Or is it extraordinaly (pun intended)? You have to wonder!
Indeed, his latest one posted to the arXiv has, among the frontpage footnotes, “Publication 1023”! There is (at least) that amount on his archive (although the 1000th is marked as ‘SAVED’, presumably for something especially noteworthy), and 630 of them are on the arXiv (as of today…).
He apparently is 68, and received his PhD in 1969 (the year he started publishing), so that’s 1023/(2013-1969)=23.25 papers per year, for 44 years! For comparison, other prolific recent mathematicians include Erdős (about 1525 papers in about 67 years) and Tao (about 250 papers and 17 books in 17 years, according to wikipedia; so just for papers he is on track for 1000 in about 68 years).
Back to Shelah, and it seems that many of his papers are moreover important, with one in JAMS just in the past couple years, while his older works have been recognised many times, including by the 2013 ‘Seminal Contribution to Research’ Leroy P. Steele prize. Slightly more information on his work pre-XXIth century is on the MacTutor, but a self-contained text (at the level of a graduate student or non-specialist mathematician) giving a moderately technical overview of his main achivements would be very welcome, I’d imagine (the slides for his Plenary Talk at the 2012 ECM are interesting, for instance).