Web services are like trainee staff

Peter Judge: Even if Web services are everything they are cracked up to be, we won't trust them till they pass their probation.

When you take on a new staff member, how quickly do you expect results? If it's a senior person, you might see some quick benefit to the share price, but it will be months before they understand the company enough to make changes -- and disaster if they try to change things before then. If it's a junior person, you get smaller results -- but you get them quicker.

When you look at Web services, remember that they are not the IT equivalent of a new CEO, or even a new CFO. They represent a new and very hard-working junior admin person, but one with a lot of potential.

The marketing people say Web services are ready to use because the standards work is more or less done. Now, this is as much of an exaggeration as you'd expect from marketing people. What it means in practice is that the basic SOAP/UDDI/WSDL trio of standards is solid, and other essential stuff like security and business process modelling is at least underway.

Like most of the other significant stuff, security standards will go through the influential OASIS group, and there should be some sort of draft next year.

But with that proviso, I'd say that Web services are ready to use. There are early adopters using them, and building your own security onto the system is inconvenient but worth doing if there is some real advantage you can see from Web services in the end.

The problem is that, to get that advantage, you have to get your people to use those Web services-based systems.

And this is where I think it is going to get interesting. Because here we have a very human issue.

Web services will not reproduce the colossal white elephant systems of the past. There won't be a huge investment, long lead-time, big bang approach (like the EDI systems of the 1980s, or the EAI engines of the 90s). They will be put together in small, manageable, chunks from the bottom up. The things that get made will be the ones with a quick payback.

But with a bottom-up system, the problem is, to pay back quickly, the new systems have to be picked up easily and brought into real use.

Web services-based systems will have to model ALL the processes associated with a business action. For example, when a furniture manufacturer receives a purchase order for some chairs, a whole string of questions must be answered. Does the buyer have enough credit to pay for the order? Can the manufacturer get all the parts? Has it got the manufacturing capacity to put them together?

Unfortunately, all these questions do not have yes or no answers. The parts might be available just after the deadline, or they might be available if a smaller customer's order is delayed. In that situation, a smart human sales person will start negotiating with the customer, perhaps offering a lower price for a few days' delay.

Web services, we are told, will be intelligent enough to handle this kind of negotiation. But will anyone trust them? "Web services will start out in an advisory role," says Bob Sutor, IBM's director of e-business standards strategy. "Once the confidence level rises, they will be given more trust. They will be treated like new employees."

If you had a software agent to buy stuff on the Web for you (and this has been tried a lot of times) you might trust it to buy a book, but not a car or a house. The same goes for Web services in business, says Sutor.

In many ways this advisory role suits where Web services are now, since they are currently (reportedly) a bit slow, owing to the overheads of processing XML and the other new standards. By the time they are ready for more widespread use, they will hopefully run a bit faster.

So Web services is like a very capable new staff member. They will pick up your company's way of doing things, and gradually win the trust of existing employees. But they won't change your business overnight.

To have your say online click on TalkBack and go to the ZDNet UK forums.