Offices: Mixing and Matching
The construction on our new building, the Gates Hillman Complex (GHC), has been moving along well. As we speak, the building is still set to open in late-summer 2009. Here is a current webcam picture, showing GHC in all of its 9-floor glory:
This picture shows the south and east-facing walls of the Gates Center. Off in the distance, covered in plastic tarp, is the Hillman Center. I’ve seen enough late projects that I can’t help but be somewhat skeptical about the completion date. On the other hand, it is a very good sign that things are still more or less on schedule. In fact, the university registrar is actually starting the process of scheduling classes to take place in GHC in Fall 2009! Everyone who has gone on one of the “muddy boots” tours of the building has been impressed with (or even amazed by) the architecture.
One potentially difficult aspect of the impending move to GHC is the assignment of people to labs (or, what we prefer to call “project spaces”) and offices. There are lots of great offices and project spaces, but in talking to colleagues around the country, a surprising number of them have gone through acrimonious moves into new buildings. I think it is clear that office assignments have to be approached thoughtfully.
There are lots of things to take into consideration, but one that I find particularly interesting is the issue of “mixing or matching”. Traditionally, we have “mixed up” our faculty and students, by which I mean that (a) students have offices in the same floors and corridors as the faculty, (b) students from different research areas share offices, and (c) faculty and students from different research areas share corridors. This “mixing” has been a great source of growth and interdisciplinary activity in our department. Going back to my start here at Carnegie Mellon, I was put into an office right next to Satya (who at the time was working on the Andrew File System) and Eric Cooper (who was working on gigabit networking). I learned a lot about systems concepts simply by running into them every day. This changed my focus on formal semantics and compilers to a broader interest in software and verification issues in operating systems and networks. Indeed, I would say that the work on proof-carrying code would not have happened without this formative experience.
This type of “stretching” is, I think, a hallmark of CMU research. To take the most recent example, just a couple of weeks ago, Rob Simmons (a CS Ph.D. student working on logical frameworks) and Shay Cohen (an LTI Ph.D. student working in computational linguistics) received the best student paper award at the 2008 International Conference on Logic Programming. I realize that every department head likes to brag about how “interdisciplinary” their departments are, but here this kind of collaboration is simply routine.
When it comes to office assignments, I definitely have a mixing philosophy.
Having said that, there are quite a few faculty (and students) who would prefer to “match”, meaning that they would like to be near other researchers in their same or closely related fields, and furthermore have their students share offices with other similar students. This preference seems to be strongest among the junior faculty who, by and large, have come from good departments where matching is the norm. There are undeniable benefits to matching. Junior students can learn a lot from senior-student officemates in the same research. And there is plenty of evidence showing that even a one-floor separation greatly diminishes the opportunities for those “chance encounters” that are crucial for sparking new collaborations.
So, what to do? I’m mulling it over, along with quite a few others in the department. Luckily, the architects have also thought hard about this, and a careful inspection of the building design shows numerous “vertical cuts” through the building, in an attempt to create transparent and convenient connections both horizontally and vertically through space. As I’ve thought about this more, it looks like it might be possible to accommodate both mixing and matching by taking advantage of the numerous cuts and “Hogwarts-style” stairways and helices.
It’s actually a fun problem to think about. And yes, I’ve also been hacking on my office-assignment-via-stochastic-optimization program…
Peter Lee @ December 18, 2008