The latter of these is probably the most important trend to watch. The need to communicate across languages is expanding fast across all types of businesses. We estimate that by the end of the decade a quarter of spending on localisation and translation (on what we call Globalisation services) will be for business information and services sector companies.

This raises the question of how the localisation industry will grow and adapt to new types of customers, beyond the now quite sophisticated customer base in the software sector.

This will be most efficiently done by vendors - of both products and services - who understand the role of information manufacturing within larger business processes.

The term information manufacturing refers to a set of processes which make it possible for companies to provide their customers with the information that supports products - or in some cases is the product. Customers have begun to realise that this information is an important corporate resource that must be managed, just like the processes for developing, making and selling a product are managed.

Core skills in manufacturing information for a global market are illustrated in Figure 1 (unavailable). Of the four key areas we identify here, two are the natural domain of the customer - product engineering and maintaining a repository of the text and documents which support the product. The other two are the natural domain, at least in the current environment, of localisation and translation vendors. These include the management of the localisation process, and the provision of language engineering skills (increasingly based on technology) that go hand in hand with the translation function.

As the figure illustrates, localisation functions are embedded within the mainstream processes of customers. This has given rise to the strong "partnership" model in outsourcing such services.

Many localisation vendors have concentrated on their partnerships with software companies. They have developed an understanding of software engineering, and documentation requirements for software products. As other sectors become important customers for globalisation services, vendors may find their focus shifting somewhat towards complementing other core skills.

Localisation in industrial manufacturing, for example, requires knowledge of completely different engineering conventions and processes. This is more than a difference of "domain terminology" - though this is, of course, vital. Vendors must be able to adapt to the specific development and manufacturing cycles of different types of customers. Acquiring a command of "domain expertise" in this wider sense will be a key differentiator of service companies.

There is one respect in which companies outside the software sector are actually more advanced than mainstream software publishers. This is their consciousness of the need for discipline and management in the creation of documentation as well as in its translation. There is a somewhat more urgent need for industry- wide standardisation, not least for reasons of safety and regulation, which in turn has given rise to controlled authoring of technical texts.

I believe the differences in the needs of customers in the industrial sector will put them on a quite different trajectory from software customers. They will focus more attention on engineering the information they publish, developing robust systems for manufacturing information that go hand in hand with their product manufacturing systems. This presents two challenges for the localisation industry:

• being able to complement true information manufacturing - e.g. slotting into a rigorously controlled workflow
• retaining a significant share of the added value in the service chain - i.e. being more than a body-shop for human translators.

The latter is a particularly notable risk, as customers who build effective systems to manufacture text into many languages, delivered in multiple formats, may outsource only those parts of the process which are unpredictable, or labour intensive.

Multilingual communication continues to expand. You can now play multilingual monopoly across a network, courtesy of Hasbro. As more people want -and need - to communicate across languages the market for translation in the service sector will be unprecedented. This is an entirely new area for localisation vendors, and one where the use of technology will be essential.