ABSTRACT
The real test of machine translation is whether or not it is effective in large-scale operations. These may be specialized applications, as in the case of Canada’s METEO (see Chapter 3), or they may involve the translation of a broad variety of text types. In the latter case, the purpose of the translation will dictate the characteristics of the installation, particularly the human post-editing component. The purpose can run the gamut from publication for dissemination to ‘information only’. For the product that has a public and undergoes scrutiny, a translator post-editor must be enlisted to eliminate all problems from the output, interpret connotations appropriately and make certain that the reader will understand precisely what the author intended to say. On the other hand, MT, by automatically generating a product where nothing existed before, has created a new mode of work in which a less than ideal translation can be provided for the consumer who merely needs to assimilate information and is not concerned with disseminating it.
