Objective view: users slow to spot benefits of object-oriented development

Most people are aware that object-oriented programming has been around for a while, but not many realise that it was first introduced in the 1960s.

Back then, the commercial world was not ready to take advantage of such systems. IT departments were still busy trying to get the basics right and building large, monolithic systems using proprietary software. Unconcerned about future-proofing their systems against the then-distant year 2000 date change-over, developers were hardly going to worry about objects.

But times have changed. It is impossible to walk into an applications development event nowadays without being bombarded by information about objects. In spite of this, many customers still do not understand the benefits the approach can offer. Reports from one or two object vendors reveal that some users are even becoming disillusioned with objects.

Like most other sectors of the industry, the world of object orientation has always had its evangelists. Jonathan Selby, UK director of operations at object-oriented design tool company Cayenne, is one such enthusiast.

As he explains, the modularity offered by object technology is of key importance to businesses.

'We are going through a major transition in the IT business today especially on the software and architecture sides,' he says. 'One of the tenets of quality is that organisations have to build their IT infrastructures to maintain an independence between business modelling at the business process re-engineering level and the IT infrastructure that supports subcomponents of that at a technical level.

'That is where object orientation has a lot to say because it wraps very clearly - it enables people to break down their business processes into components.'

The other key benefit of objects is that they can be re-used, so that in theory, as time goes on, they can cut down development costs.

However, the grass is not as green on the object side of the fence as some people would have you believe. One area causing consternation in the object world is the running battle between the two main object architectures - Microsoft's Dcom standard and the Object Management Group's Common Object Request Broker Architecture (Corba).

In its present incarnation (version 2.0), Corba has been around since 1994. It is a system that enables compliant Object Request Brokers (Orbs) to talk to each other, with the result that heterogenous distributed object systems are easier to create.

As always, some IT vendors have vested interests in their own technology.

Microsoft has flirted with the Object Management Group (OMG) in the past and has promised to give the OMG the specifications for Dcom so that it could roll it into Corba. But the OMG has said that the specifications it received were incomplete.

Microsoft's Internet development manager Mike Pryke-Smith says that Dcom is the way forward for the company's NT software and implies that there is little point in jumping into bed with the OMG at this stage.

'(Dcom) is the best way of implementing these objects across multiple machines. From that point of view we are very much committed towards developing Dcom,' says Smith. 'As far as the technology goes, to bridge Dcom and Corba - we don't see that as an area of expertise but third parties are already moving into that space. Hewlett Packard has announced the fact that it is coming out with bridging technology between Corba and Dcom.'

Microsoft's antipathy towards the OMG is perhaps revealed by the fact that, at the time of writing, Microsoft, although still officially a member of the OMG, had only just paid its subscription fee which was months overdue.

The situation is one that open systems users should be used to by now, but it makes things no less confusing. Decisions about whether to take the object-oriented application development route will be clouded by this industry agreement between two large forces in the market. Confusion over standards, however, is the least of users' worries, according to Steve Larwood, business development manager at applications development company JBA. Larwood says he has renounced object-oriented standards in favour of something he calls 'pragmatic objects'.

He argues that developing object-oriented applications based on industry standard frameworks is too time-consuming and complicated. A better solution is to package up your existing code in an object-like system, effectively parcelling it up into components so that it can be re-used easily. This avoids the need to rewrite your systems, he adds.

'You have all your intellectual copyright and intellectual knowledge and business contained in your existing code and you don't want to throw that away,' he says. 'Instead, you break out the business logic and wrap it into something de facto like ActiveX.' The technology allows you to wrap your code in ActiveX controls that can run on non-PC platforms, and the ActiveX controls can be joined together.

'You get the benefits of object orientation without having to comply with any standards,' Larwood says. He dismisses concerns that this approach is dangerous for users, even though the aim of a standards initiative is to prevent the object world from becoming fragmented.

Larwood is unfazed by suggestions that his home-grown technology could risk ostracising users from the rest of the market. Apart from the fact that users can add standards-compliant objects into their system later, he points out that his customer base does not want a 'pick and mix' approach to applications development.

Attractive though standards compliance may be, Larwood has identified a major problem with objects from a user perspective - the 'no gain without pain' principle. Users wanting to adhere to standards and reap the benefits of re-use have to undergo a costly and time-consuming rewrite of their systems first.

Mel Earp, technical director at systems integration giant Sema Group, warns that objects must be a long-term investment. 'I do think re-use is a positive thing but I am also convinced by the argument that with the first project you try to undertake, you will actually get less productivity, because in order to achieve re-use you have to make your components more robust,' he explains.

He adds: 'You will either be disappointed, or if you have been realistic to start with, then you will be gratified by the idea of some extra benefits.

It is only in the third project that you will start to reap the benefit, however.'

Alan Ratcliffe, director of development at Citymax, a consultancy and products subsidiary of Credit Suisse, appreciates the need for a long-term view, but also believes that maintaining an existing investment is important.

His answer is to parcel up older code while keeping one eye on the future.

'(JBA) is coming from a more focused, short term approach,' he says. 'The advantage of Corba standards is the openness. A lot of projects we see are taking information from a variety of sources. We need something that will give us an open standard. We are trying to comply with standards that have a long-term benefit for the customer.'

Nevertheless, Ratcliffe is often asked to give users a re-usable system with a minimal outlay.

'We encapsulate legacy code by writing some C++ code over the top,' he says.

The main difference between JBA's approach and Ratcliffe's is that Ratcliffe appears to have more faith in standards. The essential principle is the same, however.

Users have been held back from taking the object-oriented route by two things: a lack of foresight which prevents them from seeing the benefits of re-use; and an understandable reluctance to make the necessary initial outlay.

Only the users can solve the first problem, by taking a longer term view of their IT strategy. The second problem can be solved by chucking out the bath water but keeping the baby. It is now clear that altering your development methodology does not mean that you have to lose your old code.