In a roomful of weathermen it's a hard thing to give a prediction. They want hard facts, pictures of the future. All I can give them is a forecast. And I might be wrong — but they should be used to that.

Outside the hotel in northern Budapest it is swinging from rain through overcast to sunshine. Inside the hotel's conference room mood is similar in its variation; the first annual meeting of the European Meteorological Society, held last week, was hardly likely to ever break into uproar, but some here have strong points to make.

Societies are like turtles, explains Arne Spekat, executive secretary of the EMS, to me earlier in the day. And you are here in Budapest, he says with a smile, to help them speed up a little. This is a tall order.

When you're sitting in front of your TV in the evening and the weather forecast comes on, if you're in the UK you're likely to watch it with some attention. Britain has some of the most changeable weather around, which is why the British have such a traditional fascination with it. Weather forecasters know this, which is why they experience such trauma when they get it wrong.

And forecasters of other nationalities also experience that terror when they give forecasts. One Swedish meteorologist tells me over a beer that the most irksome part of his job is when he has to get on the local Stockholm radio service and tell his countrymen what the weather will be like that day.

"And when I get it wrong, they think I'm lying," he says. "They think we know what it will be like, and are just not telling them." So it isn't just the British who are obsessed with the weather.

The EMS is in an embryonic stage. Stanley Cornford, vice president, is a little nervous that everything will come off well. He's hampered, tragically, by acts of terrorism several weeks ago in the United States, and by an unrelated explosion at a chemical factory in Toulouse on Friday 21 September, in which so far 29 have died. This explosion was very close to Metéo France, the French national meteorological service.

Several people who were expected to come have, therefore, been unable to attend. Stanley Cornford is only a little worried about this, because he has a hall full of people, and the project that he and several other senior meteorologists has seen to fruition is working out just fine.

One of the surprising things about modern meteorology is that, for all the satellites, weather stations, buoys and scientists it has working on it, forecasters still cannot predict what the weather will be like anywhere in the world in 10 days time with any real accuracy. They call this 'skill' — the ability to be more right than wrong over a period of time.

Dr David Burridge, a top-ranking meteorologist and director of the ECMWF (European Centre for Medium-Range Weather Forecasts) in Reading, is trying to change this. Already, using the Ensemble Prediction System, a weather modelling system run on supercomputers, meteorologists can predict with more than 70 percent accuracy the weather over the period of a week, but at the EMS conference he predicted that, with better modelling, accuracy could be extended. By an extra day. To being pretty right after eight or nine days.

It's almost a shock to discover that meteorologists cannot tell what the weather will be like in more than 10 days time, but it shouldn't be. Models become 'unstable' at around the 10-day mark, and succumb to the forces of chaos.

This is because the weather is a fabulously complex beast. It's doubtful it will ever be possible to do much more than two weeks, because there are just too many unpredictable variables.

This is much the same as predicting the technology market. When the hour comes, I have to tell this hall full of meteorologists how I think their industry compares to other information businesses, give them an idea of what devices we'll all be using in three to five years time, and give them a few predictions about the deeper future.

So we cover connected handhelds, PDAs, mobile phones and how to reach them. Most meteorological services already provide 'products', in other words forecasts, to radio and TV stations as well as running their own websites. But what they want to understand is how to make money from selling those forecasts to us, the consumers of the world.

They seem to perk up most when the subject turns to services such as SMS, text messaging. This makes money and can be delivered over the existing infrastructure. In the question and answer period after I've tried to cram too much into half an hour, a man stands up to say his firm is already delivering forecasts over SMS, and making money from it. Harry Otten talks as if I should have heard of him, of the service provided by the company of which he is president — Meteo Consult of the Netherlands.

Perhaps I should have heard of his company's service. Sadly Meteo Consult's service is only available to Dutch mobile users wherever they are in Europe, and the website has no English version.

And this is the problem. It isn't Europe-wide. Meteorology is a scientific business that spans the globe. But so far no one seems to be delivering a forecast service over today's devices across Europe. The future of doing this looks dim if Otten's service is so good that only the Dutch, so far, can use it.

I spoke to Brian Golding of the Met Office, who admitted that the Met Office had tried such a venture, but had chosen to go down the WAP route instead of using text messaging. "Perhaps we should have done that instead," he said ruefully. The Met Office looked into the future and got it wrong.

But there's still time.