How do you eat an elephant? One mouthful at a time...

Or so the joke beloved of six-year-old boys used to go when I was a six-year-old boy.

If that is the case, there are a lot of elephants looking nervously at hungry IT professionals in the UK.

As many of you know, Computing runs a regular web seminar programme, inviting readers to hear from expert speakers on a particular technology trend or issue, and to pose questions to those experts.

Invariably, the presentations cover best practice and typically describe an ideal solution to users’ business requirements. All useful information for IT managers.

But almost every time the same question will be asked by someone in the audience, along the lines of:

“This sounds great, but my IT set-up is a long way from there. Where do I start?”

And this is the elephant in the room that too many IT vendors somehow fail to notice. It is all very well offering great new products and solutions, but most IT departments are managing legacy systems going back years and with a complexity that is difficult to unravel. If you want to get me to that place you describe, say IT managers to vendors, then I would not start from here.

I was having a similar discussion about service-oriented architecture (SOA) recently, and it is a good example of the problem.

Most IT directors in large companies have bought into the principles and promise of SOA ­ greater flexibility and productivity, easier integration, reuse of code, and so on ­ but putting it into practice is another matter. They are inundated with vendors claiming to be SOA experts, and the conversation ends up as a sales pitch.

But SOA is a perfect elephant. Too huge to get round, and not appropriate in the middle of what you have now. So, take a small bite first, and see how you get on.

IT is usually seen as a huge endeavour, a major project to be tackled. More often, it is best viewed as an iterative process, broken down into manageable parts that are well-defined and easier to implement.

IT directors need to put themselves at the head of the table, set out their plans and strategies, and fit vendors kicking and screaming into that. Then, once the cutlery and plates are in place, it is time to start eating ­ bit by bit.

Google is now the fifth biggest company in the US, based on its market capitalisation. The search firm’s share price passed $700 (£336) last week, which means the business is more highly valued than Ford, General Motors, IBM, Bank of America, Boeing and a host of other household names that ring up a great deal more in sales than Google and turn in a greater profit.

And when Microsoft paid $240m (£115m) for a 1.6 per cent stake in Facebook last month, the social networking site was valued at $15bn (£7.2bn).

With comparisons being drawn with the ludicrous company valuations that led to the huge dot com bust six years ago, influential voices are now asking the question: is a new dot com bubble starting to appear?

The question could hardly come at a worse time.

The financial markets in London and New York are in a state of high caution. In the UK, the knock-on effects of the run on Northern Rock are still being felt, and banks are worried that other companies in the sector are teetering on the brink.

In the US, a second drop in interest rates came in response to fears of an inflation-led recession as the sub-prime mortgage crisis continues to cause shockwaves across Wall Street.

Unlike the dot com era, the technology companies in question can hardly be blamed for their current valuations.

Google has historically been almost anti-Wall Street, reluctant to offer guidance on its future sales and leaving financial analysts to make guesses about fiscal performance ­ and hence to send the share price spiralling upwards in unbounded optimism. Facebook is a private company and doesn’t publish its results. A little bit of mystery is clearly not a bad thing.

Share price valuations aren’t, in reality, that different from gambling odds. Financial experts making a bet on the future of a company go through much the same analysis as a punter poring over Sporting Life.

Ultimately, nobody really knows the answers. But the after-effects of a bad share punt are more significant than the result of the 3.15 at Kempton Park.

It doesn’t actually matter if Google or Facebook are worth these apparently daft valuations. What matters is the underlying state of the financial markets. A balloon can safely fly as high as it likes, provided it doesn’t hit bad weather.

Any company that can acquire customers as rapidly as Google and Facebook has a business model worth investing in. But those investors have a long way to fall if the rain clouds form ­ and the danger is that the IT sector will again be caught up in the storm.