In 1984 I was fortunate to be hired by the Xerox Research Centre of Canada. Up until that time my exposure to computers had been using very simple personal computers to run short basic programs. I wrote those programs to read in experimental data and write out a set of results - and that was all. When it was time to write my first report at Xerox, I was introduced to the workstations that seemed to be everywhere in the building. When I logged on to one of those machines I discovered a Graphical User Interface that at first seemed strange and difficult to use. It was designed to present things to me as "my desktop" with icons for my documents and others for the various laser printers around the building. To print a draft of my report I learned to drag the document icon with the mouse and drop it on the printer icon. The icons were conveniently labeled with the room number allowing me to choose the laser printer closest to me.
As I worked with this system two things started to be obvious. After getting over the initial strangeness of the system, learning new things was relatively easy. It was not hard to guess how to do things I had never done before. More remarkable, was the discovery that my desktop followed me wherever I went. When I logged in, my documents were not only in the folders I left them in, but would also be "lying on the desk" where I left them. The state of my desktop was perfectly preserved.
As I began travelling to Xerox facilities in Webster, New York and Oklahoma City there were always workstations available where I could work. My work, in exactly the state I left it, followed me around and was readily accessible. Even better was my ability to send perfectly formatted documents to any one of my coworkers in any Xerox lab and to receive their fully formatted, graphics rich, documents anywhere I was working.
Today many companies such as Microsoft and Apple are still striving to provide the kind of system Xerox invented but couldn't make any real money on. Some people at Xerox are still thinking about extending these capabilities and refer to this a ubiquitous computing. Novell calls their vision of the future pervasive computing, and Sun Microsystems says "the network is the computer." For me the central experience can be summarized by four simple phrases:
Or at least something like that. Ubiquitous computing has been promised in different forms by people selling software, network systems, and of course computer hardware. Lately universities also want to provide their students with ubiquitous computing. William H. Graves at the University of North Carolina, in an article entitled A Framework for Universal Intranet Access describes the goal of "Universal Access" this way:
"All students, all faculty members, and most non-faculty employees will have convenient and affordable access to a personal computer connected to the institution's network at any time and from almost any place they are working or studying - a library, a home, a field location, or another off-campus location."
At some point then you have to ask:
I think there are three overlapping reasons to promote ubiquitous computing at a modern Polytechnic University.
Twenty years ago Ryerson's one centrally supported mainframe satisfied the academic needs of a small group of students and professors. Five years ago the New Systems Committee recommended a large expenditure that led to upgrading or purchase of roughly 300 lab computers, many of which CCS still operates and three UNIX servers for statistics, programming, and E-mail/Internet access. At the time, no more than 100 simultaneous E-mail users were predicted. Today the four and five year old systems we bought in 1992/1993 have passed their useful life. They do not run the two years old Windows 95, cannot adequately handle files from the latest software suites students use at home and in other department and faculty labs, and are constantly breaking down. Worse, students taking courses that had relied on those machines some years ago find they are now contending for access to them with students who would not have used them three years ago. Internet access and E-mail are just two of the "killer" applications (in a negative sense) accessed by students from every program in CCS labs that are displacing students who want to work in more traditional ways - programming and wordprocessing are two examples.
There has also been a tremendous increase in workstations on campus outside of CCS's labs. Applied Arts now operates just as many, if not more machines, and the University has roughly 1100 accessible computers across the campus. Nevertheless, even with so many students owning their own computers, there is still tremendous contention for computers in two areas:
The need for access has become universal across all faculties and even if we were to replace every machine every three years we could not keep up with demand nor could we afford the rich set of additional applications faculty are demanding.
In the context of exploding demand for networked computers and the fiscal restraint of the last two decades several universities have turned to the concept of the "laptop university" to solve the problem. Accadia and the University of Minnesota are two examples. There are many schemes to do this but to be blunt: all involve significant off loading of the cost of the networked workstation to the student. This can take the form of all students having to lease a laptop or buy one outright from the University. So, before exploring the value of a "laptop" university maybe we should ask the question:
Personally I believe the answer is yes, but under the following conditions:
The cost to buy a laptop computer with modem and Ethernet card, a suite of software, an Internet Service Provider account, and maintenance is still well over $4000. Few students will simply write a check for $4000 in their first year of University and agree to walk away with an antiquated laptop four years later upon graduation. Paying for the equipment over time raises the cost to over $1000 per year and any equipment renewal adds to the bill. For many students across the University this is an unacceptable additional burden. For others it is not.
For students in some programs a laptop is the wrong machine to buy. If they had $4000 they could get twice the storage, better performance, and better displays (critical in many applications) for that kind of investment. Also, the cost to upgrade a desktop system after two or three years is not as high as replacing an entire laptop.
But lets consider for a moment what it might be like some time in the future. Every year the computing power, available storage, and quality of the display of both laptops and desktop systems improves while the price paid for these things decreases. Imagine for a moment that a student could purchase a fully equipped laptop computer, all the software they need, a maintenance agreement, and Internet access for $500 a year for each of their four years at Ryerson. Then imagine that every student enrolled at Ryerson were to buy one. What would this mean for our students and what would it mean for Ryerson?
For students the laptop university represents an immediate shift to ubiquitous computing. Being connected anywhere and anytime changes the way work is done. Today a student working in the library must make notes and then go to a computer lab to type up a report or essay. In one report written in 1994 by Green the average workstation time available to the average American university student was 3.5 hours per week. Having a laptop with you means you can connect in the library, in class, and at home. Your work travels with you, you can send work via the Internet anywhere and you can get information anytime. The laptop also means you posses the computer. A student can configure the system in a way that makes personal sense and can store work on the system knowing it will be there when it is needed. Compare that to a lab where you must bring your notes and your work on disk and then take it away that way - where the state of the workstation has little to do with your way of working, and the working environment is stressful and not conducive to reflection. Imagine being able to review lectures, run virtual experiments as many times as needed with different parameters, and being able to access on-line databases anytime and anywhere. Finally, the system which would have to be configured for Ryerson would have the added advantage of being designed to connect to and integrate with the digital services available on campus. The E-mail system would be ready to go as would the Web browser with bookmarks to Ryerson's resources. For students, especially those who don't know about IP addresses and E-mail protocols this would be a significant advantage - as would the self support of having all their peers using the same system.
When I arrived at Ryerson I was amazed at how unconnected things were. In 1984, at Xerox, I was using a graphical user interface, networked laser printers, discussion groups, and media rich E-mail. When I taught my first day class at Ryerson in 1991 and gradually starting introducing image processing into the Photographic Technology courses I was teaching, there was no rich E-mail system, no scheduling system, no lab laser printing, and whenever I asked for a new service, such as being able to send an image file for printing over the network, the way I had to tell my students to do it was so crude it was almost beyond belief. Using CUTCP and chmod and then a special print routine on malthus. Unbelievable. Ryerson did not have, and still largely does not have, the systems integration and development resources it needs to leverage the power of its network. The financial restrictions on the University have traditionally meant that we focus on buying the hardware and software to put in labs and then try to support it. There has been little or no money for development of better integrated client and server systems - something the University desperately needs.
If our students were all to buy laptop computers the University could refocus some of its resources on:
Laptop computers offer the opportunity to have pervasive on and off campus computing, to end Ryerson's cycle of buying equipment and watching it become obsolete, to end the constant shortages of computers as more and more students find they need access, and to remove one of the most fundamental restrictions to faculty use of digitally mediated teaching methods.
Increasingly, and in a way we often take for granted, academic work is something we do using a computer. A well implemented system using laptops as the mobile network workstation can offer a level of ubiquitous computing that was not even available to me at Xerox. Its not a question of taking your laptop to a beach or mountain so you can appear in an Apple commercial curtesy of a helicopter flyby. It is the ability for our students to do work anytime and anywhere in a way that makes sense for them and is connected to the academic project of the University.
Now lets come back down to earth. Today's LINK students spend roughly $5600 over four years. In many programs this cost is too high given the income levels of the students and the limited use those departments are prepared to make of them tomorrow. In some departments, making a laptop purchase a requirement may not present a fundamental barrier to students. In fact, it may be welcomed by the majority of students who buy systems anyway and wish they were better connected to their work at Ryerson. These departments may also be prepared to exploit the opportunity that laptops represent for ubiquitous computing. The business school is one obvious possibility but its not hard to imagine interest from others.
Then there is the other problem. Laptops are more expensive than desktop systems. In some disciplines where addition storage, compute power, and display space are critical requiring a laptop purchase may seem unreasonable today. Students with CAD, visualization, and multimedia production requirements can get much more for $4000 out of a desktop system than a laptop at least for this year and next, and maybe for some time to come. Desktop systems are also less expensive to upgrade.
Finally there is the question of part-time and continuing education students. Can we ask them to buy a machine in order to take one course? I don't think so.
Where programs can show that the laptop model is reasonable the University should immediately enable those programs going forward with plans for the laptop computing model. The University should allocate significant funds for the:
At the same time Ryerson must - at least once - rebuild its existing central labs. Next year anything less than a 486 should be replaced and the following year all systems should be capable of running a robust and maintainable operating system (to reduce the cost of ownership/support) such as Windows NT. After that a budget needs to be established to replace 1/3rd of the workstations every year. In the past these investments have been made in large single year acquisitions. This is a poor strategy for buying computers. Base budgets must be established to maintain the labs at a certain level rather than spending very large sums every four or five years.
Ryerson must also redesign how its labs are used along the lines of a more efficient, user friendly, client-server model that facilitates students moving between systems on campus and their systems at home.
As laptops increase in capability and decrease in cost more and more departments or programs should be encouraged to become "laptop programs." until most of the University has migrated to this model. However, this will take some years. In the interim a mixed model is unavoidable and must be approached with a different funding model that includes the total cost of providing networked computing and maximizing the value of student purchased desktop and laptop systems.