Manichean, here are some example file names:
Chernick, et al 2011 - The Impact of the Great Recession and the Housing Crisis on the Financing of America's Largest Cities.pdf
Dalgaard 2008 - Introductory Statistics with R.pdf
Shrader-Frechette 2001 - MacIntyre on Human Rights.pdf
Theus & Lauer 1999 - Visualizing Loglinear Models.pdf
Verzani - SimpleR.pdf
1766_1103_CLT Reader Part I.pdf
Urban Stud-2007-Butler-1161-74.pdf
First title uses "et al". Subsequent ones include authors with hyphenated names, joint authors, and filenames with no publication date. The last two indicate that there's some stuff in my library which has some weird filename which probably came from the website I downloaded from.
I tested
Code:
(?P<author>[^_]+) (?P<published>\d\d\d\d) - (?P<title>.+)
using 0.8.1 and the published date came in nicely (though set to current month-day, e.g. 5/16, of the specified publication year, not a big deal). Authors with et al get an author_sort field of "al Author, et".
Thanks,
--Lee