The need for system and product assessment has multiple aspects. The most obvious one is the need that service providers communicate to system manufacturers, as they want to have a clear idea of the benefit that they can expect from some particular technology in the context of the product that they intend to develop and supply. A first question is therefore to define and measure appropriate quantities from which a service provider will be able to predict with reasonable accuracy the consequences of the new technology for its service.
Some form of assessment is also necessary to inform and protect end users. Let us illustrate our argument with the extreme example of a hypothetical ``lie detector'', a system which would supposedly be capable of detecting when a speaker is purposely not telling the truth. Should such systems be implemented and sold as commercial products without any objective evaluation of these products having been conducted?
For some products, a more or less independent standardisation institute has to deliver formal approval prior to commercial exploitation. For others, the product has to fulfil a certain number of specifications before it can be sold. At the moment there is no such procedure concerning speaker recognition technology.
It is part of the responsibility of the scientific community to provide tools and methods in order to assess the performance of systems.
The rest of this section lists the multiple aspects that have to be taken into account. However, no commonly agreed standard exists today concerning the relevant quantities to measure and a fortiori the methods for measuring them.
Guidelines and recommendations should also be made about the feasibility of applications: it may happen that the general public is not able to tell easily whether a given product is realistic or whether it remains science fiction. While scientists should be aware of new needs, they should also keep the general public permanently informed of the progress and the limitations of our current knowledge.