Pair or solo programming. Which is more productive

[rageh]

Which is more profitable; Pair programming or going solo?

http://agilesoftwaredevelopment.com/...lo-programming

I came across this article in the link above and it made me think specially whether pair programming is more productive and thus profitable than solo programming. Obviously when you have two programmers, you have to pay double. Is productivity doubled as well? In what scenario other than bug-fixing is pair programming better than solo? Can we say that there are more distractions when pair programming because the two people chat more and may take things easy than a solo programmer? What do you think?

Thanks for your input.

[Mittineague]

I guess it depends on the pair.
Do they replicate each others work?
Do their strengths and weaknesses compliment each other?
Are their responsibilities well defined and can they communicate with each other?

[rageh]

Pair programming is obviously more productive. I tried to edit the title of the thread with no success. It is meant to read which one is preferable.

When I said pair, I mean a pair that have clearly defined responsibilities. They will be more productive of course, but can they be 100% more productive than a solo programmer?

Put on your employer hat for a minute and look into the pros and cons of pair vs solo programming. In the absence of a deadline pressure, can you justify paying two people while one person working solo can in the end finish the project?

[Mittineague]

From a purely business perspective, I guess you would need to consider the labor costs (including benefit packages and taxes). I don't see where it would make much difference if time isn't a factor.

[crmalibu]

Hard to say. Specialization often yields efficiency improvements, so if the two are specialized in different areas, they can work in their most efficient mode. Inefficiency will come from when one will invariably finish before the other, and be forced to switch to tasks they're less efficient at.

Something that's maybe hard to put a metric value on though is I think you will almost certainly get a better software product out of 2 coders specializing in different aspects. Even two coders specialized at the same task, are likely to think a bit different, and offer valuable objective analysis of each others coding, improving it.

There's a lot of variables at play here, I wouldn't be able to come to a clear decision.