Sorry to say I find the alignment of ArchiMate to TOGAF in this paper is too superficial. They use similar words (because we have a limited vocabulary of workds like system, function, service, interface) but the meanings given to these words are not consistent within TOGAF and not consistent between TOGAF and Archimate. 
 
At the level of the underlying concepts, of the underlying meta models, the ArchiMate meta model appears to me to be significantly different from the TOGAF meta model, and that is level the work needs to be done.

For an extensive discussion between Graham and me on the details of mapping TOGAF to ArchiMate and vice versa, see the following series of articles on this site: 
 
Using ArchiMate with an Architecture Method, http://www.via-nova-architectura.org/magazine/magazine/using-archimate-with-an-architecture-method.html 
 
Using ArchiMate with TOGAF – Part 1: Answers to nine general questions about methods, http://www.via-nova-architectura.org/artikelen/tijdschrift/using-archimate-with-togaf.html 
 
Structural and behavioural views – and the ambiguity of “Function”, http://www.via-nova-architectura.org/artikelen/tijdschrift/structural-and-behavioural-views-and-the-ambiguity-of-function.html 
 
Using ArchiMate with TOGAF (2): http://www.via-nova-architectura.org/artikelen/tijdschrift/using-archimate-with-togaf-2.html

And if you are interested in the combination of ArchiMate and TOGAF, also check the following two Open Group whitepapers: 
http://www.opengroup.org/bookstore/catalog/w191.htm 
http://www.opengroup.org/bookstore/catalog/w192.htm