The former head of IT at Bovis, explains how interoperable computing systems could help construction teams speak the same language.
Sir John Egan told construction to take lessons in lean production from car manufacturers. But it could also follow its lead in 3D modelling.

In the motor industry, firms such as Ford and General Motors produce their concept designs in highly specialised – and highly expensive – 3D modelling systems such as CATIA and ProEngineer. Their suppliers then design the components in the same system. The component designs are passed to the manufacturers, who fit them into their overall models, and finally those computerised models can be tested in computerised test beds, wind tunnels and so on. Electronic prototyping, as it were.

By using the prototype to simulate the manufacturing process, the models can also be used to design production lines or even entire assembly plants. The design data contained in the model is then passed to computerised equipment that can press, cut, mould and assemble the final products.

It doesn’t stop there. By providing some of the model information to their dealers and other outlets, the manufacturers can give potential customers the chance to choose their own combination of configuration, components and colours. This is priced, then the order is taken and fed back to the manufacturing plant to be scheduled into the appropriate production run.

Wouldn’t it be nice to do buildings like that? Twenty years ago, we had the beginnings of that capability here in the UK. Systems such as GDS, Acropolis and RUCAPS – true CAD programs – all enabled, required even, the design to be modelled and analysed in 3D.

The models were also intelligent to a certain extent; a door “knew” it was a door, with a frame set, hinges and a lintel. Move it around in a wall and the wall “knew” how to respond.

The basics of integrated, model-based, clash-free design were all available in the UK in the 1970s, thanks to the efforts of a few visionary designers who knew how to apply computing power to the subtleties of the design process. But these systems were expensive and difficult to operate.

Then along came systems such as AutoCAD that blew them out of the water. AutoCAD made producing drawings cheap, easy and efficient, so firms throughout the industry invested heavily in these computer-aided drafting packages.

Everyone on a project should use the same software package in the same way – just as car manufacturers do

Ray Crotty

Unfortunately, in the process we lost sight of the fundamentals. It is the ideas that count, rather than the drawings. What is important is the information embodied in the drawings. The drawings themselves are simply communication devices, a means of ordering ideas and of sharing them with others.

Whereas people can infer nuance and deduce rich meaning from drawings, we live in an age of computers which are quite incapable of this sort of inference and deduction. To exchange information, everything must be expressed in an unambiguous language that the receiving computer fully understands. With a few notable exceptions, effective information sharing between firms is almost impossible, even if they are both using the same system.

There are two ways to achieve effective sharing of CAD information. The first is for everyone (at least, everyone on a given project) to use the same software package in exactly the same way, just as the car manufacturers do. This would not be easy to achieve in construction. The alternative is for someone to invent a common language and common protocols for different systems. This is the raison d’être of the International Alliance for Interoperability: to enable diverse systems to interoperate, seamlessly exchanging data.

It faces a Herculean task. The IAI has to find ways of fully specifying, in data terms, every conceivable component that could be used in a building. To succeed, it will also have to persuade every CAD vendor in the industry to conform to its standards – and software vendors are hardly renowned for their willingness to share data or to put the users’ interests before their own.

A more likely outcome would be for a software manufacturer to develop a new, easy-to-use, reasonably priced CAD system. It is surprising that this much-needed package still does not exist.

I expect such a system to become commonplace within the next five years. Although perhaps not fully interoperable, it should support fundamental data standards that will enable specialist programs to translate between them. An architect using such a system will be able to carry out most of the design of a building, including the visualisation work, on one model, thus reducing the need for interoperability between systems. The result will be that architects and clients will be able to see the exact outcome of their design.