My colleagues and I tend to apply pair programming regularly, but only when it's useful, and as long as it is useful. It helps us tackle hard or complex problems (and we only use it for that purpose). The way we apply it, it's also definitely 'pair programming' and not just 'pair coding', since we fairly often briefly move to a meeting room to talk over the problem and possible sketch it out.
Do you follow through on the problem as a pair completely to the end? That is, after the meeting room do you sit at a computer (just one) and code the solution as a pair?
I'm asking since I've regularly had programming meetings in pairs, but we've almost always broken up after that meeting. On rare occassions I've had somebody else help me with code, but certainly nothing formal.
That mostly depends on the problem and on the outcome/conclusions of the impromptu meeting: sometimes the conclusion and the decided course of action is so cut-and-dried that, essentially, the problem that was no longer is. In that case, usually only one person continues with the actual implemention. In other cases, the result of the meeting is an avenue to explore via coding, and in that case we continue attacking or unravelling the problem together, behind a computer.
My colleagues and I tend to apply pair programming regularly, but only when it's useful, and as long as it is useful. It helps us tackle hard or complex problems (and we only use it for that purpose). The way we apply it, it's also definitely 'pair programming' and not just 'pair coding', since we fairly often briefly move to a meeting room to talk over the problem and possible sketch it out.
Do you follow through on the problem as a pair completely to the end? That is, after the meeting room do you sit at a computer (just one) and code the solution as a pair?
I'm asking since I've regularly had programming meetings in pairs, but we've almost always broken up after that meeting. On rare occassions I've had somebody else help me with code, but certainly nothing formal.
That mostly depends on the problem and on the outcome/conclusions of the impromptu meeting: sometimes the conclusion and the decided course of action is so cut-and-dried that, essentially, the problem that was no longer is. In that case, usually only one person continues with the actual implemention. In other cases, the result of the meeting is an avenue to explore via coding, and in that case we continue attacking or unravelling the problem together, behind a computer.
This flexible and adaptable approach sounds entirely reasonable to me.