Open systems are crucial to networking: unless the machines are designed to a consistent standard, it can be difficult to link them together. It is a bit like having a different type of plug for each electrical appliance in the house.
Hardware manufacturers must have a networking strategy if they want large, lucrative corporate customers to take them seriously, and this will add to their research and development expenditure. "We will not stop developing hardware, but unless we have a solid way of linking machines together we won't sell them," says Digital Equipment's European networks marketing manager, Paul Evans.
While some, such as Digital and ICL, have embraced networks as an integral part of their strategy, others, notably Unisys, Wang and Data General, have been slower to take up the gauntlet. In 1987 Unisys bought a networks company, Timeplex, but has failed to take advantage of the skills and customer base which it offers.
But as customers become more familiar with computer systems they recognise that it is the applications software that determines the usefulness of the machines to their business activities. They are demanding standard hardware and specialist software. This is the second growth area of the computer industry and many companies which have specialised in designing software are now facing new competition from the hardware manufacturers that are moving in.
Many companies were caught out by this shift. Nixdorf, Apollo, Wang and Philips all have similar stories to tell. The first two have found white knights to save them; the last two are still trying to do it themselves.
Apollo was itself a niche player like Sun, but did not move early enough to build a machine that conformed to a standard architecture. Hewlett-Packard saw the Apollo purchase as a means of increasing its own share of the workstation market, but it also took immediate action to revamp the product range, and within the space of a year all Hewlett-Packard workstations were built to open systems standards.
Nixdorf also foundered because it failed to recognise quickly enough that it had to build machines to global standards. After growing rapidly in the mid-1980s with its minicomputer-based systems designed for vertical markets such as banking and retail, it ran into trouble and in 1989 reported a nine-month loss of DM 465 million.
After Siemens bought it in 1990, Nixdorf was fully integrated with Siemens' computer business. David Cliffe, international business development manager at the new company, Siemens Nixdorf Information Systems, says that the way forward is to move away from hardware to being more software and services orientated. It will also be able to build on the combined strengths of the two companies to expand geographically. "They (Nixdorf) are strong in international areas where we are weaker, such as France, Spain and the UK," Cliffe says.
Philips' computer business has similar problems, and it too is making heavy losses. "It is an entity in trouble," says Dr Jane Dawley, an analyst with market research company Dataquest. "It has an outdated product line and the R and D to keep ahead is enormous and not sustainable unless you are large."
Philips has taken action to reduce its R and D costs: it will cease production altogether of its minicomputer line, and will combine its home computer and its business PC development programmes. Like Nixdorf, it will move towards becoming a systems integrator rather than a manufacturer.
Wang, once highly successful, is probably in the deepest trouble, having made huge losses and done little to rationalise product lines. The problem is believed to lie at the top of the company, where the family of the founder, Dr An Wang, has hung on to the company and resisted making major changes.