This presentation intends to display some of the main conclusions of the thesis entitled Terminology extraction: elements for the construction of SEACUSE (System for Automatic Terminology Extraction). The article discusses four points. First, the results of the analysis of the main existing terminology extractors are presented. Secondly and thirdly, the causes of the limitations of these extractors are argued. Finally, two of the parameters are presented that are considered to improve terminology extractors: the magnifying of the extraction object and the functional point of view of the extraction.