TO: Instructional Support Committee
FROM: John Stone, Department of Mathematics and Computer Science
RE: Recent additions to the Academic Computer Use Policies
On July 21, 2003, the Director of Information Technology Services (ITS) announced several changes in the Academic Computer Use Policies that regulate student and faculty use of computers at Grinnell College. Some of the new rules limit the right of students to create computer programs and to share their creations. Others constrain the activities and publications of student organizations, interest groups, and scholarly collaborations, if the students who form them communicate primarily by computer networks instead of assembling in one physical location.
Section I.2.c. of Part Two of the Faculty Handbook says that the Instructional Support Committee shall ``assist in the development and review of policies regarding the use of computer software and hardware, the Internet, and the libraries.'' Section V.D. of the Academic Computer Use Policies says that ``if an individual disagrees with a college computing policy,'' and the disagreement is not resolved in discussion with the Director of Information Technology Services, ``the individual may notify the Chair of the Instructional Support Committee,'' who is empowered to investigate and review the concern.
In accordance with these provisions, I call the attention of the Instructional Support Committee to the new sections of the Academic Computer Use Policies, sections III.A.4 and III.A.5. I call for the rescission of these new sections on the grounds that they obstruct the educational mission of the College, placing unnecessary and undesirable constraints on the curriculum and undermining the College's core values of student self-governance and personal responsibility.
I also protest ITS's use of the new sections of the Academic Computer Use Policies to shut down Plans, an established and successful on-line community for Grinnell students, alumni, faculty, and staff. I object both to the fact of the shutdown and to the brutal and abrupt manner in which it was ordered.
I call upon the Instructional Support Committee to investigate the rationale for the addition of these new sections to the Academic Computer Use Policies; to study their effects on the curriculum of the College, on the development of student organizations, interest groups, and scholarly collaborations, and on Plans in particular; and to recommend to the Vice-President for Academic Affairs that the new sections be rescinded.
Let's look first at section III.A.4:
All students offering programs and applications hosted on a personal or group account must review and follow applicable principles of the ACM's ``Software Engineering Code of Ethics and Professional Practice'' http://www.acm.org/serving/se/code.htm. Students offering applications hosted on a personal or group account using a username and password login system have special responsibilities as ``software engineers'' and should take note of the following sections of the ACM Code of Ethics: 1.01 - 1.07, 3.03, 3.10, 3.12, 3.13, 3.14, 4.01, 5.12, 7.05.
The acronym `ACM' here refers to the Association for Computing Machinery, which is the leading professional organization for computer scientists, software engineers, and information technologists generally. Software developers who are also members of the ACM are expected, though not required, to observe the Software Engineering Code of Ethics, which also serves as a guide and a model of behavior for students training for careers in software development.
In the Department of Mathematics and Computer Science, we have until now regarded voluntary and intelligent observance of the Software Engineering Code of Ethics as a desired result of the education that computer-science majors receive at Grinnell, not as a prerequisite for any student who exercises her right to create new computer programs and to share them with the community. I should mention that this is a right that students exercise frequently: On MathLAN alone, there are currently about fifty thousand files in student accounts that are executable by other users. I do not know how many of the students offering these programs and applications have reviewed and followed the applicable principles of the ethics code, but I speculate that many of our users are in violation of this requirement. (I imagine that many of these programs use inaccurate data, are inadequately documented, inadequately tested, difficult for persons with disabilities to use, wasteful of resources, or non-standard.)
As written, the new language applies both to programs and applications that the user who shares them has written herself, and to those that she acquires from other students or by downloading them from the Internet and installing them. It thus imposes a somewhat arbitrary requirement on many students who have no background in computer science and are not likely to become software engineers.
Section III.A.4 also affects the curriculum in computer science and, probably, in other departments of the College.
In the past, my colleagues and I have often taken advantage of the fact that experimenting with software development and distribution, in the relatively safe environment that Grinnell College is able to provide, is a strong incentive to the formation of ethical thinking in software design. For instance, a student often learns the importance of good user-interface design partly by watching his fellows struggle with an interface that he has given them -- one that is poorly designed precisely because the student has failed to think through and anticipate the needs and interests of users. This learning experience is a violation of the Academic Computer Use Policies in their new form, which require even novice programmers to ``review and follow'' the principles laid down in the ethics code before exposing other users to their work.
Under section III.A.4, since lab partners and members of a programming team typically share the programs that they write, we should begin any course in which we ask students to work together on labs or team projects by familiarizing students with the ACM ``Software Engineering Code of Ethics and Professional Practice,'' so that they will be in a position to follow the applicable principles in offering their drafts to one another. Sometimes, as in my current course ``Artificial intelligence,'' this can be done without too much violence to the structure of the course. On the other hand, I find it an absurd constraint to impose on our introductory programming course, ``Fundamentals of computer science I,'' which is currently a lab course. Unless the constraint is lifted before I teach that course in the spring, I will be unable to use the workshop method around which we currently organize the course. It is simply not possible to begin an introductory course with the techniques that professional software engineers use to ensure the correctness, robustness, and security of their programs.
I cannot say how widespread the practice of sharing software among students is in other departments, nor how important shared software is in their course work. However, I know that it is common in many departments, especially when the category of software is taken (as it is in the immediately preceding section, III.A.3, of the Academic Computer Use Policies) to include spreadsheet and word-processing macros, document templates, and even image files. The ACM Software Engineering Code of Ethics requires its adherents to test and document any Excel templates and HTML markup that they create or approve for use, just as it requires them to test and document Java applications.
The rationale for section III.A.4 is apparently to prevent the use of programs that are poorly designed, poorly tested, poorly documented, insecure, unsafe, or contrary to the public good. I fully endorse these goals, but I question whether it is necessary or even possible to impose them on Grinnell College computer users through the Academic Computer Use Policies. As I have argued, there are legitimate educational reasons for sometimes executing programs that do not meet professional standards. And programs that harm people, violate state or federal laws, or consume excessive resources are already prohibited by other sections of the Academic Computer Use Policies (III.B).
Here is the other section added on July 21:
Students developing applications on College systems requiring individual login for membership and offering membership to the broader Grinnell College community have more stringent rules to follow. In the case of virtual communities* based at Grinnell College, Grinnell software authors must ensure compliance with College policies regarding accepted student community standards in addition to standards of academic computer use. For example, one of the ``Core Values of Grinnell College'' as stated in the Grinnell College catalog is that the College encourages ``personal, egalitarian, and respectful interactions among all members of the community.'' Accordingly, Grinnell software authors must publish clear, complete, and publicly available guidelines detailing acceptable behavior & membership policies for the virtual community. Due process must be ensured before any sanctions of members are carried out. The software authors must make provision for appropriate Grinnell College staff to ha ve full access to the virtual community. Communities or individuals in violation of the Academic Computer Use Policies or standards of student community at Grinnell College are subject to immediate disconnection from the campus network pending completion of review procedures by Information Technology Services or Student Affairs, as appropriate Accordingly, a system of logging of the community entries on a nightly basis must be in place in order to establish an evidential base for proper hearings of complaints brought by students, faculty or staff to appropriate College committees or boards. The content of any student or student group virtual community must not be viewable by non-members of that community, either directly or via internet search engines. The members of student-created virtual communities must be current students, faculty, staff. or Trustees, and members of these groups cannot be excluded from a virtual community on an a priori basis.
A new endnote attached to the Academic Computer Use Policies explains that the phrase ``virtual communities'' first appeared in Howard Rheingold's 1993 book The virtual community: homesteading on the electronic frontier (Reading, Massachusetts: Addison-Wesley, 1993). Rheingold applies it to any community in which the participants communicate with one another primarily by means of computer networks and rely on the infrastructure supplied by such networks as they share and develop their interests, policies, and institutions. Specifically, he applies it to on-line groups defined by common interests or hobbies (parenting, jazz, movies, boating), to political interest groups (environmentalists, civil-liberties advocates), and to research teams.
The intention and effect of section III.A.5 is to prevent students from forming autonomous on-line communities.
To begin with, under section III.A.5, an on-line community organized by students is not permitted to admit alumni, students at other colleges and universities, local townspeople, prospective students, friends and relatives of current students, or indeed anyone except current Grinnell College students, faculty, staff, and trustees. I am unable to comprehend the point of view from which this result could be considered advantageous to the College. It patently violates our students' freedom of association.
One consequence that is particularly relevant to our curriculum, which encourages students to participate in summer research programs, is that students at Grinnell cannot form and host intercollegiate research groups.
Perhaps the rationale for this restriction is that persons who are not students, faculty, staff, or Trustees of the College are not bound by the Academic Computer Use Policies and cannot be punished by the College for violations. I note, however, that many other colleges and universities seem to be able to accommodate on-line communities that are open to outsiders. These colleges and universities include many of Grinnell's peer institutions (Swarthmore, Bowdoin, Amherst, etc.) and many that are reputed to be knowledgeable about legal obligations (Cornell, Harvard Law School).
Secondly, under section III.A.5, on-line student organizations may not publish their communications to non-members or allow Internet search engines to index them. This requirement prohibits many useful and beneficial student activities. For instance, my colleague Sam Rebelsky has several student research teams working with him this summer. Each research team is a virtual community under Rheingold's definition. In the past, such teams have generally published some of their results as Web pages, and some have made their software available through the department's CVS (Concurrent Versions System) server. Non-members of the research team -- including other Grinnell College students, students and faculty at other colleges and universities, friends and relatives of the researchers, and indeed anyone anywhere on the Internet -- can view the papers and download the software. This pattern is typical of research in computer science in the age of global communication. But section III.A.5 prohibits student-run on-line communities from following it.
The presence of this clause in the Academic Computer Use Policies damages Grinnell College's reputation as an advocate and supporter of free speech and free inquiry. Whether an on-line student organization publishes its communications to outsiders, and where, when, and in what medium they shall be published, are matters for the organization itself to decide, not ITS.
Thirdly, section III.A.5 requires student authors of software supporting on-line communities to give ``full access'' to the virtual community to Grinnell College staff members. The Director of ITS understands ``full access'' to include not only the access enjoyed by an ordinary user of the software or member of the community, but also, specifically, the powers of administration and control over the software and any associated files.
Similarly, the new section requires student software authors to police the on-line communities that their software supports, ensuring compliance with College policies, imposing sanctions for violations of the ``core values'' of the College, and recording and archiving the communications of members of the community every twenty-four hours in order to ``establish an evidential base for proper hearings of complaints.''
Again, in my opinion, a student community has the right to decide for itself which records of its activities and deliberations it makes and preserves, how long such records will be kept, and what purposes the community will make of them. It is arguable that the appropriate College committees and boards have a right to subpoena such records when they exist, although in many cases it would seem prudent and even imperative to recognize analogues (appropriate to private colleges, of course) of the First, Fourth, and Fifth Amendments to the Constitution of the United States. However, it is not ITS's business to dictate to student organizations whether and how often they should back up their files.
An organization that does not control its own membership, its own records, or its own publications can hardly be described as ``self-governing.'' I conclude that ITS designed section III.A.5 to force student organizations using College computers either to allow themselves to be managed by ITS or to shut down or move off campus. Section III.A.5 is a direct attack on student self-governance.
Like most of my colleagues, I see self-governance as an important element of the educational mission of the College. It develops students' sense of social and political responsibility, gives them confidence and valuable experience in practical life, and provides vivid illustrations of many of the ideas and principles that we encounter on a more abstract level in the classroom. We computer scientists take particular pride in the successful development and widespread use of the Plans software, which demonstrates the initiative, independence, and skill of our students.
Plans was initially written by a computer-science major, Rachel Heck 2001, in response to a complaint voiced by many students in the 1999-2000 academic year. When academic computing at the College made the transition from a central VAX cluster to a distributed network of personal computers, one of the services that was lost was the FINGER program, a kind of precursor of Plans, which allowed users to read the rants, stories, messages, journals, and other kinds of texts that other users had posted in specially named files in their accounts. At that time, ITS had no interest in virtual communities and did not consider it important to provide a replacement for FINGER. Ms. Heck realized that she could meet this need, using the World Wide Web as the medium, and proceeded to implement Plans.
This was a private project for which she received neither course credit nor remuneration from the College. It was not supervised or approved by faculty or staff members, nor was it supported by the Student Government Association. It was a pure instance of altruism and social responsibility. (However, Ms. Heck did ultimately receive some official recognition for her work on Plans, among other things: She was the recipient of the Robert N. Noyce Senior Student Award in 2001.)
Initially, Plans had only a few users and was only slightly more powerful than the FINGER program had been. However, it rapidly became much more popular, occasioning some changes in its design and functionality. As Ms. Heck reworked it, she added new features, often in response to suggestions from the user community. When she graduated, she turned the responsibility for maintaining and extending the program over to three other students, one of whom (Jonathan Kensler 2004) has since done most of the work on the software. Mr. Kensler was also the nominal owner of the database containing the texts of users' Plans.
Plans served as a model for closely similar implementations of similar on-line communities at North Carolina State University and at Amherst College. The source code for Plans has been released as free software under the GNU General Public License, so that programmers everywhere can study and build on the work of Ms. Heck and Mr. Kensler.
At Grinnell, Plans was (and is) immensely popular. There are about 1600 Plans accounts -- about 1100 students and student organizations, four hundred alumni, and a few dozen faculty members, administrators, and staff members.
In speaking with prospective students, faculty and students at other colleges and universities, and various people who just want to know what Grinnell is like and what distinguishes it from other small liberal-arts colleges, I have often used Plans as my leading example of the College's values. In telling the history of Plans, I portrayed the College as an environment that fostered creative student initiatives and our students as insightful and socially responsible activists, eager to put their mastery of their discipline to constructive use.
Well, I won't be telling that history to outsiders any longer. At 3 p.m. on July 21, the Director of ITS presented Mr. Kensler with a copy of the revised Academic Computer Use Policies and a letter pointing out that Plans was in violation of them -- ``and so it must be taken off-line by 9:00 p.m. today and remain off-line.'' Mr. Kensler took Plans off-line at 8:46 that evening. He spent the next ten days finding a more reliable Internet service provider to host Plans, moving the software and the database, adapting the software to the new site, and testing it in its new environment. Plans came back to life at 2:47 a.m. on August 2, as http://www.grinnellplans.com, and has now resumed its role as the leading on-line community for Grinnell students and alumni. Functionally, it's essentially the same as before. Now, however, the College has to live with the ignominy of having tried to extinguish it.
Since it is not possible to terminate a service like Plans gracefully on six hours' notice when it has so many users, the Plans shutdown took many users by surprise and provoked a storm of outrage. I have reproduced some of the protests on a series of ``Plans comments'' Web pages, and I urge the Instructional Support Committee to read and consider them as it considers what to do about section III.A.5 (see http://www.cs.grinnell.edu/~stone/plans-archive/plans-comments.xhtml).
The Director of ITS offered the following rationale for the Plans shutdown:
VCs [virtual communities] are subject to abuse if they are not set within a system of clear rules and policies. ``Plans'' does not have a strong technical or policy framework. Consequently, the ``Plans'' VC has been subject to a abuses of College computer policy, violations of state laws, and transgressions of the values of respect and responsibility that are inherit in the College's mission and culture.
In his letter to Mr. Kensler, the Director enumerated these abuses, violations, and transgressions, as follows:
Everyone involved agrees that items 1 and 2 are indeed violations of the Academic Computer Use Policies, and specifically of section III.B.3:
Grinnell College's computing resources may not be used for any activities which violate state or federal laws. Computing resources may not be used to intimidate, threaten or harass individuals, or violate the college's policies concerning relationships between college constituencies. Such activities include, but are not limited to, using computing resources to store, print, or send obscene, slanderous, or threatening messages.
However, everyone also agrees that the violators in these cases were individual users of the Plans software, not Mr. Kensler. Indeed, in the few cases with which I am familiar, Mr. Kensler acted fairly and promptly to resolve the violations that were brought to his attention, usually by removing the offending passage or persuading the author to do so, and he was willing to work with ITS in handling similar cases in the future. It was not necessary to shut down Plans in order to deal effectively with offenses committed by users of the software, nor was it necessary to add anything to section III.B.3 in order to provide grounds for seeking Mr. Kensler's cooperation.
What provoked the shutdown of Plans was not, therefore, the misbehavior of the users of Plans, but the fact that ITS had to go through Mr. Kensler in order to deal with them. ITS prefers to have direct control over the services provided on College computing equipment, so that it can enforce the Academic Computer Use Policies without having to ask for anyone else's assistance. Convenient as it might be, however, ITS is not really entitled to this measure of control over services that are designed, implemented, popularized, and maintained by students. The Plans shutdown and the addition of section III.A.5 are just ways for ITS to make its work easier at the expense of student self-governance.
The security problems with Plans, which date back to early in the spring semester of 2003, were more troubling, because they reflected real design problems in the software rather than casual misuse. The most straightforward way to resolve them was to move Plans from the MathLAN Web server, which is an ``open'' machine on which anyone with a MathLAN account can log in, to a server of its own, accessible only by the Plans maintainers. It took several months to obtain the use of a computer for this specialized purpose, but the Plans maintainers did finally get a server, and Plans began running on that server on July 2 -- about three weeks before the shutdown. Users of Plans were, of course, aware of this move, but the Director of ITS apparently was not.
Finally, as to the ``arbitrary and opaque administrative processes for determining who cannot remain a member of Plans,'' I wish to point out that Plans never had any official status at the College, never even registered as a student organization with the SGA Student Group Liaison, and never received any funding or organizational assistance from SGA. I think of it as a large garden party for which Mr. Kensler has kindly agreed to act as the host. Having provided amenities and refreshments for as many members of the College community as care to attend, he has opened the gates and allows us to circulate and speak freely, so long as we behave ourselves. On the other hand, I recognize that it is within his discretion to bounce anyone whom he considers a troublemaker. I do not think that he is obligated to formulate and post an elaborate set of guidelines for his guests to follow or to establish ``administrative processes'' for removing anyone who abuses his hospitality. In my opinion, a lot of muddled thinking on this issue arises from a common but deplorable tendency to assume that any software that is widely used at the College is officially supported by the College. This is simply not the case.
None of these alleged problems adequately explains or justifies the decision to add new language to the Academic Computer Use Policies without consulting the Instructional Support Committee, the decision to make the new policies effective immediately (effective, in fact, several hours before they were published to the College community at the ITS Web site), or the decision to give the alleged offender, Mr. Kensler, such a brief and obviously inadequate interval to shut down his service. He was given no opportunity to bring Plans into compliance with the new policies, but was contemptuously told to get out and stay out (``it must be taken off-line by 9:00 p.m. today and remain off-line'').
Nothing in this entire ugly episode gives one any confidence in ITS as a maker and enforcer of laws for on-line student communities. On the contrary, ITS has shown itself to be unfair, disrespectful, arrogant, power-hungry, thoughtless, and ill-informed. ITS has offended and alienated the most active and influential users of virtual-community software among current students and recent alumni. Seven hundred and eighty alumni, parents, and students have signed a petition in which they promise to withhold donations from the College for at least one year, pledging the money instead to a fund to support Plans on a non-College server.
Under these circumstances, I think that the best thing the College can do is simply to remove the new language from the Academic Computer Use Policies. It's too late now to bring Plans back to campus; the current maintainers regard Grinnell College as an unreliable Internet service provider, and I expect that we will retain that reputation for several more years no matter what we do at this point. However, it may still be possible to prevent ITS from strangling future student on-line communities at birth. Rescinding sections III.A.4 and III.A.5 would be a good first step.
The Director of ITS has created an Advisory Committee on Virtual Community Issues, and I understand that the Instructional Support Committee may want to await the report of that group before committing itself on section III.A.5. However, the charge to the Advisory Committee does not include either an investigation of the Plans shutdown or a review of section III.A.4. I suggest that the Instructional Support Committee begin at once to deal with these parts of my complaint.
If any member of the Instructional Support Committee wishes to discuss these issues with me further, I am willing. I also invite you to examine the documents at my Web sites http://www.cs.grinnell.edu/~stone/plans-archive/ and http://www.cs.grinnell.edu/~stone/acup/.
[signature: John David Stone]
The Department of Mathematics and Computer Science likewise protests the use of the new sections of the Academic Computer Use Policies to shut down Plans and calls upon the Instructional Support Committee to recommend to the Vice-President for Academic Affairs that the new sections be rescinded.
[signature: A. Royce Wolf]
[signature: Christopher French]
[signature: Marc Chamberland]
[signature: David Bishop]
[signature: Samuel A. Rebelsky]
[signature: Karen L. Shuman]
[signature: Shonda Kuiper]
[signature: Charles H. Jepsen]
[signature: Henry Walker]
[signature: Pamela A. Ferguson]
created September 9, 2003
last revised September 10, 2003