A request for translation will go either to the central planning, in the case of multithematic documents, or to the department associated with the particular requester. In either case, the first operation is to see how the request fits in with planning. A first problem here is that requests come from a variety of sources, each of which has its own priorities, which may well be in conflict with those of other requesters. There is no central policy about how to decide conflicts: it is left to the person responsible for planning to make decisions based on experience and on general knowledge of what the current important issues are.
The decisions taken may well have an impact on work which has already been accepted and is waiting for translation: an urgent document coming in may push less urgent material down the queue. Not surprisingly, this can sometimes lead to strained relations with the requesters. If someone sends a well-prepared request for translation in plenty of time, it is very hard for him to understand that the deadline still may not be met because of other work coming in.
This problem is much worse for certain departments than for others, especially for those dealing with a lot of political documents or with privileged requesters, such as the Secretariat-General, whose requests have to be given very high priority.
Exactly how the planning is organised varies from department to department. In some cases a document will first be examined to see what reference materials are required and whether they are available, whether parts have been translated before, and whether modifications have been marked. In other cases, the assistants responsible for planning will do this work themselves, and in yet others, some of this work will get passed down the chain to the translation units or even, occasionally, to the individual translator. Sometimes this may happen because those earlier in the chain do not regard it as part of their responsibilities or as lying within their competence. In other cases it may be unintentional: it is not always possible, for example, to determine what all the pertinent reference material may be without subjecting the text to the sort of close examination it will only receive once it is being translated. A very negative side effect of such work being passed down the chain is that the same requester may be contacted by several people (the translators from the different units) to ask the same question or request the same reference document. Clearly, this does nothing to improve either relations with the requesters or efficiency.
One extreme case of the difficulty of finding reference material will help to illustrate its importance. One department deals fairly frequently with ``Common Position'' documents, documents of only two or three pages which report on the position of all the different Institutions. Consequently, a single document may contain fragments of texts from all the different Institutions, and the exact wording must be reproduced. One case was cited where a document of this type had almost sixty reference documents associated with it. It can take a considerable time just to find the references, and perhaps involve requesting help from the Parliament translators who may already have done the work of collecting the references together.
It was suggested during the interviews that it might also be possible to do some work on terminology when a document first arrived in the Department, rather than leaving it to the individual translation units.
Where documents come in successive versions, some groups use document comparison software to determine what modifications have been made. Others rely on visual scanning. Exasperation with requesters who fail to mark modifications was frequently expressed. It is possible too for mistakes to intervene at this stage. For example, imagine that a document has previously been translated and sent back to the requester. The requester then changes a figure somewhere in the document. Subsequently, a whole paragraph of text is modified, and the requester sends this back to the translation service marking the paragraph, but not the figure. The translation service then works on the first version of the translation, which is still available electronically and sends back a new translation. The modified figure has now been changed back, or rather has never effectively been changed. There is a high probability that the requester will not notice.
Version control and avoiding doing the same work more than once are a major preoccupation for almost everybody who was interviewed. The problem is of course compounded by the sheer size of the organisation. Many people emphasized the importance of good electronic archiving facilities, and the possibility of being able to carry out a key-word search over large archives of documents was one of the most frequently mentioned dream-tools. A Word-Perfect facility of this type, QUICKFINDER, does exist and is used by some people, but its usefulness is limited by inability to handle the complete (and very large) archives in the SdT's networking environment. The issue of access to the Official Journal versions of texts, previously mentioned, is also obviously pertinent here. The planned document server (vide Document server, section 2.4.6) is designed to mitigate this problem.
As might be expected, those responsible for planning make heavy use of SUIVI.