These suppliers found their customers battening down the hatches in 1999, unwilling to spend any of their IT budgets on new projects while they were preparing for the possible havoc of the bug. The effects of that "millennium lockdown" on this section of the IT industry were like a "nuclear winter", according to one observer.

Two US-based companies in the thick of that winter were PeopleSoft, historically a maker of human resources software which in recent years has broadened its offering to encompass supply chain systems and customer relationship management, and JD Edwards, a venerable vendor of financial, accounting and supply chain management software.

In each case, a long-serving chief executive had just retired, leaving a newcomer in his place. PeopleSoft's founder Dave Duffield, after 12 years at the top, gave way to Craig Conway, previously a senior executive at Oracle, in September 1999.

JD Edwards reached a crisis of profitability following the retirement of its founder and chief executive Ed McVaney in October 1998. A brief interregnum under Doug Massingill saw the share price collapse, and Mr McVaney returned to the helm in April this year.

Both chief executives, the newcomer and the returner, are now faced with the task of pulling their companies out of the doldrums and into the internet age.

For Craig Conway, the transition that enterprise software companies are going through to turn their applications to face the internet is the biggest since the change from a mainframe topology to client-server technologies ten years ago.

"These transitions happen every ten years, and the amazing thing is that some companies never manage to get through it," he explains.

Mr Conway points to the examples of Dun & Bradstreet and MSA, two of the world's biggest software companies a decade ago, but which most people now have scarcely heard of in relation to software. Their problem, he says, was "not being able to make the change from one architecture to another, mainframe to client-server."

Today's transition to the web is no less challenging, he believes, and involves a similar architectural shift. He expects an equivalent casualty rate, with some companies so weakened that they are losing their independence. One example he gives is Baan, the struggling Dutch-based enterprise software vendor which was snapped up this summer by Invensys of the UK.

Mr Conway is clear that the move to the web should not be taken as a simple matter of porting software from the desktop to the internet, perhaps with a new set of interfaces to the new medium. Dun & Bradstreet and MSA lost their previous eminence in software because they tried to port their products from one architecture to another instead of rewriting them from scratch, he claims.

"Re-coding from scratch takes a huge investment, but the result is that you get a pure architecture, a pure internet architecture, and by that I mean no code on the client-so that the applications can be accessed from a mobile phone, say," adds Mr Conway.

To that end, PeopleSoft has shifted its entire product development resources to rewriting its product set, involving an expenditure of more than $500m in the last two and a half years. Research and development accounted for 27 per cent of the company's revenues last year.

All this amounts to a rather big gamble if it turns out that companies do not want to access their corporate payroll applications from a mobile phone. Mr Conway, however, boldly brushes aside such fears. The point of e-business, he says, is to enable companies to run their businesses faster, which means on the move.

"E-business lets you improve processes. Companies are not doing anything today that they didn't do before-just doing it faster," he says. The immediacy of internet communication squeezes out inefficiencies in business processes, enabling companies, such as Cisco, to garner 80 per cent of product orders through the internet without human intervention.

So much for simply improving existing processes: the greatest contribution technology has made to generating new ways of doing business is in the field of analytic software, according to Mr Conway. Companies now have access to a broad range of business analysis software tools, which can spot patterns in business data that can be used to generate greater profits.

These business analytics are the fruit of artificial intelligence research. "Everyone thought artificial intelligence was going to somehow mirror human thought-that's not what it ended up being, but it did end up in pattern recognition," he says.

Mr Conway cites an example of how such analysis could be used: Firestone tyres, which Ford recently had to recall because of faults. "A business intelligence system would have said to Ford, there's a correlation between accidents and tyres: look into it!"

Companies must adopt these analytic approaches as soon as they become available, he suggests, or risk being overtaken by nimbler rivals, as the fast pace of internet life means less tolerance for tardiness than ever before.

The pace of Mr Conway's own career has mirrored that of the technology he cites: after leaving his job as executive vice president at Oracle in mid 1993, he served as chief executive of TGV Software, which made applications for corporate intranets, and OneTouch systems, an interactive broadcast network maker, before joining PeopleSoft as chief operating officer in early 1999.

Looking into the future, Mr Conway sees that pace accelerating, with the introduction of what he calls the "evernet".

The term evernet describes the condition of being wired or connected all the time, through a cellphone, PDA, or whatever. "There are personal issues here-would you want to be available 24 hours a day? But the point is that you can," he says. The challenge for company executives is never having to leave the office-how can they, if the office is in their back pockets?

Ed McVaney wears a tie bearing little sketches of sportsmen hunting, shooting and fishing-perhaps an emblem of the retirement from which he was prematurely plucked. Having left the helm of JD Edwards in late 1998, in April this year Mr McVaney returned to the company he founded to rescue it from a parlous financial state.

Shareholder confidence had evaporated as the company was seen to be failing to deliver its software, and failing to compete against the likes of Oracle. The share price plunged below $12 in March 1999. It subsequently recovered in the general ebullience surrounding technology stocks early this year, only to plummet again to a low of $10.25 following poor financial results in May. Now, Mr McVaney insists, the company is back on track.

"I take the blame: I failed at retirement, failed at picking a successor," he says. Coming back, however, has been fun. "Quite frankly, it was really reinvigorating-I think I came back a much better person than when I left."

Mr McVaney blames too many cooks for the company's lack of strategic direction without him. "When I left, I was the majority stockholder, 58 years old, founder of the company-there was just no question who the leader was. Then there was a crew of about ten 44-year-old people who assumed leadership and they started-'fighting' is not the right word-debating among themselves, and there was such a sharing of power that there was a lack of leadership," he explains.

On his return, he initiated a bout of restructuring to cut costs, which had gone "whacko", he says. And then he set about forging a clear goal for the company.

The internet would have been an obvious place to start, and JD Edwards already had an internet focus, remodelling its flagship OneWorld platform to fit e-business.

But although Mr McVaney is very clear about the huge strategic importance of the internet, he prefers to point to a much less glamorous technological advance that he believes is transforming the enterprise software business. That is enterprise application integration, or EAI.

EAI means drawing up a set of rules that software development should follow, in order that different applications from different vendors (the "best-of-breed" in each area, such as the best accounting package or best human resources software) should be able to work together.

That sounds simple enough and prompts the question of why it has not been done before. The problem has been that software companies often gave little thought to how their products would work with a multiplicity of wares from other vendors. Getting any level of integration between disparate best-of-breed products proved a nightmare for many customers.

"Best-of-breed was always a great idea that never worked," Mr McVaney admits. "And the reason it never worked was that interfaces were so troublesome."

He gives an example of a company wanting to use PeopleSoft's human resources software with JD Edwards' accounting software. It would have had to set engineers to tweak both PeopleSoft's and JD Edwards' code to let them work together, and even then they would have worked only until a new version of either product was released-when the process would have had to start again.

With a new strategy built on EAI, Mr McVaney hopes to reverse 22 years of history at JD Edwards, when it took the view that "you could get everything you could ever want to get from JD Edwards", rather than integrating with other products.

The web and e-business have accelerated the need for EAI, Mr McVaney notes, because companies are now having to deal with electronic networks of customers and suppliers who all have different systems. EAI technology forms the nuts and bolts that will fasten these systems together.

However, he will face an uphill battle in winning over customers who have already suffered at the hands of best-of-breed systems.

Mr McVaney's extra-curricular activities involve similarly uphill struggles, and mirror his professional interest in cohesion. His "biggest thing", he says, is inner-city education, the failure of which in the US threatens such a disruption to social integration that "if we don't get it right, we'll be living in gated communities, because crime will get so bad".

As he notes, the issue is important to technologists, too. Sorting out inner-city education could go some way to solving the IT industry's skills shortage.