The clock continues to tick away on the Millennium Time Bomb, and since the majority of organizations out there have not completed the absolute safest level of Year 2000 compliance protection – modification of all their data files – most now have to turn to the next best thing, bridging, or modification of code, which is either a kludge or a way to grant yourself some breathing time, depending on how you see it. The necessity for some sort of band aid is only underlined by a report published Tuesday by San Diego research firm Technology Business
By Gary Flood
Reports, which has stuck its head out even further than Gartner Group on the likely cost worldwide of fixing the problem, with a prediction of $2 trillion to Gartner’s $600 billion. The consulting firm puts the figure so high due to inclusion of a range of possible imponderables, such as the cost of rewriting existing programs, the acquisition and installation of replacement systems, and productivity possibly to be lost due to system downtime and business interruption caused by systems failing due to non year 2000 compliance. The firm believes that the average cost of fixing a line of code, often, but not exclusively, Cobol on a mainframe, is set to rise from $1.30 the second half of this year to $1.75 the first half of 1998, $2.35 the second half of next year, all the way up to $3.65 the bottom half of 1998 and $4.00 in the first few months of 2000 itself. As is swiftly becoming folk wisdom, the report confirms that the main factor behind such steep rises is the seller’s market in Cobol and Y2K skills, with consultants now being able to charge $1,500 per day on average, a sum set to rise to about $2,500 in 1999 for the top notch ones. This kind of talk is labeled scare mongering by some, who dismiss it as just a way to Chicken Little clients into paying up for big ticket Year 2000 toolsets and body shopping services; and Computergram International has consistently questioned the science behind such enormous sums being plastered all over the pages of Newsweek. Yet at the same time, there is undeniably a big problem out there, and one has to remind skeptics that they may yet be ridden out of town on a rail for exposing their company to litigation for failing to turn MM/DD/YY into MM/DD/CCYY (where C = century). Thus, back to bridging, a Year 2000 conversion strategy that depends on so- called windowing, essentially a program that scans code as it is being executed, watching out for any two-digit date fields. The programmer assigns a value whereby, for example, date fields that contain the numbers 00 to 60 should be interpreted as 2000 to 2060 and all those with 61 to 99 would be interpreted as really being 1961 to 1999. This means that you get let off the hook of having to go through all your data files and change all date fields, since existing systems can continue to run until such time that they are either properly fixed, mothballed, or swapped out for some large application such as SAP AG’s R/3. (Readers will have spotted that this will not work in context of having three centuries to worry about, i.e. 19th, 20th and 21st, which is common in the insurance. You guys will have to do it the hard way, it seems.) Of the main Year 2000 software vendors – Computer Associates International Inc, Platinum Technology Inc and Viasoft Inc – it is the latter which claims is first to market with a bridge product with Bridge 2000, aimed at Global 5000 organizations with large MVS Cobol, PL/1 and Assembler code libraries. Since Bridge 2000 merely interrupts program I/O with a single CALL statement, the Phoenix, Arizona company says, it is a non-intrusive patch that will save a few people’s skins for the time being. As Wayne Snell, its Solutions Marketing Manager states: With bridging, you don’t have to stop everything you do while you fix the Year 2000, you can do it a program at a time. It’s not a silver bullet, but it will buy you some time. And if you have already got your impact analysis out of the way it could be all you need anyway. Viasoft, which ha
s 900 client organizations, believes 90% of them are looking at bridging as the way forward, and the company was able to beta test the code with 22 companies eager to get started. Bridge 2000 is immediately available for $80,000. If bridging is a way to stave off the Year 2000 nightmare, will companies like Viasoft be in turn crushed when the market presumably moves on to the next problem in 2001 or so? We’ll hear tomorrow how Viasoft is already planning for life beyond that end of the century party cum (possible) disaster.