| Newsgroups |
gmane.comp.programming.extreme-programming |
| Message-ID |
<[email protected]> |
Steven,
> On Oct 18, 2016, at 12:19 PM, Steven Gordon [email protected] [extremeprogramming] <[email protected]> wrote:
>
> Distribution is an even worse team-building idea. So bad an idea that it is very likely that the group will effectively never become an team. So bad an idea that it is much more likely that members of the work group will leave for better opportunities at inconvenient times.
I completely agree that distribution is terrible. As is single-person code ownership.
>
> Having a "bus number" of 1 is extremely risky for a distributed work group. The reason to mandate pairing is not necessarily for the traditional team-building and quality-improvement aspects of learning to pair well, but just to increase the chance that when a person leaves the work group, there is at least one other person who has some knowledge of any particular part of the code that later needs repair or extension.
I’m sure no one is recommending a bus number of one. I fully understand the reason for mandating pairing, and if I were somehow required to create a distributed team, I could see making the point at hiring that everything would be paired. I fully understand the purpose in mandating it after the fact if it had not been mandated at hiring.
And I would still much prefer not to mandate a practice like this. It might be the best possible thing to do, but I think it is unlikely to be a very good thing.
I’d prefer to do whatever it might take to get them to choose pair programming, or mob programming, or some other collegial way of doing things. Adam’s example, I think, is better than mandating. And there might be something better to get them pairing.
Also let’s note that you’ve begun to list some of the concerns to which we think pair programming is a solution: single ownership, surprise exits, and so on. What if we listed those concerns and sat down with the team and invited them to solve the problems. They might come up with pair programming. They might some up with other things to do, maybe even better things.
Ron Jeffries
ronjeffries.com <http://ronjeffries.com/>
Before I start arguing with somebody, I try to save them a little time by letting them know that I’m usually right about stuff.
— Andy Richter