Gradnja novega korpusa slovenščine

The article presents the initial work in the construction of a new reference corpus for Slovene. This will be an upgrade of the FidaPLUS corpus and will be divided into two parts: the first containing one hundred million tokens and the second containing the difference up to a billion tokens. The taxonomy of the corpus is presented, with the approximate division of various types of text, along with a brief explanation of the principles guiding this taxonomy. Certain other key principles that will determine selection are also listed. The selection of texts on the basis of various data, from which it is possible to at least approximately make inferences about the reception and production of the publicly published Slovene texts, is still in progress.