Community (?) of programmers

In “The Social Life of Information” John Seely Brown and Paul Duguid wrote about how Xerox techs used to gather informally during breakfast and other times.  Although the purpose wasn’t for business or to exchange technical information or to share war stories about work practices, that is what they did.  In the olden days of mainframe computers, inbetween the era of punchcard machines and individual terminals in every cubicle, we programmers used to sit in bullpens where we shared terminals.  Then as now entering code was a relativity minor activity in our day.  Programming is thinking, not typing.  A side effect of sharing terminals arrayed in a bullpen is that we actually saw each other and could compare notes, ask questions, mentor and learn from each other.  Not unlike the Xerox techs. When we each have our own computers this enforced community breaks down.  Someone has to go to another person’s cubicle–or worse, make loud noises over the partitions.  Some cubicles are set up so that the terminals have to be placed in the shared intersection and the programmer has to sit with his back to the opening.  Which I personally find rather creepy but then I’ve seen too many horror flicks.  Cubicle terminals seem to make interruptions more intrusive.But in terms of the analyst – programmer relationship, I wonder if a communal environment would help foster the exchange of ideas, knowledge, identify areas of confusion and need.  Extreme programming already includes the concept of paired programming.  The idea of paired programming is that typing, again, is not programming so it’s not like one person typing is cutting productivity in half.  But you do gain two minds applied to a problem, sharing ideas, one debugging the other and so forth.  All to the good. But what if we expand the notion?  Have the analyst sit in and provide their perspective and expertise?  Swell.  Except that the analyst isn’t needed much at that phase in the process.  If they are, you’ve got serious design problems. Two people in one cubicle is rather intense.  A crowd just sucks up too much oxygen.   But what if we opened up physically, back to something like the bullpen where we could meet and talk without planning to.  For the analyst to participate, they shouldn’t have cubicles at all.  They should be in open seating just off the array of terminals in the bullpen.Two problems, at least.  The cubicle police would never go for it.  Second, when the analysts do need their own quiet concentrate time they will have to flee their own desks and all their materials.  Just a thought.