« One of My Favorite Baseball Pictures | Main | My Talk on Software Quality Available Online »

August 27, 2006

Jan Chong Talks About Workspace

On Thursday I went to a talk by Jan Chong, a Computer Science Ph.D. candidate at Stanford who is studying how XP and pair programming teams work. I'm interested in this because it offers some insight into the upcoming Workplace Advantage reconfiguration that Engineering Excellence is scheduled for.

This particular talk was a comparison of a team working solo in cubicles vs. a team working paired in an open workspace. This is not quite the comparison that Microsoft is interested in; we are considering swapping teams working solo in offices with teams working solo in--well, in something that MIGHT be open or it might be cubicles or it might just be smaller offices with more common space. But basically we are not specifically considering pairing, and Jan's work is focused on the effects of pairing, not the effects of open space.

Nonetheless, some of her conclusions may apply to solo workers in an open space. She has interesting data on how interruptions between team members happen. In a cubicle space they happen much like in an office space: someone gets up and walks to someone else's cubicle/office, interrupting whatever they happen to be doing then. With open space, there are some cues that the interrupter can draw about what the interruptee is doing, before they decide to interrupt. These cues are stronger in pairs environments (since it's easier to tell when people are working vs. taking a break) but still they likely exist in a solo environment. Also, in an open space the interruptee is often called over to the interrupter's desk, which gives the interruptee better control over the duration of the interruption. This is likely partly due to the fact that the context for the interruption (the bug or issue or whatever) is happening at the interrupter's desk, so it makes more sense for the interruption to happen there, but also because of the 2-against-1 nature of a pair interrupting an individual. The former effect would seem to apply even in a solo environment, while the latter would not.

Peter Provost invited Jan over to tour the Patterns & Practices open space in building 5, and I tagged along (Patterns & Practices was the source of the arguments in favor of open space on Joel Spolsky's forums). P&P is doing a lot of what Jan is studying (paired work in open space) so it was interesting to watch Peter and her discuss it. Peter has a theory that open space works much better with pairing because the pair gives you someone to focus on and therefore prevent background distractions. This of course made my ears perk up since we are planning to go open with solo workers (although two people in EE did recently do some "pairs course development" which went quite well, it's not clear that a lot of our activities are pairable). Jan occasionally would study a solo/open person (when the number of workers at the office on a certain day was odd), but in general those people were so discombobulated from not being paired that they weren't valid study material.

The question of how to indicate "don't interrupt me" (akin to closing the door of your office) also came up. Peter said that people use headphones for this. The problem is that I might be using headphones for a legitimate reason but also be in an extremely interruptible state (watching a recorded talk, say) while other times I might not have the headphones on but be thinking deeply about something. Peter also said that some people put on the headphones when they get to work and leave them on all day. Tie this in with something Jan said, about how in an open space people get used to expected background noise (work talk, keyboard clacking) but get easily distracted by unexpected background noise (social talk, paper rustling), leads me to a new theory that when you move to an open space you have to commit to a couple of weeks without headphones on, to get acclimated to the background noise (sort of like trying Grape Nuts for a week). If you start out right away with headphones always on, then you won't ever get used to it and you will lose the ability to cue your interruptibility by the presense or absence of headphones.

Posted by AdamBa at August 27, 2006 09:19 AM

Trackback Pings

TrackBack URL for this entry:


Currently we have 4-person cubicles for most people, and that's not great.
We are about to move to an office with less space and a completely open plan, where everybody will be directly facing somebody else...to my mind this shows contempt for the development process.
It's not *only* about not being explicitly interrupted: a key thing is to have some peace and quiet in which to think, and also some personal space just to keep the mind relaxed.
I do not see myself as a battery hen.

Posted by: Kevin Daly at August 27, 2006 12:43 PM

Before I retired, I was aware that almost all my colleagues kept their door closed, while I and the prpeople on either side of me didn't. I think I always went home when I didn't want to be interrupted and my kids learned that.

I know you are always aware of grammar, so I will point out that in the fragment "...watch Peter and her discuss it" that "her" should be "she". If the verb in the dependent clause had been an infinitive, then "her" would be correct, but it is a finite verb and needs a nominative subject. The object of "watch" is not "Peter and her", but the entire clause.

Posted by: marble chair at August 27, 2006 07:47 PM

Mea culpa, mea maxima culpa. I was wrong on the grammatical point. The point is that "discuss" is an infinitive and the subject of an infinitive is always in the objective case. By the way, the subject of a gerund is always in the possessive (in "the city's destruction", say, the city does not own anything).

Posted by: marble chair at August 28, 2006 06:08 AM