A text which has been dictated is typed by a secretary. All secretaries use WordPerfect, which was initially selected for its ability to support multilingual character sets and maximize document interchange compatibility with the Community institutions. They also use the standard facilities provided with text-processors such as style-sheets, macros and spelling-checkers considerably more than do their translator colleagues.
Secretaries are also asked to type from handwritten input. This is usually a consequence of corrections to draft translations by the translators who have dictated their work, or of corrections made in the course of revision. Inevitably, there are people whose handwriting renders this task difficult.
A problem with keeping secretarial staff was frequently mentioned. The SdT is the most typical first destination for a newly recruited secretary, partly because of the high number of secretarial staff needed by the SdT and partly because of frequent departures. Working with the SdT does have some advantages, primarily those of working with fellow-nationals and of comparatively flexible working arrangements in some cases. Some people too, for a variety of reasons, are happy to have work which does not vary a great deal and which does not imply a heavy load of responsibility. However, audio and copy-typing is not a very interesting occupation, and career-oriented staff tend to leave for posts with more varied work as soon as they can.
It was suggested that secretarial staff could do part of the preparation of a text: in certain units, for example, secretarial staff already carry out preprocessing of repetitive texts so that what is passed to the translator is a text where standard repetitive phrases are already replaced by the target equivalent. This technique is being used with considerable success for some texts, of which the Bulletin is one noticeable example. The degree to which preprocessing can be used to lighten the burden of translation or of typing depends on the identification of suitable types of text.
Although one of the questions put during the interviews aimed at getting an impression of how much repetition those working on them perceived in the texts they dealt with, it proved impossible to formulate any clear opinion. This perhaps can be partially explained by the fact that repetition can occur at any level of a text. There are the obvious clear cases where large stretches of text are repeated, but, at the other extreme, there may be heavy repetition of quite small text elements, which is harder to spot. Also, there is quite a strong difference between literal repetition, where a large or small segment of text is repeated in exactly the same form, and what one might term intuitive repetition, where repeated segments are almost, but not quite the same. Literal repetition lends itself to computer processing in a very straightforward way, where intuitive repetition may imply recourse to more sophisticated techniques. Nonetheless, for the translator, both kinds of repetition feel very similar, and he may be disappointed if a tool which deals very well with text exhibiting a high level of literal repetition fails to perform equally well on texts with a high level of intuitive repetition. This in turn leads to the reflection that there may be a place for the intervention, at least in the early stages of trying to identify suitable texts, of advisers specializing in the processing of repetitive text (Language Help Desk) and software tools designed to confirm intuitive judgements of text repetition (for which TMan is, in effect, a pilot). The experience acquired through the new Translation Workshop (q.v.) may also help with this.
Some units make use of freelance typists. This can prolong the work-chain: there are cases where a text is sent out for initial typing, comes back for proof-reading, is sent out again for correction, comes back again for revision by the translator, goes out again, comes back for correction, goes back again to the freelance, and so on. Freelance typists return the final version of their work on diskettes, but do not always return early versions which are to be corrected in electronic form. One of those interviewed pointed out that this means that the time taken over the whole chain of typing and correcting can become unnecessarily long: if the typed version of the first draft were returned on diskette, the translator's and reviser's corrections could be done directly on the electronic version and one or more steps in the chain avoided.
A number of those interviewed also remarked that the typed final version tended to arrive on the last minute, so that there was little or no time to do final corrections. This was found frustrating, especially since there was no guarantee that the freelance typist had succeeded in deciphering the translator's or reviser's handwriting correctly. Handwriting can sometimes be appalling, and when typing is done by a freelance, the author of the correction is not easily available for clarification.
Whether a translation is prepared directly in electronic form by the translator or whether it is typed into a text-processor by a secretary or freelance, the electronic file is periodically sent to a central archive, where it becomes available for consultation.