As the mainstream software industry passes into its fourth decade, it is increasingly clear that the power to deliver the backbones of computing systems is becoming concentrated in the hands of a few very large, very powerful companies - including IBM, Microsoft, Oracle and Sun Microsystems.
Historically, these companies have maintained their own, almost discrete market segments, differentiated through their approaches to solving problems; but over the past five years or so something interesting has happened to these players and their views of the world.
Globalisation and market saturation have forced them to become more homogeneous in their market positions. IBM now makes a determined effort to deliver infrastructure to all-comers, not just Big Iron lovers; Sun and HP are busy building their own mainframe technologies (but calling them something else of course); and Microsoft claims that the PC is no longer its chief fixation. Network-based computing is the target, instead. Oh, hang on, though, wasn't it Sun's chief executive, Scott McNealy, who originally said, "The network is the computer"?
The Internet has provided a major catalyst for this coming-together of views. Microsoft is desperate to ensure that it isn't caught out by the Internet again; Sun claims it has powered the development of the World-Wide Web since its inception, and wants to keep up that image; HP is desperate to develop a software business and knows that such business has to be Internet-focused. IBM's revitalised focus on an e-business strategy is widely acknowledged as being a great success.
And it's the Internet - and these companies' desires to offer compelling visions of the Internet's future and how they will play in it - that is doing some very strange things to the IT "backbone" supplier landscape.
What is strange about the recent rash of "Internet vision" announcements from the major players is that despite their different names and subtle twists in their content, these "technology roadmaps" have two core shared beliefs:
n That the Internet will evolve into a global middleware platform which enables online software applications and components (variously called "Web services", "e-services" or similar) to collaborate to carry out services for users
n That the Internet will soon reach way beyond today's universe of PCs - not only to mobile handsets; but also to TV sets, cars, consumer appliances, elevators and vending machines.
Behind the rhetoric the major software players are finally starting to build bridges between their products and strategies.
It seems that the strategists at Microsoft, IBM, Sun, Oracle and HP have realised that the colossal market opportunity that Internet MkII brings will be squandered unless they can deliver interoperable technologies.
The technology tool which companies are using to build these bridges is an emerging, seemingly innocuous software specification called Simple Object Access Protocol (Soap ). This XML-based specification maps out how a software program should accept requests from other programs, and respond to them, using the Internet as the communications medium. Never before have these protagonists agreed to share such a technology. Indeed, most of today's middleware is based on technology that is explicitly designed either to run on Microsoft platforms or on anything but Microsoft platforms.
Until now, building complex distributed systems required you to back one or other side; you either built using Microsoft's Com/DCom, OMG's Corba or J2EE technology from Sun. If the suppliers continue down their current paths, Soap and the other "Web services" standards such as UDDI and WDSL could sweep that uncertainty away. Skills will be more determined by the environments in which you operate, rather than the technologies that suppliers dictate to you as part of their product architectures.
Neil Ward-Dutton is research director of e-infrastructure at Ovum.