Motoring: The way forward, or mere diversion?: Christopher Columbus] Electronic route planning] But an atlas will do just as well for Phil Llewellin

Maps of one sort or another have been guiding travellers ever since the Babylonians recorded 'how to get there' information on clay tablets about 5,000 years ago. Then, in 1675, the first road atlas was published by a surveyor, John Ogilby, whose basic equipment included a good pair of boots.

Nevertheless, the makers of Columbus insist that it is 'the world's first portable route-planning system' - and technology's answer to the old book of road maps is bound to interest anyone who spends a lot of time at the wheel.

Columbus is a keyboard-and-screen instrument only slightly bigger than a mobile phone, but packed with comprehensive information about Britain's road network. Even your 53-year-old reporter, who has only just got to grips with a fax machine, found it easy to use.

The index, which appears at the touch of a key, contains names of almost 40,000 places - nearly twice as many as the average road atlas - ranging from Ab Kettleby to Zouch, taking in Wimbledon Chase railway station and the junction of the A298 and B279 between Weybridge and South Norwood.

What you do, in essence, is to key in your starting point, destination and any detours, then ask Columbus to work out the quickest or shortest route. The details (precise or simplified, as you choose) are screened on a 'page-by-page' basis, and include place-names, road numbers, the length of each stage in miles and, more often than not, an arrow indicating the approximate direction of travel.

Other functions can calculate the estimated time of arrival (which depends on your 'reasonable average speed'), and revise the route in the light of traffic problems. Truck drivers can program Columbus to steer them clear of low bridges and to suggest the most efficient routes for collections and deliveries.

The difference between the shortest and quickest route can be dramatic. I asked Columbus to get me to London Bridge from my home on the northern Welsh marshes. The quickest route, 182.3 miles long, would take 3hr 44min; the shortest, 171.6 miles, 5hr 4min. At the other end of the scale, the complete absence of motorway options explains why Columbus calculated the same time and distance for the shortest and swiftest routes between my Shropshire home and Milford Haven.

Most routes are calculated within a few seconds of pressing the 'go' button. Less than a minute was needed to figure out the quickest route for an unlikely marathon from Dover to Kinlochbervie, a few miles south of Cape Wrath: 752.2 miles and 14hr 31min.

But several weeks with Columbus convinced me that, for the average motorist, it is far from a serious alternative to the road atlas.

It lacks the intelligence to appreciate that an apparently illogical journey can be better than the quickest or the shortest route. Both of its suggested routes from my home to Manchester airport involved roads (via Ellesmere, Whitchurch and Northwich) that I stopped using umpteen years ago when dual-carriageways and motorways were built. The route that bypasses Wrexham and Chester may be much longer, but it is also much quicker.

Not only is the display not always easy to read in a moving car, but it also switches itself off after less than three minutes - to save the batteries - unless you react within five seconds to a warning beep; and fumbling for buttons is not the safest task while driving. The switch-off problem can be avoided by fitting Columbus to the dashboard and connecting it to the car's electrical system, but the kit required costs pounds 29.95 - three times the price of a road atlas.

It is the price, however, that is the real cruncher: the hardware costs pounds 299.50, plus pounds 99.50 for the data-packed 'routeing card', without which it is useless. I wanted to like Columbus, because anything that makes my 50,000 miles a year a bit easier is to be welcomed. But it strikes me as little more than a very expensive toy.

(Photograph omitted)