In this article we discuss the problem of annotation of historical languages for which less, up to no resources are available, and which do not follow the standard paradigm of Indo-European languages. We show that the development of a tool adequate to the data model (and not the adaptation of data to given tools) has to be considered in particular cases.
