Using the WWW for a Team-Based Engineering Design Class

Jack Hong, George Toye, Larry Leifer
Design Division, Department of Mechanical Engineering, Stanford University

Abstract

ME210: Mechatronic Systems Design, is a graduate-level engineering design class offered by Stanford University's Department of Mechanical Engineering. In Autumn '93, the World-Wide Web was first incorporated into the curriculum of ME210 in service of project team formation. Based on lessons learned from this pilot experience, we introduce a "breakdown"-based model of WWW-services migration in the classroom, as well as a new set of Web services developed to enhance the educational mission of ME210.

Introduction

Each year, ME210 provides an environment for 15+ three-person teams to tackle a wide range of industry sponsored projects. The educational goal of the course is to create a shared learning experience in collaborative mechatronic product design and development. Student teams benefit directly from working with a $10,000 design budget on their sponsored project, and indirectly from the design experiences of peer project teams. The deliverable for each team is "living documentation" of the sponsored product, its design, fabrication and testing. Delivery requires the development of a social framework for team management. Thus, students are given the experience of working in several ad hoc teams on mini-projects prior to selecting a team for the year.

In 1992, the SHARE Project, a collaboration between Stanford's Center for Design Research and Enterprise Integration Technologies, started using ME210 as a testbed for internet-based design collaboration services [1]. The most visible outcome of this collaboration can be found in the Web pages for ME210, also known to the class as the 210 Web. This Web site contains complete, WAIS searchable student design reports from the past two years, slide presentations, local vendor directories, past student profiles, and the full course syllabus.

As an archive of ME210's collective project knowledge, the 210 Web is being actively incorporated into this year's curriculum. The dynamic nature of the class, however, calls for the 210 Web to be more than an library of past knowledge, but also a sifting ground for "live" information. Given that Web server authoring and maintenance is still more complex than desktop publishing, current teaching and administrative staff are ill-equipped to deal with Web maintenance. To deal with this problem, we are developing community-serviceable interfaces to the 210 Web to help with the transition. In what follows, we provide background for our efforts by describing lessons gleaned from a pilot Web experiment, and giving a preview of new services planned for the 210 Web.

Pilot Web Experience

While students of ME210 are generally not computer-naive, computers are considered tools in service of design. Little time will be invested in tools with no perceived benefit to design tasks. When new computer tools were introduced into ME210 in the preceding year, we found that the learning curve for these tools competed with time students would rather have spent on completing the end-quarter report. Consequently, students reverted to less-complex tools, which they already knew how to use [2].

In September '93, the SHARE project decided to explore using the WWW in ME210. From the preceding year's experience, it was clear that students would have to consider the Web a worthwhile investment to even try it, especially since we were not allowed to grade students on research related activities. Since this was something very new to the engineering community, we decided to look for a context early in the quarter that would deliver high returns on student time invested, require minimal overhead to set up, and be aligned with educational goals. Such an opportunity was found in the team formation process.

One of the major experiences in 210 is how to design in teams. Finding the "right" people to work with from 50 or so near-perfect strangers, however, can be difficult. While 210 helps reduce some degrees of freedom via cognitive testing [3], the basic stance on team formation is laissez-faire. Using the Web to provide additional information about peers fit our requirements for a pilot experiment.

Halfway through the Autumn Quarter of '93, a pilot experiment was conducted using the WWW as a support tool for team formation. At the time, the only Web client available for the Macintosh was NCSA's Mosaic for the Macintosh B6, which did not support forms. Consequently, the exercise was limited to using the Web as a read-only medium. Students were asked to submit "self-marketing pitches" to a single coordinator via e-mail. Each "pitch" was to be fashioned after a resume, to answer the basic question, "Why would you want me on your team." After all pitches had been received, they were put on the Web for perusal and cross indexed by name and photo. An additional index was provided which grouped students by cognitive styles, as revealed by an in-class questionnaire.

Findings

Students who tried the technology immediately revealed unforeseen instabilities in the client software, server software and campus networking; some even questioned whether paper should have been used to serve the same goal. A survey was administered on the day teams proposals were due to find out how people had used the system and found positive responses from 70% of the students. Relative to the preceding year, students were better prepared to form teams by the deadline and required less last-minute match-making from the staff. The effort expended to make this gain behind the scenes, however, was enormous. Upon reviewing problems that occurred during this pilot test, a "breakdown" or "fuse" model of WWW Migration was constructed to better characterize our experiences.

Figure 1. Fuse Model of WWW Migration in the Classroom

In this model, the traditional channels of communication between teacher and student are being replaced by a computer-supported channel mediated by researchers. While this new channel is being developed, there are many opportunities for failures, or breakdowns, to occur. These breakdowns are depicted as fuses in the chain. In the pilot test, lack of communication between the researcher and teaching staff was one type of breakdown, while software and hardware problems caused breakdowns between the Web and the student. Regardless of where the breakdown occurred, students saw a less-than-functional system, and resorted to old channels of communication. At some point in the past, the old channel of communication may also have been laden with failure modes, but as teacher and students became accustomed to their mode of interaction, the channel became more robust.

New Services for '94

In order to construct a WWW-based channel of communication between teacher and students as easy to use as "analog" channels, we must remove the "researcher note" in Figure 1, as well as potential failure modes in this channel. We decided that the best way to eliminate potential failures was to hide unnecessary steps from the user. Three new services for the 210 Web are described here.

1. The PENS Web Authoring Client

PENS (Personal Electronic Notebook with Sharing) is an off-line authoring client for the Web, which simplifies posting of Web information for both students and staff. It is currently being used by the teaching staff for curriculum development notes and for posting FAQs to the 210 Web. It is slated for student usage in team mini-projects. Technically, PENS is a HyperCard-based notebook database (see Figure 2) which sends a MIME-encoded message via SMTP to the 210 Web, where the message is received by ServiceMail(TM) and placed in a predetermined HTML directory. PENS eliminates the need for direct Web server write access to edit HTML documents, and eliminates the need for HTML markup by incorporating limited style parsing with pre-structured document organization schemes. The user only needs to pay attention to note content, as with any conventional word processor, and click the "SEND" button to post contents to the 210 Web.

2. WWW-Based Class Registration

To familiarize students with the Web from day one, and simplify routine administrative tasks, the first class assignment asks students to register for ME210 using the Web. The 210 Registration Program is accessible from the 210 Home Page and only active during the first week of classes. It uses a forms-interface to collect the following items:

Name and Contact Information
Learning Objectives Statement
Personal Web account name and password.

Items 1-2 are assembled into a home page for each student, while Item 4 is used to set up accounts for the 210 Web, class e-mail server, and class AppleShare server. Upon submitting the form, all accounts are set up, and an "Edit" button is added to the student's home page. By clicking on the Edit button, and entering appropriate account information, each student can edit their home page information as well as access the 210 Sounding Board, described below.

3. The 210 Sounding Board

Because 210 students learn through immersion and interaction with peers, experiences vary and are difficult to measure by traditional lecture standards. Assessment of design learning tends to be a time consuming task, making it hard to give or receive frequent feedback. The "210 Sounding Board" is designed to improve the current channels of communication by automating administration and tabulation of questionnaires via the Web.

Students access the SB through the personal portion of their home pages, which requires password access. This user information is automatically carried over to the SB Interface, such that subsequent questionnaire responses are all tagged with a user ID. The SB Interface shows a listing of all questionnaires, past and present. Clicking on the title of an "active" questionnaire leads to an interactive form which can be answered in installments, any time before the "tabulation" date. After the tabulation date, the questionnaire becomes "expired" and tabulated results can be obtained by clicking on the "outcome" button, adjacent to the questionnaire title. Clicking on the title of an expired questionnaire pulls up a read-only version of that form for review, complete with the student's old responses.

A Questionnaire Construction Kit provides staff with the ability to customize a variety of questionnaires for SB. Applications range from simply voting on the time to hold a tutorial session to rating the effectiveness of a particular guest speaker. Two applications are of particular importance and described in more detail below.

Team Placement Questionnaire

One of 210's tenets is that design thrives on diversity. To realize this, project team placement is assisted through a cognitive preference questionnaire derived from the Meyers-Briggs personality test [3]. To give students a better appreciation for individual differences, four versions of the questionnaire are available through SB, to be taken preceding each group design exercise. By correlating group indices with design performance across groups, exercises and time, the whole class can share experiences on how to better work together and maximize individual assets. Without SB, administering and tabulating results of questionnaires would be very time consuming, making it a chore to reach the same level of understanding.

Class Feedback Questionnaire

Classes at Stanford are usually on a very tight schedule, and 210 is no exception to the rule. This makes it difficult for the teaching team to get feedback from the students except at the end of each quarter. By this time however, any constructive feedback can only yield benefits to next year's class. As long as the questionnaires are short, SB can be used to collect frequent feedback from students, allowing the teaching team to track teaching effectiveness in real time and make timely path adjustments where needed.

Closure

We have presented key components behind the ME210 Web, with an attention towards services which are serviceable by non-HTML literate students and staff. These tools and services are continuously being developed and evaluated in the classroom environment for constant refinement. As an added incentive to make sure these services work, we've taken on the additional challenge of distance education. This year, ME210 will be be broadcast via the Stanford Instructional Television Network (SITN) for the first time, resulting in an anticipated 20% of the class attending remotely. It is expected that the these new Web services will enable remote students to share a sense of community unattainable by video alone.

Acknowledgements

Credit goes to Jay Glicksman of EIT for helping set up our first HTTP server and making it work, and Jeff Aldrich for making sure it continues to work.

References

[1] Toye, G. et al. "SHARE: A Methodology and Environment for Collaborative Product Development," IEEE Proc. of Second Workshop on Enabling Technologies: Infrastructure for Collaborative Enterprises, Morgantown, April 1993, pp. 33-47.

[2] Hong, J. "ME210 Focus Group Transcript of June 11, 1993," Center for Design Research, Stanford University.

[3] Wilde, D. "Mathematical Resolution of MBTI Creativity Data into Personality Type Components," ASME Design Theory and Methodology DE-Vol 53. 1993, pp. 37-43.

Authors

Jack Hong is a Ph.D. candidate in the Design Division of Mechanical Engineering at Stanford University. He holds an M.S. in Product Design Engineering from Stanford, a B.S. in Mechanical Engineering (Honors) and a B.S. in Electrical Engineering from the University of Houston at University Park. His current research interests are in collaborative design tools,interdisciplinary design team dynamics and human-computer interaction.

George Toye, Ph.D. has a background in mechatronics and redundant control systems for complex computer-based systems. He is currently the associate director of the Stanford Center for Design Research, and is centrally involved in the center's study of engineering design activities, processes, and interactive learning. His current research activities include the ARPA funded project ("SHARE") to develop computer network based technology in support of collaborative design and manufacturing activities via the Internet, and the NSF's National Engineering Education Delivery System ("NEEDS") project, in which courseware modules can be found and distributed via the Internet.

Larry Leifer's formal academic training was obtained at Stanford University. He holds a Bachelor of Science degree in Mechanical Engineering (1962), a Master of Science degree in Product Design (1963) and a Ph.D. in Biomedical Engineering (1969). He has published in the areas of diagnostic electrophysiology, functional assessment of voluntary movement, human operator information processing, rehabilitation robotics, design team protocol analysis, design knowledge capture and concurrent engineering.

E-mail Correspondence:
hong@cdr.stanford.edu