An Open Letter to the IT Review board at the University of Waterloo

To:

David Barnard, President, University of Regina;

Larry Symes, University of Regina, and

John Mather, Manulife Financial;

CC:

Imprint Publications, University of Waterloo,

The Gazette, University of Waterloo,

uwstudent.org,

Amit Chakma, Vice-President, Academic and Provost, University of Waterloo, and

David Johnston, President, University of Waterloo.

Sujeet Chaudhuri, Dean of Engineering,

Alan George, Dean of Mathematics,

Jay Black, Associate Provost, IST, and

the student members of the University Senate.

My name is Simon Woodside. I appreciate the opportunity to address you in your IT review of the University of Waterloo. I consider myself well qualified to comment on the following issue regarding your review: if you're interested in my background, I have summarized it at the end of the letter[1].

I am writing to you today because I believe that the university would benefit substantially from improvements to the way that IST works with students.

As you know, IST (Information Systems and Technology) is the IT department at Waterloo. They implement and manage ("own") all campus-wide systems, co-op, and faculty systems except in Math and Engineering. Those two faculties have their own (excellent) IT departments.

My experience with IST/student interaction culminated in the summer of 2000, at a meeting with the director of IST and the deans of Math and Engineering. But my account begins about eight months earlier, in the fall of 1999. At that time, many students were dissatisfied with the system used to access Co-op Education job listings and schedules, called "Access". A project to replace Access, called CECS.Online, was scuttled by the contractor (ASI) in the United States in a complicated series of dot-com takeovers, leaving Co-op without a system and the Y2K deadline looming with Access officially non-compliant. At this point many students began to wonder whether they could assist IST in developing a replacement. I co-founded a group with this express purpose — to develop an open-source replacement — called Open CECS Online (OCO). We approached IST and began a series of meetings with their staff.

Initially IST was highly resistant to the plan, due to their stated misgivings about student volunteers. Among their concerns were future support issues, the privacy of the information in the student database, and the reliability of student volunteers[2]. We felt that all of these concerns were valid, but not blocking issues. However, the talks moved slowly and I believed at the time (and still believe) that IST was not fully committed to any student volunteers working on their projects in any capacity.

I should note that IST stated at times that students were involved in their projects. However, this is not the case, as the people they referred to as students were in fact co-op students engaged with IST as employees during their co-op terms. Student volunteers are free to pursue their own interests, speak out publicly and activate when they are unsatisfied. Although I recruited and coordinated dozens of student volunteers over the years, none of them were ever employees of IST.

In any case when I returned to campus in the summer term of 2000, no significant progress had been made. This is despite the work of George Roter, a member of OCO: then a graduate student who founded and is now president and CEO of Engineers Without Borders (www.ewb-isf.org). George is also an extremely competent and well-spoken individual. He attempted to develop a framework for working together with Jay Black, the director of IST. The IST proposal presented on March 17, 2000, for example, included the following.

Constraints: UW is prepared to assume some additional risk on the project by involving students more extensively than before. ... This suggests that an employer-employee relationship will be necessary for most significant contributions.

The framework was never completed and there did not seem to be practical interest from IST in getting to work. Meetings were scheduled with weeks of delay, and generally I felt from the level of bureaucratic issues that IST was not genuinely interested. This combined with my experience in the fall led me generally to feel that we were being stonewalled. This contrasted with our discussions with Co-op who I felt were interested in engaging and moving forward.

At some point I talked directly with Co-op staff and asked if we could work with them directly. The blocking issue was that IST owned all of Co-op's computer systems; thus I could not work on any of them until IST approved. Even a system that would only interface in a limited way with an IST system was blocked. I suggested that we could develop a system wholly independent from IST, and as you will see, this is what eventually happened.

General student opinion about Access, the Co-op computer system, was very low. The job market was very hot, at the peak of the dot-com boom, and Co-op started to post new lists of jobs every day. Due to various constraints, each posting was only available for 10 hours, and once a maximum of 90 users was reached, the system could no longer handle any further connections. It is difficult to imagine, but there were over three thousand co-op students actively looking for jobs, and less than half were able to access Access during that 10 hour interval, assuming that they all spent the entire day trying (and clearly that is not the case).

The user limit, technologically speaking, was a legacy of the system. Access was a command-line system developed between 1988 and 1994, very basic and in the words of the director of Co-op (I paraphrase) "already out of date when it was introduced". It kept an open connection to the database for each user that was logged in, a relatively antique technique for managing user sessions. Modern web-based systems keep the database connection only briefly to download the data, and the result is a much more economical use of the system resources. Some computer science and computer engineering students, using the system, were aware of the new techniques as they were widely available even in open-source software by that time. You can imagine their frustration with Access. As I told the student newspaper, many students were eager to help fix the system.

In May a student who I didn't know well, Shandy Brown, took matters in his own hand and wrote a program that used a technique referred to by programmers pejoratively as "screen-scraping" to download all of the data from Access to his own system. He then wrote a web interface to the data and offered it to other students from his own web server theoretic.net. Needless to say, word of Shandy's hack spread rapidly -- and students immediately and enthusiastically used it. Shandy's system remained online during the entire first round of job postings. Most astonishing of all, he wrote the entire system in a single night of coding, and published the code on his site as open source. In one night a single student accomplished what OCO had been trying to do for months — Access was on the web!

I have two notes to make at this time in my letter. First, IST had told us in our fall 1999 meetings that they had their own Web Access system in development and it would be deployed in the winter. I believed them, but in retrospect this estimate was not met. Also, open-source is tangentially mentioned in this letter many times, and we all take it for granted today. But at the time, open-source was still relatively obscure outside the computer community. Today, highly-visible projects such as GNU/Linux, the operating system that competes with Windows; Mozilla, the web browser from Netscape deployed to AOL users; and the Darwin project that I helped initiate at Apple have helped to bring knowledge of open-source to the general public. Open-source projects are predominantly volunteer efforts that succeed or fail on the basis of an open, public, collaborative method of development.

Going back to the events of summer 2000, one might reasonably expect that IST would embrace and adopt Shandy Brown's system since it accomplished the most immediate goal — it broke through the user limit and gave students web access to the job postings. Or perhaps, at this point in my letter, you will not be surprised to hear that IST's response was less favorable. It was in fact negative. Anticipating this, I quickly contacted Shandy and asked him to join OCO, and he agreed. This was a purely political maneuver on my part, although Shandy was also an excellent programmer, and we already had a dozen or so members who were ready to write code. I intended to use the opportunity to force our problems with IST into the spotlight. OCO also corrected a flaw in Shandy's code to restrict access to UW computers. The whole story was reported in the student paper.

In response, Jay Black wrote Shandy and OCO a formal letter requesting the site be removed ex post facto. In it he said something that frightened me at the time: "I've copied the Deans of Mathematics and Engineering since I believe all the students involved are registered in those faculties." George Roter however did not seem concerned.

The conflict culminated in an early June meeting between myself and George Roter, representing OCO, Jay Black and Dave Kibble, representing IST, and the deans of Math and Engineering, Alan George and Sujeet Chaudhuri. The deans were invited by Jay Black, who opened the meeting by commenting that (I paraphrase) "It's impossible to get any action without involving the deans". I believe that he was referring to disciplinary action, although I cannot remember if he stated this outright.

The deans however were not interested in discussing discipline. They both proceeded to suggest that student involvement was a thing to foster, and that IST should work together with the students and OCO in particular. The meeting, which I believed would be very negative towards OCO, went in a completely opposite direction. It ended without any further mention of discipline.

This was the high point of my interaction with IST, and the highlight of my letter to you. After the meeting, IST paid lip service to working with students, but to my knowledge no partnership ever resulted. OCO and IST met again, but I, at the helm of OCO, at that point refocused on working directly with Co-op. Working with Co-op, OCO developed a system for students to post their resumes online. The system, called Précis, used in production in the fall of 2000 for the continuous phase of co-op. It was completely independent of any IST systems, as it covered an area that was never previously automated, so there were no pre-existing systems that needed interface. In effect, I would say that in developing this project with Co-op, I performed an "end-run" around IST since I no longer had any serious hope of working with them.

This is the first time I have publicly disclosed the events of the meeting with Jay Black, Dave Kibble and the deans. I felt at the time that further criticism of IST would be impolitic and only harm OCO's chances of working with Co-op. Now that I am an alumnus, and no longer associated with any student group at Waterloo, I do not feel so constrained.

Thus, I also will give some speculation as to the source, internally, of IST's resistance to working with students. At first it seemed inexplicable to me, but I had many meetings with IST, opportunities to observe their reactions, and also I spoke in confidence with students who worked on their co-op workterms for IST. I do not fully understand IST's internal politics, however it is my experience that resistance of this type is sometimes the result of internal opposition that is kept internal and never disclosed publicly. Certainly IST was never forthcoming in explaining their actions or decision-making process. However I would not wish to be accused of implying that this is the case with Jay Black, since as a professor he worked and probably still works cooperatively with graduate students.

Fundamentally I want IST to work cooperatively with students. The students at Waterloo are extremely bright and practical. Waterloo is considered by many the finest Computer Science and Computer Engineering school in Canada. If the situation today is as I found it in 2000, then I hold it as a mark of shame on my degree and a detriment to the university community, present and future students, the clients of IST systems, and to IST staff who might not only help to advance students exploration of technology but possibly learn something from enterprising students at the same time.

I want IST to be more open. I want them to discuss their systems readily with the UW community, and be open to discussion and debate on their practices, policies and implementations.

I hope that I may leave the implementation in your hands. I have presented just one idea that students wanted to work towards. There are most likely as many ideas as there are students, and some of them could flower into enriching projects, given the chance.

Sincerely,

Simon Woodside

sbwoodside@yahoo.com

Postscript: wireless networks, a current effort

During the writing of this letter, I solicited comments from the Waterloo Wireless group by posting on the uw.wireless newsgroup. I have a particular interest in wireless networks and promoted them back in 1999. So, I wanted to see what kind of progress has been made. My posting lead to several postings on the newsgroup, as well as personal emails that I followed up on. Ascertaining a current view of the status of wireless at UW was not originally a part of my letter, so I cannot make any claim to represent a complete picture. However, I find the comments to be disturbing in that they indicate that the situation with wireless network deployment at UW mirrors my own experience in some ways. Thus, I include some of these comments here in this postscript. I also include a description of wireless research in order to show their relevance to the academic side of the university.

Waterloo Wireless

Ian Howard, a member of Waterloo Wireless, recently appointed by IST "to help coordinate some of our wireless installations", indicates that UW is not leading in wireless:

Though we are certainly not leading in the sense of making innovative use of wireless technology today, we (UW and IST) are responding to demand to have wireless in key areas throughout campus.

indicates that current wireless networks at UW were implemented and driven not by IST:

A number of recent wireless deployments have taken place, including: Myriad net in DC and MC [implemented by BUL, MFCF, MEF], Wireless Willie in Engineering [implemented by Engineering Computing, WEEF]. [...]

and outlines the difficulties between Waterloo Wireless and IST:

Well, there has been resistance to adopt wireless.

In particular there has been resistance to off-campus wireless connections, though some experiments have taken place with cooperation from Engineering Computing, Waterloo Wireless and IST.

(Waterloo Wireless is an organization that aims to provide broadband wireless access to the UW and nearby community. Wireless networks allow users to use the internet at high speed from any location with a laptop computer.)

Wireless research

The Centre for Wireless Communications research web site:

[CWC] aims at developing a comprehensive graduate research program in wireless communications.

At the CWC there are twelve faculty, fifty-one graduate students and fifteen courses.

The Broadband and Wireless Communications Research Group run by professor Jon W. Mark has seven graduate students.

St. Jeromes University Wireless ResNet

Richard Crispin, IT/WLAN technician for SJU is succinct:

[Cooperation with IST has been] abysmal. They were so unco-operative that we went with an external service provider. Their reaction was "A wireless network will never be connected to our backbone".

He describes the SJU network he implemented through an outside ISP:

The wireless network [at SJU] is designed to provide internet access for the 2 residences. [...] Currently there are around 255 students (out of 276) using it. There are about 60 using laptops and 200 using desktops. This is almost double last year. I expect the number next year to be even higher.

What is wireless networking?

Wireless networking is one of the major technological trends at the moment. It is one of the few sectors of the tech industry experiencing growth and innovation. Wireless community networks may eventually put a major squeeze on the traditional telecom model. VoIP applications roaming on WiFi networks may replace cell phones. Mesh networks may replace traditional telcos. Location-based services are going to be really hot as soon as the technology is developed to locate computers remotely.

Second postscript: a brief biography

I'm a UW alumnus, having graduated with a BMath degree in Honours Computer Science / Co-op in 2001. During my co-ops and following graduation I worked at Apple Computer at the research and design headquarters in California. At Apple, I performed program management duties to open-source the operating system, Darwin, in the 1.0 release (credited here). Later, I served as a program manager for the core OS/kernel team for Mac OS X 10.1. I also initiated the program of another project that has not yet been made public.

During my time as a student at Waterloo, I served on the writing staff of the student newspaper (Imprint) for four years, covering a variety of issues but focused especially on computing issues. At Imprint, I covered the collaboration between UW and Research in Motion in a two-story feature that included a cover photograph of UW president David Johnston with his RIM pager. At Imprint I made a specialty of covering Co-op Education (CECS) and computing issues, and wrote a series of stories covering the convergence of these issues with regards to the "CECS.Online" project (which failed). I served as the computing director at both Imprint and for the Math student society.

I also participated in the formation of several student-computing organizations. One of these was the group Open CECS Online (OCO) which I co-founded. This group is particularly relevant to this paper as it intended to work co-operatively with Waterloo's information technology department, IST. The goal of OCO was to enable student/IST co-operation in the development of an online computer system for Co-op. Myself and the other founders formed this group following the surprise cancellation of the CECS.Online project by the contractor. I will detail the results of the OCO efforts below. Briefly, they were a failure with respect to co-operation with IST. However, OCO did move forward separately with Co-op to develop a system for online resumes called Precis that was used in production for at least one term. Another organization I helped form is uwstudent.org, an incorporated, non-profit, student-run online news source which continues to cover student-related news at UW today.

Also worth mentioning, I was the founder of the Co-op Society movement which resulted in a student referendum. The referendum was held in conjunction with the first fully electronic vote held by the student federation and developed technically by IST. Unfortunately, this first test of online voting did not result in a sufficient voter turnout to achieve quorum necessary to make the (positive) referendum results binding on the Feds. Currently I am back in Canada taking a break from work for personal reasons.

Third postscript: notes of a meeting between OCO and IST

These notes were taken by Colan Schwartz and annotated [in square brackets] by George Roter, both OCO members. I received these notes in an email on the evening of 2000/03/27 (I was in California at Apple). Those present at the meeting were: Colan Schwartz and George Roter for OCO, and Jay Black and Dave Kibble for IST.

JAY BLACK: A CECS advisory council report has been completed and the next step is to develop a "project charter" for a long-term approach. This will be done by the CECS.Online Steering Committee (B. Lumsden, D. Kibble, G. Walker, J. Black, R. Roach, D. Thomas).

GEORGE ROTER: What about the short-term plans to put Access online?

J: Oracle tool-set was selected [Oracle "Web DB" -- some sort of Oracle web-based forms]. Implementation requires Oracle specialist for a week and some training [Jay hinted that this is not a major priority and probably won't be finished until at least mid summer].

G: How can students help facilitate this implementation more quickly? Surely the web forms are not that difficult to learn that a student couldn't figure it out relatively easily?

J: There is not enough work to involve students in the development. I have, however, committed to using students for the testing and debugging.

G: Lets move on to student involvement in the long-term plan.

J: [He admitted that my first point above was entirely valid, but holds that it is not only him that thinks it so] Ok, Dave Kibble sees the project as not large enough to warrant slicing up into various sub-projects, but nicely large enough for some student involvement.

G: From what I have heard IST has not been able to attract top students in the past as co-ops.

J: Well, recently we have been implementing higher and more current technology so the calibre of students has risen somewhat. We feel this project has a lot to offer in design challenges, management of human resources, development of training material, and a considerable amount of responsibility.

G: I still am a little bit cynical of IST's ability to attract top students with a lower pay and somewhat less dynamic working environment. Where might volunteer programmers come into play?

J: Well, that might be determined further into the project by the co-op students who would be responsible for organizing such sub-projects for volunteer involvement.

G: I really see a need for more student input into the design of this system. Since this project is one where an application is being built from scratch it would be a missed opportunity to at least have student focus groups and design brainstorming sessions, both which are perfect for student volunteer involvement. Additionally, if you are worried about security or technical problems with possibly integrating with a toolset such as PeopleSoft then you could set-up a volunteer project room for involvement. I'm sure significant portions of the project could still be undertaken in such a setting, and also, where appropriate, using the open-source model. I still believe that an open-source model is the future of software development and that approaching this project from that standpoint can only be positive.

J: (Still has reservations about deadlines and driven, specific activities actually getting accomplished with the open-source model)

G: Additionally, more volunteer involvement would definitely address the resource problems that IST faces with taking on this project.

J: All good points [still not convinced that it can work]

G: Lets find a project that can be taken-on immediately where we can set-up a satisfactory model for volunteer student involvement and have it serve as a testing ground for the CECS.Online project

J: Let me think about finding an appropriate project.

G: And we will develop a model for volunteer involvement that is compatible with most of the framework you presented to me today. However, you must know that we will suggest a few changes!!!

Fourth postscript: call for confidential submissions

I have received communications from a number of people within the UW community who are well placed to comment on the problems I experienced with IST. In addition, I have communicated with people who are currently experiencing problems with IST. Some of these people are hesitant to disclose their concerns publicly due to their stated concerns that they might "ruin a relationship" among other concerns. Indeed, although you must take me upon faith alone in this statement, there are people with seemingly legitimate concerns that I will not disclose because they have requested me not to.

From these letters, I am concerned that there may be a "chill" effect discouraging public statements in opposition to IST. In order to encourage full disclosure of any problems facing the university's IT programs, I want the review board to call for confidential submissions from the university community. A properly guided and fully confidential process could greatly help to alleviate any potential chill that may exist.

Final postscript: this document on the web

This document is http://simonwoodside.com/projects/itletter.html and the online version contains embedded links to supplementary information. It is also printable from a Web Standards compliant browser.

Updates

2002/11/6 - Final letter.

2002/11/8 - Corrected the first paragraph of the postscript.