Describe how Literal Types are Resolved

The description of the CSV extractor shows a nice example how a trivial data set is converted. In the description I'm missing how the rdf:dataType is derived for literals. For example the cell 30 is an integer but how can the extractor know this? Is there some kind of classifyer? Please describe that in the docs.

Also a hint: The mailing list link on the web page is dead.