The internet changes everything and nothing. The means of delivery may have been updated but the business decisions which determine how an organisation manages its information technology remain much the same as they have since the advent of the computer.

Look, for example, at the emergence of Application Service Providers (ASPs) - companies which, for a fee, provide customers with most or all of their computing requirements. Software, generally of an industry standard variety, is stored centrally and downloaded over the internet as required to customer sites.

A broad range of organisations, from large telecommunications operators such as AT&T and British Telecommunications to specialists such as Esoft Global, are investing heavily in ASP technology in the belief that it represents a large part of the future of business computing.

It is one of a number of outsourcing options which are now available to information technology managers. Others include web hosting, where a service provider takes responsibility for mounting and maintaining a customer's website and storage service provision (SSP), where a customer's large scale memory requirements are met by the establishment of networks of storage systems.

There are powerful reasons why outsourcing software and memory provision to third parties makes sense. Modern computer hardware and software is complex: making it all work together, keeping it maintained and dealing with upgrades is a daunting task best left to experts.

Most companies these days would rather focus on their core business than support large information technology departments. And their IT directors are tired of trying to cope with backlogs and maintenance problems.

From the ASPs' point of view, economies of scale can be found in offering industry standard systems to a diversity of customers. From the customer's viewpoint, it could mean the end of dumping the company's entire complement of desktop computers every two years or so.

ASPs, however, are simply computer bureaux by another name and will be prone to the difficulties bureaux experienced in the 1970s and 1980s. First, there is the question of which kind of companies will make satisfactory service providers. Some are experienced at providing telephone services but have little experience of computing services provision.

This raises a question central to the whole issue of outsourcing: can corporate customers be persuaded to trust an outside party with their crown jewels - their corporate data and communications.

In practice, many have proved willing to trust a services group with their bread and butter applications - payroll, bought ledger and so on - while keeping competitive edge developments in-house.

The fact remains that standard software provided by computer bureaux rarely delivered the promised benefits. Few companies were so standard that they felt comfortable using the same software as their competitors down the road. Bespoke solutions were invariably more satisfactory. But they were expensive and took time. In a world where time to market is critical, bespoke may be simply too slow.

To date, the evidence is that small- and medium-sized companies rather than large corporations are attracted to the ASP option. This may change as the difficulty of managing IT increases and the credibility of ASPs improves.

The key, of course, is the quality of the service level agreement between the company and the ASP and the provisions within it for penalties for failure to provide the promised benefits.

There are also physical and environmental considerations. The equipment used to run the internet uses much more power than conventional telecoms kit. According to one estimate a 50,000 square foot facility can consume as much power as a small town. The dangers of power failure or terrorist activity are obvious.

In the end, however, IT is a cyclical business. Computer bureaux failed because they could not compete with the flexibility and economics of the microcomputer.

The ASP route seems attractive at present because of the complexity of modern computer systems and the internet. There is huge appeal in leaving the responsibility to a third party.

That said, within a few years it seems likely that computers will have become so powerful that much, if not all, of that complexity, can be hidden from the user without extra cost. At that point, the attraction. once more, of in-house computing may be too great to resist.