Development of A Digital Corpus of Portuguese Didactical Texts For Language Research

The Corpus de Manuais Escolares consists of 2000 texts from Portuguese schoolbooks, annotated for genre pedagogy research. The extant form of the corpus is a tree of directories and files maintained entirely by hand. We are currently revising this form into one more suitable for computational processing, publication, and evolution. This work has included computational processing itself, with programs written in an incremental way to support the necessary reverse engineering and the envisioned re-engineering. Here we describe the overall process, the programs developed, and some new corpus data already found.