A Reply from Ovum

Last time we got it wrong — as Tony Whitecombe rightly re members (“Did Ovum Get it Right?”, March-April). In 1985, even though Tim Johnson discounted some of the more optimistic claims of the nl researchers of the time, he still predicted much more rapid growth in nl markets than has actually materialized.
    We now know that nl processing is really a very long-term business; it takes not years but decades for promising research activities to grow into commercial products. That is why, the second time around, in Natural Language Markets: Commercial Strategies, we have concentrated on products which already exist, or, in at least one case, have reached a beta testing stage. The fact is that, apart from Machine Translation systems, these are all English-language products. Interest and goodwill clearly exist in continental European countries, too, but the markets for each individual language are so much smaller that potential developers hesitate to make the massive investment necessary to bring an nl product to market. We predict, therefore, that European language products will follow successful English-language products. The market is there, but it will be realized much later.
    Machine Translation is the exception. There are simply many more foreign language speakers in continental European countries than in the us or the uk. The CEC’s 1988 Language Industry Survey showed that for every ten persons involved in foreign language activities (reading, writing, translating) on the continent there was only one in the UK. The US situation is very similar. us companies have realized the need for documentation translation, but they get the work done in Europe where the translators who are needed to get the job done live. The major research programs are in Europe, too, including the CEC’s Eurotra project, which is described in the report, on page 97.
    Because of the long-term nature of nl product development, the very laudable CEC NL initiative of the last few years still has a long way to go before it will translate into commercial products. Its funding will contribute to making the nl market happen in Europe, but not within the next few years.
    Concerning talkwriters, there is solid practical experience of talkwriter use by non-handicapped people, although so far only for special-purpose systems, such as for medical reporting and law, and despite the obvious shortcomings of present-day systems. Field tests are underway in other industries and these include French, German, and Italian systems. Once affordable continuous-speech systems become available, today’s two-finger typists will find them difficult to resist!