As you may imagine, we work both with unilingual and bilingual dictionaries. I am responsible for the French bilingual dictionaries (and now the Swedish, Hungarian, Bulgarian, and yet to come Russian dictionaries). But when deadlines approach for a given project, we all work together. Unilingual and bilingual are obviously very different. Unilinguals deal more with definitions, and Bilinguals with translations.

When I first came here, we had a data base on Access, with all the entries, and information about each one of them. So, for instance, if we wanted to treat all numerals, we could make a query and have all numerals and treat them in the same way (the same kind of structure, etc.). It's really good when we want to find discrepancies, typos, etc. and want to uniform. The problem, I think, was memory. Access "eats" memory, and our computers blocked quite often.

Then, we had this dictionary editor, created within our company. We were lucky to have a computer engineer who loves linguistics. His thesis was on a parser, so you see he is both computer science and linguistics freak.

We teach our contributors and authors how to use the editor (it's part of my work, actually). With this editor, they don't need to worry about bolds, italics, fonts, size fonts, etc., if they put the information in the appropriate field, we can treat it the way we want. With Word that would not be possible. And later on, if we want to change the graphic layout of a dictionary, we just need to make a design study, and then apply the new styles to the editor. When the work goes to the Pre-Print department, everything is already done, and the process go really fast. Twenty years ago, a dictionary would take ages to be done!

We do need a corpora. We have been talking about it for ages, but somewhat, there are more important things to be done before.

Now the idea will be to use Oracle all over the Company.. and that will be another story.