The rush to create viable e-commerce systems during the dotcom frenzy resulted in the overthrow of software engineering disciplines which had taken the best part of 40 years to establish. Now, although the dotcom boom is essentially over, it is unlikely that the industry will ever return to the traditional methods of systems creation. This tacit revolution represents the most dramatic change in IT practice since the development of the first business computer. And it should, for the most part, prove beneficial. The new approach, for example, emphasises a novel kind of teamwork in systems development: business managers and computer professionals working together on systems specifications. That alone could prevent some of the notorious computer disasters which occurred because managers, short on systems skills or experience, demanded more than the technology was capable of delivering. The changes in IT practice forced on the industry by the e-commerce phenomenon are tracked and analysed by Michael Earl, professor of information management at the London Business School, and Bushra Khan of Computer Sciences Corporation in the autumn issue of the MIT Sloan Management Review*. They conclude that in e-commerce companies, the separation of "business" and "IT" has become obsolete: "First, some of the new skills are not yet codified into specialist jobs. Second, the boutique organisation (of small e-commerce companies) deliberately encourages collaboration. And finally, because systems development is business development, interdependence is logical," the pair write.
A discipline
Software development began as a craft. From the first, however, programmers were anxious to see their speciality accepted as an engineering discipline in its own right. Computer jobs became classified by activity: programmer, systems analyst, team leader, and so on. Proprietary methodologies for efficient software production were developed. Project management software was written. The overall aim was to find "best practice" in developing robust software on time and on budget. It was partially successful. Several of the methodologies were successful in bringing some discipline to what had been an anarchic business. But too many computer projects continue to be completed over time and over budget with too little attention paid to documentation. The emergence of e-commerce companies seemed to herald a return to the anarchic early days. The principle difference was that IT, rather than being an expensive and remote supplier of data services to the core company, was the company, so important was IT as a channel of delivery. Prof Earl and Mr Khan say of the pioneers: "Because e-commerce could not survive without technology, they believed, IT was no longer merely on the critical path of business development; it had become the critical path."
New imperatives
There were other changes. Because speed to market was an imperative for e-commerce groups, time rather than cost became the key attribute for the new systems, with structured engineering methods giving way to a craft approach where aims and objectives were assessed and changed monthly if not weekly. Stability was sacrificed for speed: "When we look back at early versions of what we built and took to market, we shudder," one company told Prof Earl and Mr Khan. Work style changed from structured to entrepreneurial, work spirit from serious to fun, the attitude to the technology from essential to the business to the making or breaking of the business. Nothing was built to last; companies learned from mistakes made while developing live systems. A novel, three-tier architecture evolved to provide some order. The first tier was the ephemeral, customer-facing or front office systems which gave the company its identity, the third, the older, more stable back-office applications, such as book-keeping and transaction processing. In the middle, the second tier translated and co-related information between the front and back office systems. With the collapse of the dotcom sector, the urgency, the need for ever-shorter times to market has largely disappeared. Much of what happened in that period must seem retrogressive - flaky systems built without regard for engineering principles, poor documentation and little maintenance. On the plus side, business managers were forced to work with IT specialists on a regular basis as systems specifications evolved while three-tier architectures suggest a way in which companies can adapt to rapid change. E-commerce is merely hibernating. All the IT industry learned in its first, brief flowering will be needed when it returns to dominate the commercial landscape.
* http://mitsloan.mit.edu/smr/
|