1. Expand all database date fields by two bytes to include the century (i.e. YYYYMMDD) and

2. modify and re-compile every single program and piece of code that references the date field or includes the database definition (schema).

This approach has many drawbacks:

1. Increased DASD requirements.
   Even though DASD is relatively inexpensive, many applications have literally millions of dates. A date field increased by two bytes factored across millions of occurrences could have a substantial impact on the database size. Many times databases are sized to fit completely on a single device or exist with many other files on a completely full device. Even a minor increase in the size of a database can mean repositioning files, splitting up existing files, etc.

2. Increased buffer utilization and i/o.
   When record sizes increase, less non-date data exists on a given database page. Now the CPU works harder reading and writing additional database pages and buffers become more heavily utilized. All of this degrades performance.

3. Substantial database administration effort to increase the record field formats.
   This approach requires a re-structure of existing database records and data. Besides the obvious drawbacks of database administration preparation time, the sheer volume of data that must be manipulated may make this seemingly simple task very difficult.

4. Every program and dialog that uses the schema definition must be re- compiled.
   Since the record formats are changing, the programs and dialogs must be re-compiled to account for the increased field size. This is true even of programs and dialogs that do not directly reference the date fields (or even the records that contain the dates). Although no company or organization would like to admit it, re- compiling programs and dialogs almost always turns up missing code and inconsistent source libraries (i.e. test and production source libraries are rarely 100% compatible).

5. Programs and dialogs that reference the date fields must be examined and modified to account for the YYYY value in the following circumstances:
   • date compares
   • storing and/or modifying records
   • retrieving CALC records (with a CALC date field) or VIA or indexed records with a sort value (sorted on date field)

There are several disturbing trends that have developed lately regarding companies attempts to resolve the Y2K issue;

1. There is a VERY limited amount of experienced CA-IDMS talent in most companies. There is a huge demand for anyone that can even spell IDMS. Salaries on the open market place are getting higher every day, and will continue to spiral out of control until the year 2000.

2. Many companies want to just "wash their hands" of the Y2K conversion effort. They are willing to hire big name consulting companies to come in and just "fix" the problem. Many of these Y2K solution companies also have a very limited experienced IDMS staff. Some organizations are hiring new college grads, giving them a couple of weeks of COBOL training and then billing them out at excessive rates.

3. Unfortunately many Y2K solution companies are often just looking at any way to maximize their billable hours. Usually the Y2K effort is measured in lines of code, and sometimes conversion efforts bids approach (or exceed) $1.00 per line of code. This is the effort that most of the companies feel is justified to re-structure the database.

4. Some good questions to be asking are;
   • Can these guys get the job done?
   • Have they done it before?
   • If so where?
   • Can we get references?
   Assuming you can get references - CALL THE REFERENCES!! Then get scared. Promises are being made that can not be kept!!

5. One of the most common Y2K solutions appears to outsource the Y2K effort overseas to a Y2K factory.
   • Who are these guys?
   • Do they have IDMS?
   • Do they know IDMS?
   • What happens if in 6 - 8 months the code isn't starting to return?
   • What happens if the code returns, and dialogs will not generate or programs will not compile?
   • Who do you call on Jan 1, 2000 if something doesn't work?