> The people in this facility measure, as a matter of accepted practice, the output of individual programmers and programming groups and have, over the years, developed rather accurate techniques for estimating the time and resources that a project requires. Therefore, any substantial deviation from one of their estimates can be considered a true variance and valid comparisons of group performance are possible.
Looks like no. All the citations generally go to OP. It appears to have originally circulated as "IBM technical report GE20-0752-0"; one book in Google Books notes it as 'available from the author at IBM'. There might be an IBM website somewhere which has it, but if so, Google doesn't know about it, and the one IBM research database I checked didn't turn up anything for that ID.
Seriously, that's the paper we should be reading!