The identification of anomalous code measures with conditioned interval metrics Carlos López 1 , Esperanza Manso 2 , Yania Crespo 2 , 1 Área de Lenguajes y Sistemas Informáticos, Universidad de Burgos EPS Edf. C. C/Francisco de Vitoria S/N 09006 Burgos, Spain. clopezno@ubu.es 2 Departamento de Informática, Universidad de Valladolid ETS de Ingeniería Informática. Campus Miguel Delíbes 47011 Valladolid, Spain. {manso,yania}@infor.uva.es Abstract. Anomalous measurements are identified in the software measurement process using valid metrics intervals. In the particular case of code measurements, the same intervals are used independently of the nature of the problem solved by the entity being measured. Our proposal is to condition the measurement intervals according to the nature of the problem solved by the said code entity. By ‘nature’ we understand that which is expressed through standard UML classifier stereotypes. This paper identifies the requirements needed for a code measurement support tool to be able to take on this new perspective. Using these requirements as a basis, some existing tools are reviewed and the difficulty of applying this proposal with its current functionality is recognized. To this end, we present the adaptation of one of the reviewed tools (RefactorIt) and, in addition, the measurement process is applied to ten real projects, obtaining some initial intervals conditioned by the nature of the code entities. Key words: Code metrics, Use intervals, Code measurement tools, Measurement process. 1 Introduction Since the 1990s, software metrics and their associated measurement process, have attracted great interest in the software engineering community as a means of quantifying and controlling software quality [1] [2] [3]. According to [4], measuring is part of a process (see Figure 1) which consists of obtaining a numerical value for an attribute of a software product or process.