To reach these figures telecom companies would need to start key WLL projects this year. However, according to Mr May, the projections will have to be reduced by 70 per cent for 2000 and by 50 per cent for 2002.

WLL is an application of radio technology. Links from the local switching system to the home or company user can be completely wireless, or partly wireless.

For example, a home telephone with a small antenna communicates with a base station, which is connected to the access node of WLL and then via conventional wiring to the public switched telephone network.

WLL is suitable particularly in rural areas because it is quicker to install than laying fixed copper cables underground.

In Asia, WLL will be a fast and inexpensive way to bring telecommunications services to rural areas where there are no telephones.

Studies show that in areas with low population density - less than 500 person per square kilometre - the wireline cost per line, estimated to be between US$600 and $3,600 depending on the population density and telephone penetration, is too high.

By comparison, WLL cost per subscriber is calculated about US$600.

Additionally, in some areas it is impossible to lay copper lines.

Despite the tough economic climate, Mr May believes telecom operators should still invest in infrastructure projects such as WLL, as they are encouraged by the World Bank and supported by local governments.

'These networks are 10 to 15 years investments, so even if the region is affected for three years they are still better investments than those in coconut juice bottling plants,' he said.

However, investors have to take a long-term view. Systems proposed so far are based on telephone lines or narrow-band WLL, offering the most basic voice capabilities plus some enhanced services such as call waiting and fax.

Vendors believe the technology can be upgraded to wideband WLL, to support high-speed multimedia and Internet.

'Although vendors say their equipment will be able to support these features, they have not spelled out how this is going to be done and when,' Mr May said.

A leading competitor to WLL is microwave, another radio technology. Microwave also is quick and easy to implement, offers flexible capacity and already can support multimedia capabilities.

However, Mr May believes the technology is immature and lacks products.

So if WLL is so promising, why did it not take off earlier? The first reason is cost. Vendors are producing systems at between $700 to $1,000 per line while operators are not ready to pay more than $500, especially in poor, rural areas.

The second reason is political. For a long time, developing countries considered WLL to be a second-class technology. It has taken vendors two to three years to convince them otherwise.

Large-scale projects such as Motorola's system in Hungary involving 200,000 lines and other projects involving up to 4,000 lines in Indonesia and India sent a positive signal.

Mr May believes operators would have turned to WLL this year if not for the financial crisis.

In Indonesia, the per subscriber price is three times what it used to be.

'Plans that were in place are being reviewed,' Mr May said.

The third reason is the uncertainty about which wireless technology will be used as a single standard for wide-band communications, that two international bodies have agreed to deliver by 2000.

'Vendors talking about future developments are killing the market. A lot of operators do not want to jump on a technology that may become obsolete in a few years,' Mr May said.

He believes WLL's best prospects are in the mainland and India, where a newly emerging middle-class is demanding telecom service, and where vendors can offer systems at keen prices.