The Text Encoding Initiative will officially release the P5 edition of its guidelines at the end of October 2007. In this context I will provide an overview of what the guidelines can actually provide a scholar with, in the domain of text representation and annotation. I will present both the global architecture and principles of the guidelines, with some insights into specific mechanisms and tagsets tat are available. I will finally focus on the capacity that the TEI framework offers to customize the guidelines in the context of specific encoding projects.