category: "Too Late"; June 19, 1997, entry.) But very few organizations are taking this advice. Well, they take half of it: it's too late to fix legacy systems. Conclusion: it's best to do nothing.
Here is the problem with the "switch now" advice: it isn't cheap. Switching may not even be possible for a firm in the time remaining. Most important -- and never said in print -- the computer hardware/software companies cannot produce new systems and install them if every y2k-threatened organization should decide to switch. It's another of many Catch-22 aspects of y2k.
This assessment appeared in a special October issue of SOFTWARE MAGAZINE.
* * * * * * * *
Replacing existing systems is no simple matter, however, and may not even eliminate Year 2000 problems. Most companies still need to inventory in-house systems, migrate legacy data, and interface with myriad heterogeneous systems. Additionally, installing packaged software is not a plug-and-play endeavor. New apps need to be integrated with existing systems, and then phased into production. . . .
Furthermore, introducing new software elements to existing environments requires various types of testing activities, including regression testing, unit testing, integration testing, and user acceptance testing. Many organizations also require a quality assurance step. So, depending on the size of the installation, the number of functions being installed, and the overall complexity of the existing environment, migrating to packaged applications can take months or even years.