I was perusing the second-hand bookstall at our local church fete the other day. Among the dog-eared children's books and long-discarded recipe collections, I found a forlorn-looking tome entitled Year 2000 Solutions for Dummies. What an irony that last year's life-or-death challenge for the IT industry has become this year's fodder for the belltower fund!
Y2K for "dummies"? I don't think so. It may be possible to guide the cerebrally challenged through the basic functions of Word or Excel, but Y2K solutions demanded the combined skills and dedication of a workforce facing a problem of unprecedented and incalculable scale. If the year 2000 problem passed by with relatively few disastrous consequences, it's a tribute to those who brought a whole range of technical and managerial skills to the table.
I don't want to dwell on Y2K; we seemed to talk about very little else last year. But among the things it taught us (or at least reminded us) was the need for technical continuity, formal maintenance and change management procedures, tight application integration, consistent system management, and a good IT/business relationship. In some companies, these things were achieved for the first time as a result of the Y2K exercise, and have provided a sound basis for integrating their back-end transactional and database applications (CICS/Cobol or RPG or whatever) with the new demands of e-commerce.
But there is a great threat to this kind of technical continuity: the skills shortage. The industry is facing a dearth of specialists with the right mix of technical disciplines, and this problem shows no sign of abating. In Xephon's latest IS Plans survey, which analyses the priorities and challenges facing IS management, 78 per cent of respondents rated "shortage of staff" as one of their three most serious problems, a similar figure to last year, but considerably higher than the 67 per cent figure in our 1998 research. Some sources quote an IT skills shortfall of 300,000 people in the USA alone, and Europe is no better off.
Quite apart from the obvious problems of spiralling salaries and "poaching" of key personnel, the skills issue highlights the imbalance created by fashionable technologies. Today's market naturally puts its highest premium on e-skills - on XML and Java and similar tools - and anyone with support or development experience on these platforms is clearly very marketable. Small wonder that the IT workforce is becoming more and more "mobile", and that relatively mature areas of specialisation, such as OS/390 internals, are losing staff to the more sought-after positions. To compound the problem, as we found in some earlier research, the skills problem is forcing companies to outsource many of their "core" e-business applications to ASPs and other service providers, while keeping "chore" systems in-house. A necessary expedient, perhaps, but hardly the most favourable conditions for nurturing a stable, knowledgeable in-house IT department.
For the average large enterprise, which typically involves a broad mix of OS/390, OS/400, Unix and NT system support; Cobol, PL/I, C++ and 4GL development and maintenance; DB2, Oracle, SQLServer and numerous other databases; SAP, PeopleSoft and other business packages, this ebb and flow of technical skills is reaching nightmare proportions. However, with such an unpredictable workforce and little opportunity to fall back on the expertise of "old hands", we can at least make sure that the formal structures outlined earlier are in place. With rigorous project management, meticulous documentation, integration and adherence to formal standards wherever possible, and attention to the top-down management disciplines that have always characterised centralised enterprise computing, we can perhaps reduce the impact of this rapid turn-over of staff.
Year 2000 Solutions for Dummies may be resting peacefully in a country churchyard. But, dummies or not, we mustn't forget the broader issues raised by the Y2K problem.
Mark Lillycrop is director of research for large system researchers and consultants Xephon, at xephon.com