There are so many interesting things that can be discovered here (if
not enlightenment) even without a citation to the paper that Casey
read. I can't resist making the following observation, though the only
link back to Smalltalk is that marketers once tried to sell both
Smalltalk and Java as silver bullets.

The first thing that comes to mind is that the Java developers -
particularly the bosses - probably need to learn the lessons of "No
Silver Bullet" and "The Mythical Man-Month". Though these short pieces
weren't about BCPL or COBOL, they were written in the heyday of those
two languages, and may be sufficient to explain a high rate of defects
per capita.

David (who wonders if he should apply for a research grant)