There's a debate going on over at the Scientist about the thorny issue of self plagiarism, and when/ if it constitutes poor practice. I used to work for someone who started every paper, research or review, with the same short paragraph and we could all recite it like a mantra - and by general agreement it was felt to be the best, most succinct way to introduce the topic, which noone else has yet bettered.
Of course self plagiarism is not to be confused with duplicate publication, where there is clear evidence of results, for example, being reported more than once without appropriate reason or reference.
The original article by Jef Akst and the comments are worth a look - including one suggestion that self-plagiarism as a term is just not used any more because of confusion with the serious issue of (non-self) plagiarism (which COPE defines as "When somebody presents the work of others (data, words or theories) as if they were his/her own and without proper acknowledgment"). A case of self plagiarism which was brought to the COPE forum in 2009 is discussed here.