One other issue of utmost importance is the support of different file formats. In modern translation business it is difficult to imagine working on text files only. Translators are no longer surprised to receive spreadsheet or power point files to be returned translated in the original format. The modern CAT tools usually support a number of data formats including not only the above-mentioned ones but also database files, help contents, a number of DTP formats, java properties, html, sgml, xml, and other. SDLX 2004 also offers a solution for the formats unsupported by the tool itself. It is possible to copy the text to the clipboard and translate the clipboard contents in the program.
One important development is that there are no limits to the number of termbases which can be created by the user. The only limitation can be the size of the hard disc. It is often possible to consult a number of termbases at a time, however the number is not great - about five termbases. Usually also the lookup order can be set up by the user. Furthermore, the number of languages of the termbase can be greater than just one source and one target language. The tools subject to testing within this project offer a possibility of selecting at least twenty languages for a single termbase. The users may switch the target and source languages freely and sort the terms according to a particular language selected as source.
Terminology management tools now offer a number of statistical data including termbase size, the number of entries and the number of perfect and fuzzy matches. However most statistical analysis is carried out by the translation memory module and therefore should not be included in the description devoted primarily to terminology management modules. What these modules do offer, are basic quality management functionalities, such as warning about duplicate entries, project status, spellchecking11 and terminology validation. In one of the programs tested it is possible to restrict the access to the project or termbase by setting up a password. Another operation which can definitely improve the functioning of the termbase is data compression. Most tools also record the time and name of the user who introduced any changes in the termbase.
Unfortunately none of the tools presented in this thesis creates backup files automatically. Nonetheless, all of them support long filenames, which was not the case with the earlier versions. Years of gaining experience and perfecting the tools subject to study in this project make them stable and reliable. However, problems may occur with each of them and the tools signal them to the user by showing error messages.
3.6 Entry model and structure
There are three possible entry models: free, fixed or quasi-free. The difference between them is primarily in what margin for manipulation is left to the user. In fixed record structure the user can only select one of the available record templates without the possibility to modify it. Using this record structure we may decide to only to omit certain fields. The scope of manipulation is greater in the case of quasi-free record structure, which can be defined as a fixed record structure, with one or several user-definable fields. Finally, the free record structure means that the record structure can be defined by the user entirely, including also field naming and arrangement. When designing a free record structure, the users are usually allowed to create picklists of values for the different fields. Most commonly these include subject, customer, project, as well as grammatical information or usage labels. Unfortunately different tools use different field naming conventions, as do large termbanks.