Sentiment is an awesome space to be in, the linked article is a nice overview but doesn't exactly break any novel revelations.

If you're interested in learning about sentiment and the machine learning behind it, two of the early 'seminal' papers (somewhat outdated by current techniques, but really understandable) are Turney 2002 (http://acl.ldc.upenn.edu/acl2002/MAIN/pdfs/Main425.pdf) and Pang and Lee 2002 (http://portal.acm.org/citation.cfm?id=1118693.1118704). Wiebe's work is also worth a look (she's quoted in the OP).

Good places to start, at least; ping me if you want some more.