RAD Lab

Some believe that if they put cubicles into open space, they have RAD Lab-like Space. We disagree. Now that we've been living in the space for 6 months, we have strong opinions on why the space is so effective:

* No faculty offices Faculty offices occupy prime window space, are underutilized, and lead to poorer meetings (see below). In addition, since space is a zero-sum game in Soda Hall, offices for some means less space and meeting rooms for everyone else. * Optimized Meeting Rooms Floor-to-ceiling white boards, big displays or projectors, teleconferencing phones, power plugs and cables built into table (see below). * Open space with low partitions If you stand, you are be able to see everyone * Maximum white board space Floor-to-ceiling on every available surface, with lots of pens. * No landline phones on desks Rely on cell phones thereby reducing noise pollution from ringing and talking on landline phones. (The policy is that you move your cell phone conversations into a meeting room or outside the lab.) * Big desktop displays and laptops Bigger than home to attract you, laptops to reduce noise. * Card key entry So that can feel comfortable leaving things on desks, occupants use rooms. (Faculty office hours are held in an office outside the RAD Lab.) * Free drinks and coffee Such support gives the space a different feeling, making it more inviting. As an anthropologist once remarked to Dave Patterson about how a free meal makes groups happy: “Food is love.” * Adding a kitchen and a lounge Standard additions that are good ideas when remodeling space.

Faculty find that meetings in the RAD Lab meeting rooms work much better than in their old Soda Hall faculty offices. Many have suggested theories as to how this could possibly be true: * Neutral Territory Rather than enter the professor's domain with the professor sitting on his or her throne, everyone is equal in a meeting room. * Large, available white boards Whatever little white board there is in the faculty office, its surely covered with potentially important ideas that you need permission to erase, as you'd hate to destroy an important proof or observation. The floor to ceiling white boards of the RAD Lab rooms are both much larger than in any faculty office and you can erase anything you want when you come into the meeting, giving the group much more space to express ideas. * Staying in the flow Between phone calls, email, people dropping by, … it's easy for faculty to get distracted in their offices and not pay attention to the meeting, or for these small interruptions to disrupt the flow of ideas. Meeting rooms can keep everyone in the flow by significantly reducing disruptions. * Ready to go Since all the cabling, power, and projectors are already set up on the top of the tables in the meeting rooms, you sit down, plug in, and start the meeting 2 seconds later. This skips than the traditional crawling around on your hands and knees to find power and then the read the guidelines (unique to each room) to figure out how to turn on the projector, go back to get the video dongle you forgot to bring, and connect to your laptop. * Projectors Every meeting room has a built-in projector or large display which the group can easily see. Faculty offices have neither. * Windows and an audience Jordan quotes a famous study with cockroaches running a track [5]. Race 1 has opaque partitions, so no runner sees anyone else. Race 2 is with transparent partitions, so the runners see each other. Race 3 is with transparent partitions and spectators. The fastest times come in Race 3. The sociopsychological observation is what you normally do well, you tend to do better when other people are watching, especially with a supportive audience. Hence, the meeting room windows give you this sense of an audience that is watching you, which enhances your performance.

An early evaluation of this is experimental space is found in a survey taken in December 2006, 2 months after we moved in. It showed that 70% of people come into Soda more as a result of the RAD Lab and that 85% have had multidisciplinary interactions. People like the easy access to people, the pleasant space, free drinks, meeting rooms, big displays, whiteboards, and a sense of community. Concerns were about the blinds/lighting at desks and knocking at the door.

One of the bets of the physical RAD Lab was that faculty accessibility would attract students to Soda Hall to increase the chances of innovation due to spontaneous, unplanned meetings.

What we didn't realize was that access to faculty would be attractive to potential graduate students as well.

Historically, we get land about 40% to 50% of admits. This was even true in Spring 2006, which was after the RAD Lab project started but before we built the physical RAD Lab. We had thought that a new research project that was sponsored by Google, Microsoft, and Sun would be an effective recruiting tool, but the yield was within historical norms.

In Spring 2007 we hosted meetings in the new RAD Lab for the OS/Networking graduate students, and they said they really liked faculty accessibility. The 2007 yield backs up that observation: it was 85%, the highest in memory.

Hence, faculty availability via open space helps bring both current and future students to Soda Hall.

One HCI paper studied the impact of space on productivity.[4] They found that radical collocation could double productivity. Here is the abstract of their paper:

//Companies are experimenting with putting teams into warrooms, hoping for some productivity enhancement. We conducted a field study of six such teams, tracking their activity, attitudes, use of technology and productivity. Teams in these warrooms showed a doubling of productivity. Why? Among other things, teams had easy access to each other for both coordination of their work and for learning, and the work artifacts they posted on the walls remained visible to all. These results imply that if we are to truly support remote teams, we should provide constant awareness and easy transitions in and out of spontaneous meetings."//

As in the RAD Lab survey, they highlight the importance of easy access to others, and reference other work on the importance of having people nearby and within sight to encourage interactions.

//One of  the  main  drivers  of  success  was  the  fact  that  the  team  members were at hand, ready to have a spontaneous meeting,  advise on a problem, teach/learn something new, etc.  We  know from earlier work that the gains from being at hand  drops off significantly when people are first out of sight, and  then most severely when they are more than 30 meters apart."//1]

They also saw the value of whiteboards, as in the RAD Lab survey,

//Team  members used much of the whiteboard space to support their  problem solving meetings..."// 

They conclude that the productivity improvements are astounding.

The correlation between distance and communication is called the Allen curve because Allen found that the frequency of communication between engineers drops off exponentially as the distance between them increases. After about 50 meters, there's very little communication between groups of engineers. Allen recently co-authored a new book that examines the relationship between space and innovation. [2]

Some wonder whether telephones and email change the observation. Allen studied the question, and they don't. As he says [2]

//We do not keep separate sets of people, some of which we communicate in one medium and some by another.  The more often we see someone face-to-face, the more likely it is that we will telephone the person or communicate in some other medium."// [p. 58]

Moreover, complex conversations need to be face-to-face:

//Discussing anything that is complex or abstract by telephone or email is very difficult. We need to meet directly with the person. ... [A study concluded that]  the telephone was used extensively for communicating on simple topics, even if distance was short. ... Face-to-face was used extensively for communicating complex topics, even if distance was extended."// [p. 60-61]

And we need subtle clues to have effective communication:

//First, many things ... are difficult to communicate with words alone. We need the assistance of diagrams or sketches. In addition, we often need the feedback that comes from looking into the other person's eyes, which communicate understanding. ... Body language ... provides unspoken feedback that is very powerful. ... This feedback system is invaluable in guiding communication."// [p. 60]

He categorizes communication three ways: 1) for coordination, 2) for information sharing, and 3) for inspiration. This last category is important:

//In an organization that relies on creative solutions to problems, communication for inspiration is absolutely critical. It is usually spontaneous and ofter occurs between people ... [drawn] from different disciplines."// [p.28]

// we found that 80% of the information underlying new ideas came through personal communication."// [p.86]

and most strongly affected by the space:

//communication for inspiration is the type most affected by the type of physical space. Most communication of this type occurs during //chance encountours// which create the possibility for inspiration and creativity--the sources of innovation. It is very obvious that the ways physical space is configured it can strongly promote or impede the occurrence of chance encounters."// [p. 86]

What space encourages chance encounters? Being close

//First, it seems obvious that an organization ... needs to ensure that the distances amoung [staff] are minimized. ... to minimize distances, it seems a square single story building is most desirable. We say single story because evidence indicates that vertical separation always has a more severe effect..."// [p. 71]

and being able to see people

//In the realm of communication for inspiration ... visual contact is probably the most important. If people do not see each other, they will not have the opportunity to interact and create that knowledge."// [p. 73]

The lack of offices in the RAD lab helps shorten distances and enables visual contact, as do the windows on meeting rooms. The extensive white boards make it easy to do sketches in unplanned face-to-face meetings.

This same book [2] provides data that supports skepticism of the effectiveness of geographically distributed centers. Allen's premise is that face-to-face communication in general and specifically communication for inspiration are necessary for innovation.

First, there is by definition almost no opportunity for face-to-face encounters when people are scattered about the country. It will be a long time before telepresense can match the important feedback from looking into people's eyes and reading body language.

Second, the book measures probability of communication of any kind. (This data is comes from a study of a single company, so in some sense its more integrated than government funded research centers, and so these numbers could be high; Table 3-1, page 65)

Here are three models for such centers:

- Single project, integrated location - probability of regular communication 0.95 The center is located in the same wing of a building at a single location, and everyone in the center is working on a common project. (In Table 3-1, its same wing of the building, same department, same project) - Single project, distributed locations - probability 0.23 This is a genuinely intergrated project with common goals located at multiple sites. (Same project, different departments, multiple sites) - Multiple affiliated projects, distributed locations - probability 0.002 This is a loose federation of related projects at multiple sites, who band together–just speaking theoretically–to present the appearance of a coherent common front for a funding opportunity. (Multiple projects, different departments, multiple sites)

This research supports our experience that a multidisciplinary group in an integrated single space will be significantly more creative and productive than people distributed about a building or campus. Fortunately, the RAD Lab took advantage of being at a single site to create a small integrated space with easy visual contact to enable inspirational communication that enhances creativity and innovation.

Allen's book mentions the design of space for communication and space for concentration While it seems that technology cannot yet overcome distance and walls to be as effective as open space for communication, the success of the RAD Lab may be because technology allows concentration without private offices Here are specific technologies that help:

* Cell phones mean phones don't ring (with a sound designed to disrupt break your concentration) and you don't have to overhear conversations on land-line phones

* iPOD and headphones reduce unwanted noise and plays music that inspires you without bothering others

* Big displays provide a larger set of images to capture and sustain your attention

* Instant Messaging status lets you tell others in the lab if you're interruptable or not

* Laptops/home broadband make it easy for people to do part of their work at home in a normal day, which offers some concentration time while still putting in a full day in the RAD Lab

[http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&tid=6100] Allen, T. J. (1977) Managing the Flow of Technology: Technology Transfer and the Dissemination of Technological Information Within the R&D Organization : MIT Press, Cambridge, MA.

[http://www.elsevier.com/wps/find/bookdescription.cws_home/709115/description#description] Allen, T. J. and G. Henn (2006) The Organization and Architecture of Innovation: Managing the Flow of Technology : Butterworth-Heinemann, 152 pages.

[http://www.google.com/url?sa=t&ct=res&cd=1&url=http%3A%2F%2Fwww.cs.virginia.edu%2F~robins%2FYouAndYourResearch.html&ei=O1jgRaScI6PAhQOl1qDeBw&usg=__df4KaX_XGsSQykAx51uv9cu7gzY=&sig2=bAenBCT0YVg8M6ke8GJs2g] Hamming, R. You and Your Research, presentation, Morris Research and Engineering Center, Bellcore, March 7, 1986.

[http://www.google.com/url?sa=t&ct=res&cd=2&url=http%3A%2F%2Fpossibility.com%2FMisc%2Fp339-teasley.pdf&ei=Y1TgRdKoCafIgQPi5MymBw&usg=__MZ5FIaCqIRo4VWi0G2ynpGC009U=&sig2=4VFs6aMWCbZ7_DtYLYgJbw] Teasley, S., Covi, L, Krishnan, M. S., & Olson, J. S. (2000). “How does radical collocation help a team succeed?” Proceedings of the 2000 ACM conference on Computer supported cooperative work 339-346.

[5] Zajonc, R.B., Heingartner, A. & Herman, E.M. (1969). Social enhancement and impairment of performance in the cockroach. Journal of Personality and Social Psychology, 13, 83-92.

[jon@haribo ~]$ cp /dev/null !$ cp /dev/null space [jon@haribo ~]$ vi !$ vi space [jon@haribo ~]$ cat space Overview of survey results taken December 15, 2006. Responses from 44 people (out of 44!)

* Our hope in opening the RAD Lab was that people would both be here more and that people from different areas of expertise would get together to innovate across disciplines, either with formal meetings or serendipitously when, say, getting a coffee. In the 10 weeks since the lab has opened, have you had such interactions, either in a formal meeting or informally?

• ~ 10% Never
• ~ 0% Once
• ~ 5% Twice
• ~ 60% A few times
• ~ 25% Many times

* Testimonials on interactions.

Change since RAD Lab opened:

• 30% come in the same amount as before
• 35% come in the a little more than before
• 35% come in the a lot more than before

Below are what people like about the RAD Lab, which I categorized roughly by topic in order of popularity. As many people make multiple points, I count number of times its mentioned in any comment, and the I picked a category to put each comment, usually because it was mentioned first. Thus (17,10) means mentioned 17 times, and I placed 10 comments as primarily in this category. Click to see each set of comments.

• Easy Access to Others (17, 10)
  • Being able to talk to other students easily and more often is excellent. Same for professors but to a lesser extent since they are usually not there (but it's still much better than before the RAD Lab). It's better socially too. Free drinks, great monitors are a nice touch.
  • The ability to interact with other people.
  • Easy access to students, and their easy access to me. I no longer schedule student meetings, and let them come get me when they want to talk. I end up spending much more time with my students, and the time is more fruitful.
  • Lot of opportunities to interact with other people.
  • I get to see other people. I feel more activity at Soda. The environment is better for working on papers (grab a meeting room, grab a diet coke, walk over and easily talk to people). It is nice to have different areas where you can work.
  • Easy to locate/talk to colleagues. Easy to find an expert in another area to ask quick question. Shared work area promotes sense of being part of a team. Meeting room facilities are best I've ever had - less time futzing with equipment, more time working.
  • Easy interaction between people. Nice workspace.
  • Open, everybody pops out here and there, easy to interact

• Pleasant Space (14, 10)
  • Open space. Nice working environment. Easy collaborations from the proximity. Plenty of board space for impromptu discussion. Good free coffee (and now juice! - I hate soft drink by the way, that's why I put N/A above). Now we are only missing great healthy snacks to keep us working more all the time (like at Google)! I like the big screens. I like the environment setup. The sense of community with the social, meetings, etc.
  • Open environment, the awareness of who is around, the closeness of professors, the meeting rooms work great.
  • Some things about the environment are conducing to meeting with others and working – high visibility and access to other people, meeting rooms, etc. I like that the sysadmins are generally available. I like that if I need to see my advisor, I don't need to walk up two floors to see if he is in – I just need to stand up.
  • The open space. The meeting rooms are available most of the time. The soda+juice+tea is a huge plus. Convenient+clean kitchen.
  • Great space, brilliant people, access to faculty, excellent perks, glad I'm in the lab.
  • Quiet, pleasant space. Availability of interesting people to talk to. Coffee and soft drinks. Don't have to go to 6th floor. More stimulating conversations in last 2 months, than prior year.
  • Open space and refreshments
  • Open space. Desk. Easy to discuss things with other grad students. White board. Meeting rooms. Coffee machine.
  • It's flexible, informal/casual, and comfortable.
    1. No closed doors, everyone readily available
    2. The white boards everywhere (with pens, big innovation!)
    3. The screens (my laptop screen feels small now)
    4. Clean (no accumulated junk from generations of past students)
  • Cozy space, easy interaction
  • Open space, light, meeting rooms, lounge
• Free Food/Drinks (10, 2)
  • Free juice drinks
  • Has a kitchen, with free drinks :) o Open area, easy to deploy research equipment. No need to knock on office doors and beg for permission. o Many whiteboards.
• Meeting Rooms (8, 1)
  • No phones on desks, but phones in conference rooms.
  • Refreshments.
  • Isolated printer (less noisy)
• Big Displays (6, 2)
  • Definitely LCD display!!
  • The big monitors have increased my productivity tremendously while writing code. Have 6 big size windows simultaneously open is great even while grading programming assignments for TA work. It's great to have someone to chat with after hours of continuous work.
• White Boards (3,0)
• Sense of Community (2,0)

When asked to rate specific features of the lab, from highest average rating to lowest (Great, Good, Neutral, Irriating, Terrible):

- Big displays (Average rating Great) - Soft drinks (Great) - Refrigerator (Great) - Coffee machine (Good+) - Printer (Good+) - Meeting room availability (Good+) - Light at your desk (Good) - Noise at your desk (Good-) - Blinds (Good-)

While the average rating for the blinds and light were good, they had the 9 of the 10 ratings of “irritating.” The range on rating noise was neutral, good, or great.

We asked “What are issues that could be improved in the RAD Lab?” We got 21 responses, but compared to above, there is surprisingly little repitition in the concerns or suggestions. There are 4 comments of light and blinds mentioned above and 2 about visitors having to knock at the entrance. I put those at the top, but its probably best to just read the comments yourself to lead to a discussion about what can we do make things better in 2007.

• ~ 60% Normally in lab, + weekends/evenings at home
• ~ 25% Primarily in lab, but reserve 1 weekday at home
• ~ 5% Primarily at home, come in for meetings or 1 day a week
• ~ 10% None of the above

* Distribution of number of days per week poeople spend in the RAD Lab, self-reported vs. impression of labmates

FIX TABLE {| border=“1” align=“center” !Days !Self report !Labmate impression

* See plot of use by time of day and day per week . Highlights: Noon-4: >80% Mon/Wed, >70% Tue/Thu, ~ 60% Fri 5PM -7: 67% Mon/Wed, ~60% Tue, ~ 50% Thu Saturday: >20% Noon to 4 Sunday: >20% Noon to 2

• ~ 30% Networking grad students
• ~ 25% Machine Learning grad students
• ~ 20% Systems grad students
• ~ 15% Faculty
• ~ 5% Staff
• ~ 5% Postdoc/Other

* Add an extra row of chairs to board room * More people need to sign-up for Tuesday presentations