::scr Extreme Programming's Pair Programming

Simon Batistoni scr@thegestalt.org
Mon, 1 Oct 2001 12:52:22 +0100


On 01/10/01 12:36 +0100, David Cantrell wrote:
> XP advocates would, I believe, claim that it doesn't double cost, as you
> get more than one person's worth of work out of each pair.  However, I
> remain to be convinced that you would get *two* peoples' worth of work
> out of the pair.  I imagine that the quality would be substantially
> improved, but yeah, not by enough.

I think there's a lot to be said for collaboration. For example, I
find there's nothing more valuable than having someone else cast
their eyes over a piece of code that's been turning you cross-eyed
for half an hour. They'll often find the problem in seconds.

If you set up a good working partnership between pairs, where they
work separately, but collaborate on difficulties, or on big
decisions about the best way to implement a feature, I think you can
maybe hope to get 2.5 peoples' worth of work out of 2. You end up
with happier programmers, too, since they spend less time smashing
their brains on the keyboard.

But a pair of people doing one person's job would have to be going
some to double the speed of completion, and I doubt they'd come up
with many quality improvements that couldn't be handled by a
periodic and less time-consuming peer-review process.

By the by, I also think XP is absolutely doomed in the current
financial climate. Unless you've got a manager who's hooked on the
cult, and all his superiors up the chain of the command are
similarly infected, someone's going to worry that you're spending
2xsalary for 1xproductivity, and nix the idea :/

-- 

Simon Batistoni            Penseroso Ltd
----------------------------------------
simon@penseroso.com     +44 20 7242 0570
----------------------------------------