Good times are the best times for producing rosy scenarios. Things always look rosy right up to the point that they turn unrosy. That’s particularly true in the advertising world. It’s a business that can never stop forgetting that it’s as cyclical as the pork-belly trade.

In an article today, CNET’s Elinor Mills surveys some of the current rosy forecasts for internet advertising. Parks Associates says online ad spending in the US will rise to $23.5 billion by 2010, up from about $15 billion today. Piper Jaffray trumps that with a forecast that worldwide internet ad sales will jump to $55 billion in 2010, from the current $19.5 billion. “By most accounts,” writes Mills, “the online-ad market barely has been tapped.”

Back in 2000, the professional forecasters were equally giddy. In August 2000, Jupiter predicted a continuing boom in internet ads, with spending in the US growing at an annual clip of 30% through 2005. Online ad spending would rise, it said, from $3.5 billion in 1999 to $9.5 billion in 2002 to $14.3 billion in 2004. Veronis Shuler’s Communications Industry Forecast, released at about the same time, painted an even prettier picture. It projected that the US online ad market would grow 40% per year, reaching $24.4 billion in 2004.

Of course, as we know now, these rosy forecasts were made at precisely the point of a cyclical peak in internet ad spending. Rather than going up, internet ad sales plummeted by 25%, from $8 billion to $6 billion, between 2000 and 2002. The cycle then turned upward again, with sales climbing to $9.6 billion in 2004.

Yes, things have changed between 2000 and today, and, yes, internet ad spending will likely increase over the long haul. But two things haven’t changed: ad spending will continue to be cyclical, and we won’t see the peaks until we’re past them.

Here’s my favorite line from the CNET piece: “It’s also not a given that companies won’t again cut their online-ad budgets, experts said.” That’s why they call them experts.