The vision of the distributed classroom calls for distributing the classroom experience across space and time. In order to get started with this vision, we need to have a clear understanding of what we actually mean by “space” and “time.” Confusion about these definitions can sow miscommunications at critical junctures. For example, in discussing MOOCs, we often describe courses as either synchronous and cohorted, or asynchronous and self-paced. A “synchronous” MOOC has a cohort of students moving through the program with shared deadlines, aiming to create more of a community feel deriving from going through the same experience at the same time as others. However, in almost every “synchronous” MOOC, there is no true synchronicity: students still complete work on their own time, interacting with one another via asynchronous forums rather than synchronous classrooms. Even those that offer some sort of more synchronous mechanism leave it optional; many MOOCs use Slack or Discord for real-time chat, for example, but typically as an optional additional communication channel. So while these courses are not truly self-paced due to the shared, cohort-based deadlines, they nonetheless are primarily asynchronous, at least as compared to true live synchronicity. This differentiation is just one of the places where descriptions of synchronicity and remoteness may take on different definitions.

THE TIME/SPACE MATRIX

One of the theoretical tools employed by the field of computer-supported collaborative work (CSCW) is the time/ space matrix, which looks at interactions among individuals from the perspectives of the two dimensions of time and space (figure 2.1).1 Each dimension is made of two relatively discrete categories: time is divided into synchronous and asynchronous, and space is divided into co-located and remote. This leads to four quadrants.

We can apply this framework to many existing educational environments. A typical classroom, for instance, belongs in the synchronous co-located quadrant, where students interact with one another and with their teacher live in a shared location. That synchronous co-located time can be used for a variety of purposes, including lectures, class discussion, group work, and testing; the important element here is that synchronous co-location is available to be used.

Remote and distance-learning classes very often implement the synchronous remote quadrant, a synchronous experience where people are not physically in the same place. Often—especially in the wake of COVID-19—this is an instructor teaching via a teleconference, where rather than coming to a physical classroom, everyone signs onto the same link at the same time. This might be performed using existing teleconferencing tools like Zoom or Microsoft Teams, or it might use dedicated synchronous remote classroom tools like Harvard’s HBX classroom or the Minerva Project. Sometimes this synchronous remote distance-learning experience is added on to a synchronous co-located experience, where a remote audience can watch a live class. This is our first glimpse at a distributed classroom: a single class is distributed across a synchronous co-located and synchronous remote experience. Viewed this way, all of televised sports and live entertainment can be seen as a distribution across synchronous co-located (for in-person attendees) and synchronous remote (for those watching on TV); shows recorded in front of a live audience are distributed to the asynchronous remote quadrant as well. In the wake of COVID-19, some schools experimented with other variations on the synchronous remote quadrant; one model that emerged was a sort of “study hall” approach where remote students were assigned work to complete during a specific class time but their efforts were largely individual. A teacher would be available to answer questions live via a chat tool or teleconference, but the teleconference was not the main driver of the experience; instead, students worked more independently, albeit at an assigned time.

The obvious candidate for the asynchronous remote (AR) quadrant, then, is the typical MOOC model. Whether students are cohorted with shared deadlines or allowed to pursue the course in a purely self-paced model, these sorts of courses build heavily on material that is persistently available via the internet, breaking the necessity for synchronous or co-located interactions. The AR quadrant is not absent from traditional education, however; the very notion of homework can be viewed as leveraging asynchronous remote environments: the teacher asynchronously provides directions to the student to complete later in their own space. As technology has taken on a greater and greater role in traditional education, asynchronous remote interactions have become more significant: even in classes that meet in person, teachers may use learning management systems to post announcements, collect homework, and return grades, shifting more of the classroom administration into the asynchronous and remote sphere.

The final quadrant, the asynchronous co-located quadrant, is a bit of an anomaly: How can you be co-located without being together at the same time? But it nonetheless plays a significant role in the design of traditional classrooms. Lower school classrooms may tape students’ work up on the wall, thus providing a passive mechanism for communication among different classes. Schools may employ idea walls to gather input or give students room to express themselves. Museum kiosks may let users see the creations of earlier informal learners. From the perspective of the distributed classroom, though, asynchronous co-location will take on new importance as we reinterpret “co-located” to instead carry implications about the features of the space: it may not be important for asynchronous learners to be in the same actual space, but rather to individually be in spaces that support the same activities and behaviors.

The time/space matrix has already been a useful tool in rethinking how we use different constructs in designing the student experience. Blended learning, for instance, looks at the fact that the scarcest resource in designing a course is synchronous co-located time; a class may meet for only three hours a week. Those three hours, thus, should be spent on whatever can be accomplished only within the synchronous co-located quadrant, as one would find in a flipped classroom model. We may then look at what typically happens during that time—such as a lecture—and determine that it can be shifted to the asynchronous remote quadrant to better use the scarce synchronous co-located time.

The matrix on its own is a valuable way of thinking about interactions, although many of its implications for the design of learning experiences are rather straightforward: often the quadrant within which we may design is given to teachers rather than selected by teachers. We know at the start, for example, that we are teaching in-person or remotely, and whether we can expect students to sign in at a particular time; then we move on from there with our own designs. The goal of the distributed classroom is to design across, rather than within, these quadrants—not to design a separate class experience for each quadrant, but rather to design an overall class structure that covers multiple quadrants.

EXPANDING THE MATRIX

In the pursuit of the distributed classroom, we make two additions to this general structure: one structural, one semantic. These changes are made to reflect the reality of what we are designing for in an educational experience, as well as to shift focus on what we hope to realize with the distributed classroom.

First, under the standard interpretation, synchronicity and co-location are relative to the original delivery of the course, which is typically framed from the student-instructor point of view, especially whether students are with the teacher live in the same room. However, what we often want to preserve are interactions among classmates; or, we may want to preserve interactions not necessarily with the teacher, but with what we would commonly call a teaching assistant. Teaching assistants, under this model, are individuals with less content knowledge and experience than the teacher but more than the students in the class, acting under the designs and direction of the original instructor. They may facilitate group interactions, give individual feedback on work, answer questions that do not require the expertise of the instructor, and curate questions to deliver to the instructor that do rely on that expertise. Teaching assistants are common in large programs—our OMSCS program employs over four hundred of them each semester, and our on-campus program employs dozens as well. Analogous roles are common in smaller programs and K–12 classes as well, usually in the form of paraprofessionals with more specific roles like supporting students with Individual Educational Plans (IEPs) or other accommodations. We come back to the relevance and role of teaching assistants throughout this book.

The important point of this observation is that there ought to exist a middle layer of the classic time/space matrix. Both synchronicity and co-location can be thought of not only in terms of the student’s positioning relative to the original delivery of the content, but also relative to a cohort of classmates and teaching assistants. This relationship is automatic if one happens to be a student in a traditional model, but attendance in a traditional model need not be a requirement for a student to have some synchronous, co-located classroom with classmates and instructional staff.

Thus, our structural modification of the time/space matrix in support of the distributed classroom is to insert an additional level along each dimension. In terms of synchronicity, we have three levels:

• Synchronous with the original class: Students participate with the original delivery of the class, whether in person or remote.

• Synchronous with a cohort, asynchronous with the original class: Students participate with a synchronous cohort of classmates and teaching assistants, watching or following along with the material as it was originally presented, pausing to engage with their own cohort at moments when classroom interactions happened in person.

• Individually asynchronous: Students consume the class material entirely on their own time, without a synchronous cohort of students to interact with in a live, designed way.

Then, in terms of co-locatedness, we have three levels as well:

• Co-located with the original class: Students are in the same place as the teacher for the initial delivery of the material.

• Co-located with a cohort, remote from the original class: Students gather in a room together to participate in the learning experience, but that place need not be the location of the original delivery of the material.

• Individually remote: Students may attend from wherever they are, so long as the technology is available to support their attendance.

With these additional layers, we establish the distributed classroom matrix. Chapter 4 expands on how each element of this matrix may function in the real world. However, even before examining the individual elements, it is possible to see the potential power of this approach. In high-demand fields, a key limiting factor in delivering cutting-edge content is the availability of instructors to teach that content. Even in more well-trodden fields with established training regimens for bringing new teachers into the fold (such as K–12 education), individual variation is strong, and teachers are asked to serve far too many functional roles at the same time. Part of the power of this approach is that it may separate out elements of the experience—such as the initial classroom delivery—into reusable commodities. Rather than limiting the number of students an excellent teacher or a subject matter expert can instruct, their delivery of the content becomes something that can be leveraged toward the education of an unlimited number of students without increasing the workload on that teacher individually. For example, each semester, we together teach over three thousand students in classes we originally filmed; delivering live lectures to that number of students would be two full-time jobs on their own, but instead we teach the classes in addition to our roles as executive director of the program and dean of the college because we are able to leverage this commodity. Perhaps more interesting, professors who have moved on from Georgia Tech continue to have an impact—in some cases, continuing to teach their Georgia Tech classes from new positions at other universities, and in others allowing a new instructor to continue to leverage their content. These approaches allow classes to be taught that might otherwise be retired when the professor with the necessary knowledge moves on.

This on its own, of course, is nothing new: it can be traced back to the broadcast of lectures via radio and television through initiatives like the United Kingdom’s Open University (which has been ahead of its time in a multitude of ways). Those efforts, however, fall squarely in the asynchronous remote quadrant (or the synchronous remote quadrant before the advent of convenient individual remote communication) where classroom interactions are difficult. The distinguishing factor brought by this new dimension is the potential preservation of classroom dynamics and interactions coupled with the distribution and reuse of the initial classroom presentation. It may no longer be necessary to sacrifice the authentic classroom experience for the convenience of mass distribution; a compromise between the two, preserving their benefits, may remain possible.

That shift on its own, however, would be somewhat limited in its impact. If we merely viewed the distributed classroom as a way for a teacher to teach a cohort of students that is separated in space and time, then it might seem to be of narrow usefulness: How often is it really the case that a cohort of students can gather together at a time and place an instructor cannot reach? There might be use cases for training workers in remote areas, but that does not intersect with the larger majority of education.

This leads into the second modification of the time/space matrix, one that is more semantic than structural. While it is not deliberate, the majority of applications of the time/space matrix we have seen situate initiatives squarely within a particular quadrant: an interaction is synchronous co-located, synchronous remote, or asynchronous remote. The goal of the distributed classroom is different: rather than positioning teaching in a particular cell of this matrix, its goal is to design learning initiatives that span many cells with as little additional labor as possible.

Every cell of this matrix represents a compromise. Given the choice, we have no doubt that most students would choose to participate in the original room with the teacher instead of with teaching assistants if both options were equally accessible. However, in many subject areas, synchronous co-located participation is impossible for the vast majority of students who would benefit from the experience. For many of these students, the existing alternatives—asynchronous MOOCs or playlists of prerecorded lectures—sacrifice far more of the experience than those students need to give up. The focus of the matrix is on offering a suite of potential compromises from which to select, such that students may maximize their benefit and minimize their sacrifice, without adding tremendous extra work on teachers to accommodate each set of constraints.