The Friday interview: Michael O’Duffy, Centre for Software Engineering

25 Jun 2004

As someone so close to the software business, it clearly irks Michael O’Duffy (pictured) that major IT projects often come in for bad press or are unfairly perceived as failures. “There’s a tendency to say that people in the IT world are just not able to manage projects. I think it’s unfair for a number of reasons. We don’t hear statistics associated with major projects outside the IT world and there’s a much higher rate of failure there than is generally stated.”

A very public real-world example was the construction of the Central Bank on Dublin’s Dame Street. The cost of excavation ended up costing double its estimated budget. It was also built 30ft higher than originally designed and the copper cladding intended for the roof was never added. “If that was an IT project it would be regarded as a failure,” O’Duffy wryly remarks.

“The IT project is very different in a number of ways, starting from the base that what is being built is something intangible, a piece of intellectual property. It’s not physical and because of that it’s far more difficult to understand what exactly is required and because it’s something that’s in the mind it’s far more complex.

“It’s like an organic thing, it’s changing, it’s evolving, so the idea of comparing it with physical building projects is unfair and I think it’s also unfair to say that people who do software are nerds and they’re possibly partially qualified and they’re working with things maybe they’re not completely on top of (laughs). There is that, but at the same time, they must be particularly good at what they do to achieve success with the challenge that is in those projects.”

O’Duffy has seen more than a few large-scale IT projects in his time. He is chief executive of the Centre for Software Engineering, which has provided high-level consulting services to senior management in the Irish software sector since 1991. It also offers specialist programmes such as software project management as well as providing training to industry in Ireland and abroad.

“Our primary role is to assist software-based organisations to improve their capability,” O’Duffy explains. Breaking down the role further, he says there are three client sets: Irish software companies; multinationals that have set up with a view to selling their software here; and organisations that develop bespoke software, in the public sector organisations as well as banks, insurance companies and food providers. “If you go into your local GP, for instance, very likely he’s using a computer system that we’ve certified,” O’Duffy notes.

O’Duffy is understandably keen to emphasise the positives when it comes to IT projects and he welcomes recent changes in the way these jobs are being managed. “You’ve got to get people in the organisation who know exactly what they’re doing – and it’s difficult to do that nowadays because an awful lot of the really important systems are forging new areas of business.” That novelty factor means it can be tricky to get people to identify what they want. “Part of what happens now with those kinds of projects is that you kind of work through an evolving requirement, so the requirement becomes clarified and elaborated during the course of the project.”

A successful project needs very clear outlines for what is to be provided – whether that be a government department rolling out a new service to citizens or a bank introducing a new product. O’Duffy observes that facilitated workshops – a new term for what is effectively a good old brainstorming session – are coming into vogue and these can help organisations to form much clearer ideas about what they want to do and how their IT can help them do it.

At the next stage of the process, software developers return with a prototype and there is further discussion to see what features are present or missing in the code. This in itself throws up issues around managing how the scope of the project stays within proper boundaries. “There’s got to be a way of managing scope and there are techniques for that. The biggest of all is the area of setting priorities,” O’Duffy says.

In software development, as with politics, there is an element of compromise. The business leader and the project leader need to work together to decide what elements of the work are critical and which others can be left out. “You might say that [a particular feature] is absolutely crucial, the system can’t work with integrity without incorporating that particular requirement, so they see the detail of that and then you get down into the detail and you see that three of those are crucial to making this crucial thing happen, then the others are not so important. And that’s usually the way in which you manage the project. And it’s not two different factions that are arguing out a case – individuals are empowered to make decisions without having to go to a supervisor. The project is never stopped or held back by the inability to make decisions.”

For all this advancement, not everyone adopts this methodology and there is still room for improvement, O’Duffy stresses. “You’ve got to bring in a much greater amount of skill sets than perhaps we do in practice. Some people would say maybe the best thing to do with a software project is to bring in highly educated software people, but what we need are skill sets from other disciplines, including the social disciplines. Because a lot of IT systems engage with the users, they must understand the social engineering aspects. That’s quite important.”

Irish software projects often overlook the issue of usability. O’Duffy cites one prominent Irish organisation that had supplied its customer base with a data entry system that was designed for expert users, despite the fact that it was only meant to be used once per quarter for a couple of days at a time. “The person using that system may not necessarily be familiar with computers generally, is only using this system for a day or so every three months and will have forgotten about how to use it from the last time. There’s no guarantee when it comes to three months forward that it will be the same person in the organisation. So the nature of the use of that system should have been for what might be described as the very casual user as distinct from the expert user. So we tend to still make mistakes on that.”

O’Duffy has also seen projects come asunder due to criticisms of one party by the other. Poor organisation is often to blame here, he says. “You have this breakdown of saying: ‘you’re responsible for testing’, ‘you’re responsible for development’, ‘you’re responsible for business analysis’. And because there is this sense of walls between them, there isn’t one of pulling all together, so there can be disputes.”

For these reasons and more, O’Duffy clearly feels that the CSE still has a role to play in encouraging best practice. When the organisation nearly folded last year, there had been whispers that it had outlived its usefulness but he argues otherwise.

The role has changed, he acknowledges. When the CSE started up, it was a young industry but with good practice now in place, the attention has turned to improving strategic focus. “It needs a greater sense of maturity about the processes that are used. And the projects in the public sector, in the banks and all of that, they need to be developed such that the failure is minimised and that they are much more in tune with their organisation’s needs and that they are keeping up to date and able to deliver projects very quickly. So it’s a different type of approach now, but absolutely needed.”

By Gordon Smith

Loading now, one moment please! Loading