Thanks to David Megginson, who implemented the original Java version of SAX, which provided such a good place to start this project.
John Cowan spent a good long time developing TagSoup, an HTML parser with a SAX interface. TagSoup parses HTML as you find it in the wild, balancing tags and restarted elements as appropriate to give a well-formed output that you can process with normal XML tools. I took only a few hours to port his code directly to C++ for Arabica, and he was extremely gracious and encouraging about it.
Thanks also to the many people who've used Arabica over the years, particularly those who've emailed thanks or patches or bug reports. They're all much appreciated. I should also mention a chap called Damon, who seemed to completely hate Arabica, because it prompted me to move the build to Autotools.
Finally, thanks to Andreas Grimm, of AMS in Düsseldorf, who showed me what an XML parser interface absolutely shouldn't look like and thus inadvertently kicked this whole project off.