Pair programming: There is no silver bullet

19 Sep 2004

java.net Weblogs: Pair programming:
Everyone’s favorite argument

Just finished reading an interesting short piece and comments debating pair programming.

In my working environment, a smallish biotech software company, I belive there potetially could be a place for pair programming. However, I do not think
that everyone is equiped to participate in a pairing and its important to take this into consideration. There are likely benefits to pair programming in
terms of improved code quality and the followup efficiciencies. However, given that these benefits are not gauranteed, is it worth the effort? I don’t co
nsider pair programming to be a replacement for a code review.

My ideal situation for pair programming involves the pairing of a senior person with someone more junior. However, given the likelyhood of not having a di
rect 1:1 mapping of senior to junior, should you ever consider pairing a junior with a junior or a senior with a senior? Is it fair to a senior developer
to force him to dedicate time away from his tasks to work with others?

I agree in the need for communication between programmers. Does it follow that developers should work hand-in-hand, I’m not convinced. Programmers shoul
d be encouraged to ask for help when required, but not start depending or relying on someone else to solve a potentially difficult problem.

In short, there are no magic silver bullets, regardless of the appropriateness or inappropriateness of pair programming and XP.