Text::CSV_XS is written (mostly) in XS code: that is the C API to perl and can thus be optimized at a very low level. Text::CSV is a wrapper over Text::CSV_XS and Text::CSV_PP, the Pure-Perl implementation that does exactly the same, but up to a 100 time slower. The documentation should be identical. The wrapper will automatically choose Text::CSV_XS when it is installed and available.
If you are now using Text::CSV and you install Text::CSV_XS the code doesn't have to change at all but will immediately gain speed.
Where are the "blanks"? I think that by now the documentation is quite elaborate, and comes with clear examples.
Enjoy, Have FUN! H.Merijn