If anyone out there still hasn’t started work on their Year 2000 projects, alarm bells must surely start ringing in even the most laggardly CIO head following a warning from the Meta Group that the next six months marks the ‘point of no return’ for the IT industry.
By Gary Flood
The challenge is two-fold, argues the Stamford, Connecticut based consultancy: there is grave danger of a shortage of programmers, a gap no off-shore programming option can hope to solve, plus the actual time left to complete the conversion and adequately test the changes is now too short – four years was the optimum. Some of the figures bandied about when discussing the Year 2000 problem, especially the $600 billion price tag Gartner has hung on the bug, rightly draw sceptical fire. Even so, denying there is *some* potential problem means every responsible IS organisation must clean out that data garage, if only to make sure. Meta’s cut on the size of the problem is the following: take 300 billion lines of Cobol code and add some other n-billion lines of Assembler, PL1, what-have-you. That number, incidentally, is US only; add another 120 billion (40% of the US total) for the rest of the planet. On the other side of the equation, put the 600,000 US programmers who know how to spell Cobol, and have them each convert 150,000 LOC (lines of code) each per year. Add the 50% testing time needed for such necessarily complex application development, and you end up with four years. Problem. So add in third-party conversions. But Meta reckons these service providers – and that business, as an interesting sidebar, is now up to a $300 million market for last year alone – can only hope to convert five to ten billion LOC by 1998, a year often taken as a red line for many of the financial applications most at risk (given the need to calculate, say, a five year loan interest rate beyond the century boundary). That will make all these guys’ shareholders very happy with the fees earned, but that will still only be 4% of the whole problem solved. (In a note that may chill some people’s blood, Meta hears many Year 2000 fixers are already in a position where they feel able to turn down a projects that promises less than $1m-$2m!) And as for all those clever, and cheap, hands supposedly waiting to fix the problem in India – they offer a mere 5% of the US’ native programming capacity. In all cases, statistics indicate a shortage of personnel for conversion work, says Meta. There is a danger only 50-75% of US year 2000 projects will actually get done. The consequences are potentially quite dramatic. Apart from grossly inflating salaries of qualified professionals, perhaps rising by 10-15% year on year, other application development projects will have to be de-emphasised – Meta says several large shops have told it they have mandated all non-Year 2000 work must be on hold no later than the first quarter of next year so that all hands can be put on deck! If that spells bad news for in-house development, third party suppliers can be no happier – for there will be a knock-on slowdown if not actual cessation of purchasing of major packages, yes, even SAP, because of the complex migration from legacy issues involved. And so users will try and brute-force their way out of the problem by adding fixes (like data interpretation software, aka ‘windowing’ patches) at the OS (not application) level, increasing the size of their S/390 MIPS farms. The latter has interesting consequences for DASD prices. Meta advises customers considering outsourcing is to keep hold of as much of the storage capacity they can, rather than lease it to prospective third parties only for them to sell it back at a premium at crunch-time. All in all, it seems no part of the IT industry will escape the breath of the wind of date change (or will it be a tornado?) building up these next three years. Maybe the Chinese knew something we didn’t when they coined the curse, May you live in interesting times?