Updated 3/27/1998 By Brad Cox

How

This plan proposes a Modular, Interdisciplinary curriculum delivered to classrooms, homes and offices according to the Use the right tool for the job philosophy within an organizational structure that features Specialized Roles and Responsibilities. These highlighted terms will be defined in separate sections of this document. A final section, Coordination Technology, will explain a distinction between communication tools and coordination tools that has been helpful in developing this approach.

This document explains these terms by showing how they've been applied in a particular course. A separate document describes the web-based infrastructure upon which this approach is based. The example course is Taming the Electronic Frontier (hereafter TTEF), a televised, web-based introductory core course for the MA in Telecommunications program at George Mason University. The web-based student assessment techniques used in this course are described in Evaluation Methods Used in Web-based Instruction and the Online Course, Taming the Electronic Frontier by Donna Potter (a student).

Interdisciplinary

Academic reform initiatives often run aground on either-or disputes between advocates of what is often called academic rigor on the one hand and relevance on the other. Such debates are unnecessary and harmful. The world expects both rigor and relevance, together and inseparable. Why require students to choose between learning narrow skills at a technical academy and receiving a broad liberal arts education at a university when a new university might provide rigor and relevance at the same time? The traditional separation of theory from practice is an artifact of the disciplinary structure of traditional academia. It is part of the problem and not on the path to a solution. Theory and practice should be integrated so intimately that the question, "Which is best", never emerges as worthy of serious consideration.

The example course demonstrates one way of doing this. This course combines:

• technical skills in web publishing
• writing
• philosophy (What is Quality? Value?)
• Interpersonal skills (group dynamics).

This shows one way in which theory and practice might be combined in the curriculum as a whole.

• Students demonstrate their writing skills initially by writing recipes for others to follow, describing how they accomplished technical steps. The course infrastructure publishes such work to everyone, thus creating peer pressure to write well. I critique the answers with a special web form whose output appears annotations alongside each answer when the student reviews the task again. Thus, the course teaches basic writing skills implicitly, in the context of everyday problem solving, not divorced from practical concerns.
• One of the tasks assigns Robert Pirsig's Zen and the Art of Motorcycle Maintenance; An Inquiry into Values. It requires the student to provide essay-style answers to how this books treatment of Value and Quality relates to the frontier-taming topic of this course. The course addresses questions of value and quality in the context of web publishing, not philosophy as a disciplinary subject divorced from practical concerns.
• A subsequent task requires students to publish a student portfolio on the web. The apparent purpose is to demonstrate newly acquired technical skills. However students gradually become aware that there's a larger purpose, which is to raise the theoretical question, "What is Quality", in connection with the practical question of what should be included in, and what excluded from, a published work. Since quality and value (particularly objective, subjective and intersubjective meanings) have been introduced in advance, students have a theoretical vocabulary to apply to a practically grounded task.
• A subsequent task gathers quantitative measures of the value of each work. This is done by dividing the students into "markets", in which each student "inherits" a sum to be spent on other students' portfolios according to the value each portfolio provides them. This is accomplished via a web form that provides a menu of amounts to be paid and a box for narrative comments, critiques and suggestions. A subsequent page of this task presents the comments others have written about their portfolio.
• A subsequent task presents the bank balance earned by each portfolio, thus reflecting their quality in a fashion that is obvious to all. The task includes several essay-style questions that invite the student to integrate what they learned from the ranking process (and the customer feedback questions) with the practical and theoretical components of the course
• The students are given an opportunity to improve their portfolios based on what they've learned. Then the ranking process is repeated, this time knowing that the revenue bell curve will be used to assign grades. Although the actual weighting of this peer-assigned grade is quite low (less than 2%), it is sufficient to cause considerable reflection on how quality relates to grades and their lot in life outside of academia.
• Several tasks are used to forge strangers into cohesive teams. A Desert Crash simulation has each team members rank 15 survival items according to importance. Students then meet as a team to arrive at consensus answers to the same questions. Each individual enters these as before. The task prevents each member from proceeding until all members have submitted precisely identical numbers. The tool doesn't help them determine who is responsible if the team is blocked from proceeding (which invariably happens). They must locate the reason and solve it as a team by using the best tool for the job. This leads students to explore the web conferencing and synchronous chat facilities so naturally that I rarely have to cover this in class.
• Thomasina Borkman, a Sociology professor who has graciously volunteered to help me teach the group dynamics part of this class, contributed several standard tests that she uses in her group dynamics classes. I've converted these to web-based questionnaire tasks to give the teams further insight into its members tastes and preferences.
• Once teams exist, they are responsible for planning and carrying out a semester project. The instructions are simply to choose a breakdown and eliminate it. Team grades are determined by a customer that the team designates. Team member grades are determined by flowing this grade to each member in proportion to their contribution. Contributions are determined democratically, by a web form that each member completes. The form collects a numerical contribution for each member and a narrative about their contribution. These are reported in a grade report at the conclusion of the course.
• I've also used Digital Product projects. The requirement for these is to build a web site with enough real value that other people will pay money to acquire it. These have been remarkably successful at turning unskilled strangers into cohesive (and usually profitable, at least on a small scale) virtual enterprises in just 15 weeks.

This combined approach, integrating web publishing, philosophy, writing, and group dynamics into a seamless whole, has been quite successful according to student evaluations of the course. Course evaluations are gathered each semester by a web page at the conclusion of the course and maintained online thereafter.

Although this interdisciplinary approach to integrating theory and practice might be used in other courses, this remains to be proven.

Modular

If we expect to serve several diverse markets in a cost-effective manner, an alternative to the "course" as the unit of reuse is crucial. This section describes such an alternative by using the terminology of the software reuse community.

A traditional 15 week brick-and-mortar course for undergraduates is unlikely to be attractive in a corporate education context. Employers cannot spare critical employees for so much time. Although traditional classrooms acceptable for campus-based students, off-campus employees find commuting to campus classrooms quite unappealing.

The traditional unit of granularity, the 3-credit course, is so large that reusing course materials within academia and externally within industry isn't possible. A solution is to define a smaller unit of reuse whose modules can be combined to create larger-granularity products tailored to the needs of each particular market.

Thus, we distinguish between the unit of product marketing and the unit of internal reuse. The unit of product marketing is the product as seen by the external world (the individual or corporate client). The unit of internal reuse is the reusable modules that are assembled to construct a marketable unit.

Tasks as the Internal Unit of Reuse

The unit of internal reuse is the task. A task is a self-standing unit of instruction.

• Students experience tasks as a link that automatically appears in their locker when the task is assigned.
• Task authors experience tasks as a text file. Each task file is subdivided into a number of task pages. These task pages are not static HTML documents. Each page is a computer program, written in perl, which emits the HTML commands to display that page in the student's browser.
• As students submit tasks, the answers they provide in html forms are stored in a database on the server.
• Task evaluators ("graders") experience task submissions as a hotlink that appears in a special form that is accessible only to those assigned to that role. The form shows, for each assigned task, the number of submissions in several states: Assigned (work not begun), InProcess (but not yet complete), Submitted (awaiting review), Revise (unacceptable for any reason) and Accepted (perfect work). In most cases, a perfection-based grading policy is used, in which imperfect work is never accepted. It is simply returned for rework. Perfect work; e.g. work that is accepted by the deadline automatically earns 100%. The grade decreases by 10% a week thereafter.

This is explained at more length here. The point is that each task page is an active program, not a passive data file. The programs emit the HTML commands that determine what the students will see in their browser. These commands may and generally do include HTML forms commands to create menus and text entry boxes for the student to complete. The student's answers are automatically stored in a database for each student. This database is accessible to these programs. Programs can (and often do) access the data to determine what the next student sees.

This capability is typically applied in the following manner:

• The initial pages of a task would typically deliver some unit of instruction. For example, the initial pages of an introductory skill-building task in TTEF explain how to how to transfer files from the student's PC to the student's web server.
• Subsequent pages would typically instruct the student to put what they've learned into practice. For example, the task mentioned above instructs the student to transfer a file to their web server using the instructions they've received and then to enter the URL of the page they've just transferred in a web form. When they submit the form, the URL is recorded in that student's database. The evaluator examines this record to see whether the unit of instruction has been absorbed; for example, by using a web browser to verify that the URL actually works.
• Students often find expert instruction is difficult to understand. The causes are deep, much deeper than can be addressed by expecting experts to "try harder" to communicate in terms beginners can understand. Therefore, such tasks generally include an essay-style question in which each student explains how they did it. Since tasks are active programs, not passive data files, they retrieve this information from the database and present it as part of the instruction. Thus, most tasks provide "how to do it" instruction written by both experts and beginners. The beginners answers are obtained by providing the answers written by those who've completed the task successfully. Students have been most appreciative of this approach to collaborative learning.

Web-based tasks will probably be the predominate module of reuse. However, web-based tasks are by no means the only reuse module we envision. Others are explained in the Use the right tool for the job section of this plan. Web pages can encapsulate educational material that is not web based. For example, web page can encapsulate a CD disk by providing instructions on how to purchase the required disk. Thus, this discussion equates web-based task to mean any reusable educational module at all, regardless of whether the actual delivery mechanism is web-based or not.

Products are assembled from tasks

If tasks are reusable units of instruction, educational products are the result of assembling them into a product for any given customer. For example, undergraduate courses might be assembled from some set of tasks and consulting-style organizational change programs assembled from others.

I find it useful to distinguish between the content of a product such as a course and the wrapper that surrounds that product and presents it to a given market (think of the packaging that surrounds products in stores). For example, a syllabus is part of the wrapper for an undergraduate course, whereas the wrapper for an industrial consulting program would be the marketing collaterals used to advertise it.

The tools for assembling specific products from a collection of reusable tasks and custom wrappers are simple. A new course begins with an empty window with boxes for course title, syllabus links (URLs), start and end dates, and so forth. The course designer fills in this structure with a sequence of lesson dates and titles. Then tasks are chosen from the task database, assigned to each lesson, and deadlines are assigned to each task. The hard part is designing meaningful tasks. Assembling them is routine.

The result of this assembly process, as seen by a student, is called their "locker". This is simply a dynamically-generated web page that is the first page they encounter when they enter a course. As explained in the Coordination Tools section, this page does not display everything they might ever need to know about the course (signal + noise) but just what they need to know to be successful this week. In particular, it shows only tasks that have been assigned but not accepted as complete. To see the complete schedule, their grades on accepted tasks, ongoing discussions with their classmates and the like, they must visit a different page.

Use the right tool for the job

The "Use the right tool for the job" philosophy simply means actively avoiding any tendency to focus on any single educational technology and choosing technologies most suitable for each task at hand. For example, academia's tendency to use classroom-based technologies for everything is an egregious violation of this rule.

• For part or occasionally all of some courses, the right tool for the job is a traditional face to face lectures in classrooms. For industrial clients, we anticipate that these classrooms will usually be conference rooms at the employer's work site. For individual students, the classrooms will be centrally located conference facilities owned or rented by the university.
• The best tool for some jobs is course ware delivered over the internet. For industrial clients, students will use office intranets to access course materials on our web servers. For individual students, they will typically use modem connections to achieve the same effect. We expect that this mode of course delivery will be the most common. I have found that this mode of course delivery is often more effective at building experiential, collaborative distributed learning communities than would ever be possible in a classroom.
• The best tool for some jobs is video technologies such as television, videotape and teleconferencing facilities. For example, the internet-based course mentioned above also involves lectures by the instructor and presentations by visiting speakers delivered via television to local audiences and via videotape to overseas. The primary change we envision is to raise the production quality beyond the existing budget-constrained "talking professorial head with presentation slides" production quality to approximate that of commercial television.
• The best tool for some jobs are highly interactive multimedia materials based on Authorware, Shockwave or similar competing products. These can be delivered via the internet or via CD disk when the volume of information to be exchanged (audio, video, etc.) exceeds the capacity of modem-based internet connections.
• We have also found that Folio Views and similar tools are highly effective at supporting collaborative close reading, critique and discussion of difficult texts. Folio Views features digital property protections that are sufficiently strong that publishers are quite willing to license entire books for delivery on CD or even over the web.

Specialized Roles and Responsibilities

Very few content experts have the technical skills to develop tasks unaided. The solution is an evolutionary course-development process in which content experts aren't expected to develop courses unaided as faculty does today. They are supported by experts such as these:

• Content experts develop courses as text files. They play a role analogous to the role script-writers play in movie production.
• Producers transform scripts into terms that course developers can understand, for example by (in collaboration with content experts) choosing delivery mechanisms (internet, audio, video, CD, computers). Responsibility for delivery mechanism choices resides with the producer. The content expert's role here is advisory.
• Developers specialize in the various media mentioned above. Examples are video producers, internet producers (graphic designers, computer programmers), audio producers, presentation slide designers.
• Instructors review task submissions, add comments, assign grades, and provide encouragement and support. They play a similar role to faculty once a course enters the deployment phase.
• Reviewers review each semester's results to spot what is working well and what not. They revise the material to build on successes and to update material that has grown obsolete.

Notice that, with creative attention to academic rewards issues, some of these roles could be played by students, especially in an academic course setting. For example, I have seriously considered using prior semester's graduates in the evaluation role for the next semester's class should the enrollment in the TTEF course grow beyond what I could handle unaided.

Such specialization of labor is rare only in academia and is quite the norm everywhere else. These activities might be organized like any traditional organization. Less traditional organizational structures will be entertained as we discover what works best.

• Sales and Marketing: responsible for identifying new commercial opportunities and designing new educational products to address market needs. Responsible for designing product wrappers (syllabi, marketing collateral's) and arranging their development and deployment. This function must not be underfunded as it typically is in universities: should be at least 30% of overall expenses.
• Product Delivery: Responsible for operational affairs; e.g. teaching. Quality is crucial here. The research and development departments are primarily to build robotic support systems (tasks, courses) that eliminate the unproductive (mechanical) parts of faculty's work load, freeing time to be spent on helping the student to succeed. The robotic support systems are not a substitute for strong faculty-student relationships but a means to that end. Faculty incentives are based on quantitative and qualitative (objective, subjective and intersubjective) measures of customer satisfaction. There is no such thing as "appointment without term" (tenure).
• Product Development: acquires the services of world-class content experts. Builds tasks (as distinct from courses) to be assembled into products that convey the subject matter to others. This is largely an engineering function, not an academic one. Content experts are hired on a limited term contractual basis, subject as always to experience with how this works in practice.
• Administration: Registers new customers, maintains records, and so forth. A clerical, nonmanagerial, function; heavily automated.
• Management: Provides coordination support services to the organization as a whole.
• Research: The state of educational technology is advancing so rapidly that we'd need proactive research just to keep up. The Hyperlab initiative by Peter Denning, Lewis Perelman, and Danny Menasce is analogous to MIT's MediaLab. Hyperlab's emphasis on web-based instruction and ubiquitous self-assessment could be quite relevant in this context.

Coordination Technology

In "Understanding Computers and Cognition; A New Foundation for Design", Winograd and Flores introduce the term "Coordination Technology" in connection with the work of Paul Cashman and Anatol Holt. I worked very closely with Holt at ITT, who used the term to mean almost the opposite of communication technology. I will explain that opposition here.

The technologies we associate with the internet today: email, the web, and conferencing tools such as Caucus or WebCrossing, are communication technologies, not coordination technologies. Their goal is to haul every byte to every participant without losing a single one. The goal is to empower individuals to communicate more effectively than they ever could before.

Although such empowerment of individuals has obvious advantages, I will emphasize the drawbacks here. Giving everyone a megaphone creates as many problems as it solves, as the proliferation of useless web pages and junk mail abundantly shows. By amplifying both signal (timely, relevant) and noise (untimely, irrelevant), the signal is lost in the noise.

Coordination technologies have the opposite goal, of amplifying signal alone. Of course, for computers to distinguish signal and noise, they must be given precise information about the organizational roles and responsibilities of each user and the organizational context within which they are acting. A coordination tool that expedites the yearly budget cycle, for example, would "know" the precise roles and responsibilities for each participant, the schedule, and the organizational process for budgeting. It would guide and schedule their goal-directed activities in a structured manner, leading to a defined result within a predefined schedule.

A mechanical analog is the spring-loaded wires and baskets that used to carry department store transactions whizzing between cashiers and back office clerks. The wires were organizational "pipes and fittings" that embodied the knowledge that cashiers and back office clerks needed to interact frequently. A communication tool (hiring an office boy to carry transactions from any point to any other) would have been more general. However generality wasn't the problem they needed to solve.

The following differences distinguish the two kinds of tools:

• Communication tools amplify both signal (timely, relevant) and noise (untimely, irrelevant). Coordination tools omit noise and amplify signal alone.
• Communication tools are general purpose. Coordination tools are special purpose. The latter cannot be written once and reused by diverse organizations. They must be highly customized, or rewritten, for different organizations.
• Communication tools are common. Coordination tools are rare. This is a corollary of the preceding point.
• Communication tools are unstructured. Coordination tools are structured. The former is better for exploration of new opportunities, the latter for action within existing opportunities.
• Communication tools are democratic. Coordination tools are authoritarian. They put someone in the privileged position of deciding what is timely and relevant for everyone else.

The point, obviously, isn't that coordination tools are "better" or "worse" than communication tools. Both have important roles to play. It is simply to point out that communication technologies are common and coordination technologies are rare. A consequence, of course, is the feeling that the signal is being swamped by the noise on the internet as we know it today.

Academic products (courses, degrees or certificates) obviously involve both kinds of technologies. Communication technologies range from email to web conferencing to face to face encounters in the halls and classrooms. Such activities are clearly an important part of an education.

However, there is also another part implied by the structured progression of scheduled events. Enrollment, task deadlines, timely completion of courses, and graduation are all structured schedule-driven events of the sort that fall well within the coordination tool umbrella.

For examples of efforts to balance these two approaches with respect to online courses, see http://virtualschool.edu/now under Taming the Electronic Frontier. The first page of an individual's locker presents only what's timely and relevant to the role they logged in as. For example, students are only presented with information about what they must do to be successful in this course this week.

Everything else is in the background, accessible via a hotlink. For example, the semester schedule is available via a hotlink with grades, upcoming lecture outlines, presentation slides and so forth. All this is available but in the background, there if its needed but irrelevant to the "what must I do now" decision. Several communication technologies, such as email, web conferencing, chat, telephones, are also accessible when they're needed but not as the main topic of interest.

Author: Brad Cox