Comptroller of the Currency Eugene Ludwig presented remarkably frank testimony to the Senate Banking Committee on July 30. In it, he described the number-one problem facing banks (and all other systems). This is the problem of interconnected computers. All of them must be compliant in 2000, as well as consistent with each other. The programmers must test their repairs by interfacing with all of their customers and all other banks they deal with.
What he did not say is that in order to do this, programmers must deal with a number of possible interconnections that approaches infinity as a limit.
The y2k problem cannot be solved. There is no agency with the sanctions required to police such an international coordinated conversion. It is a worldwide problem. There is no U.S. government agency that has imposed a plan of action on everyone. Also, there is no agreement as to the methodolgy of the repair: windowing, expansion, or encapsulation. The Federal Reserve has recommended a 2-digit approach (windowing), while the Office of the Comptroller of the Currency has recommended a 4-digit approach (expansion) (p. 8).
Here is THE problem: everything must be fixed for anything to be fixed in an interconnected system. Until there is a suggested theoretical answer -- there is none so far -- there is no practical solution to the y2k problem.
If the banks go down, the West goes down. The division of labor will collapse.
* * * * * * * * *
There are two aspects to solving the year 2000 problem. First, users of technology must solve this problem with respect to their internal systems. That is, they must make sure that their internal computer systems properly handle date-dependent transactions and computations in the new millennium. Second, they must make sure that the systems they use can exchange date-dependent information effectively and efficiently with other, external systems (p. 2).
It is important to keep in mind that the transition to the Year 2000 will pose the greatest challenges to banks and vendors that perform their own programming. Hence, they need to be particularly vigilant about testing and re-testing their systems. Furthermore, testing is quite a complex matter. Banks must test all their computer systems to make sure they can process dates after the year 2000, and then test them again to make sure that the method used to process dates is compatible with the methods used by the customers and correspondents with whom a bank exchanges data and funds. This requires that banks work closely with their customers and correspondents -- whether they be other financial institutions, vendors, government agencies, or foreign entities -- to coordinate a smooth transition to year 2000 compliant electronic-transfer interconnections (p. 9).