Todays project leaders are unsettled.
Should they use programmer pairs or should they rely on single
developers for the next project?
So far, empirical software engineering research gives no clear
advice. One study reports on a doubled personnel cost when using
programmer pairs instead of single developers while another
study reports on a substantial increase of development speed and
But, what would be if project leaders add a separate review
phase to single programmers' development?
Then things change in terms of development cost and program
quality. Two experiments conducted at the Universität Karlsruhe
revealed that there is almost no difference in cost if both
programmer pairs and single developers with reviews are forced
to develop programs with comparable quality.