The Ethics of Riding the Front Wave of Technology Thomas J. Scott Professor of Computer Science J-Scott@bgu.edu David Beveridge Dean, College of Business and Technology D-Beveridge@bgu.edu Western Illinois University Macomb, IL 61455 Abstract To adopt a new computer technology, you should consider the effects of hardware, software, and training costs and develop an implementation timeline. What are the ethical issues of a new technology decision, and how can they be quantified? The authors [8] discussed a procedure that can be used to develop a feasible set of computer goals and objectives. Neil Postman [7], writing about the spread of technology, advances two theses: 1) that inventors often fail to understand the effect of widespread use of their inventions, and 2) many groups have simply surrendered to the promise of a new technology without weighing its costs. We present seven ethical issues that should be considered when adopting any new computer technology. Next, we present two case studies on adopting modern computer technology. The first case study, adopting Next workstations, shows a failed attempt. Many are currently considering the second case study: whether to adopt Windows 95. We discuss intrinsic issues involved in these two decisions. Next, we discuss known rationales used to justify adopting computer technology. We also present a checklist, which can serve as a guide for those attempting to "ride the wave of computer technology". Section 1: The Surrender of Culture to Technology In Technopoly [7], Postman discusses how inventors are often unaware of the ramifications of their inventions. He discusses how medieval monks invented clocks to regulate their daily prayer times. Today, the clock serves Mammon more than God. He also discusses how physicians invented a stethoscope to listen unobtrusively to a patient's heart and lungs. Today many physicians believe tests more than what the patient says. The subtitle of Technopoly is "The Surrender of Culture to Technology". Postman points out that the printing press replaced an oral culture. In the western culture that preceded the printing press, the ability to listen carefully and speak clearly were marks of an educated person. After the printing press, the ability to write became more important. With an increase in writing, the ability to read became more important. The educational process underwent a radical change. The desire to satisfy society's reading habits caused trivia and advertizing to fill daily newspaper columns. If you talked to Gutenberg, he would have had no idea printing presses were going to wind up printing mostly trivia and advertizing. In the last fifty years, television has changed American culture even more. Newspapers needed to fill their pages once a day, but television is a 24-hour per day experience. Approximately 100 years ago, normal American citizens sat and listened to the famous Lincoln-Douglas debates. The Peoria debate lasted for six hours, with only one break. Modern, 20th century Americans cannot fathom sitting for six hours listening to two people discuss current events. An early television pioneer would probably have no clue about how television has changed our society and caused fundamental changes in the way Americans think, act, and form opinions. Section 2: Universities and Computer Technology A university normally has a mission whose goals and objectives provide rationale for its technology decisions. Administrators who face the ever-present need to upgrade computer facilities must consider human costs and ethical issues. A common university goal is to provide a computer on every faculty member's desk. Published surveys have rank ordered universities by the percentage of faculty who have computers on their desks. Most universities have met this common goal and now must replace less desirable, but still functional, workstations. The older workstations often trickle down to others just getting their first computer. Workstations are commonly replaced every four to five years. In Estimating the Marginal Cost of Adding LAN Workstations Scott and Calhoun [9] showed that the cost of adding a single LAN workstation could cost between $2,000 and $28,000. These huge variations in cost estimates reflected actual LAN scenarios, such as having to add a new LAN server, network operating system license, network hub or router, software licenses, and even having to train a staff member in new hardware, software, and implementation techniques. Other studies have shown that the real costs of adding workstations are much greater than the initial hardware costs. Today, most academic computer administrators are well aware of these extra costs. The ethical issue is, "which cost figure do I budget for new technology, the hardware costs or the real costs"? Section 3: Ethical Issues with New Technology and Change In this section, we present seven ethical issues for those considering adopting new computer technology. These issues form a basis for the two case studies in Section 4, and a starting point for the Riding the Ethical Wave Checklist presented in Section 5. The first ethical issue for any new computer technology is "what will it cost?" Computer technology costs more when it is new. In the worst case, a megabyte of ENIAC (1945) memory would cost $80,000,000 today, as opposed to $40 for a megabyte of current PC memory. In 1982, a 20 meg hard disk drive could be had for a measly $4,500, today a 1.2 gigabyte drive costs less than $300. Ten years ago, a network card was $300, now a modern 10-BaseT card is less than $75. In 1984 an Hewlett Packard laser printer was approximately $4,000, today a much better model costs less than $800. Similar reductions apply to nearly all computer component. From the same vendor, the quantity-one price for a nicely equipped Pentium 100 workstation was $2,700 in June, 1995. In January, 1996, it was $1,900 for the same equipment. A university purchasing one hundred similar Pentium 100 workstations in June, 1995, could have saved $80,000 by waiting until January, 1996. The cost of training and implementation is harder to quantify. Many know how to use older computer technology, such as wordprocessing. How will training be provided for a new computer technology? Good lab workbooks that trainers could easily appropriate will be scarce. It takes time and hands-on experiences to become proficient in a new computer technology. Thus, the initial training will often be less effective than later training. Finally, after a technology is widely used, at least an informal sharing group exists so that faculty and staff can often get valuable help from within their own office or department. Thus, training costs are highest when a technology is new. The second ethical issue is "what are the benefits to our university of this new computer technology?" Like monks, physicians, Gutenberg and early television pioneers already discussed, many users will misjudge the new technology's impact. It is difficult to quantify the benefits of new, unproven technology. Some universities launch a new technology using the "gotta have it" mentality. Because of newness, the only tangible benefit may be the hope for better things. The third ethical issue to be considered is "how stable is the new technology?" Any realistic look at computers shows that new computer technology has problems that will shake out in a year or so after initial release. The old adage is "if you want a really bad computer product, get the first release of a product." The fourth ethical issue is "what is our implementation plan and timeline for the new technology?"" In Management Information Systems, James McLeod [5] discussed three methods of implementing new technology: 1) phase it in gradually, 2) introduce it in parallel to our existing technology, and 3) do a complete cutover. He also presents various forms of prototyping for new technologies. New technology implementors should provide detailed plans describing how the new technology will ramify through the campus community. One well known model for computer technology ramification was given by Richard Nolan in "Managing the Crises in Data Processing" [6]. Nolan presented a six step model, with steps called: Initiation, Contagion, Control, Integration, Data Administration, and finally, Maturity. Contagion occurs when people see others in their community doing something, and they "have to have it." Nolan's Control, Integration, and Data Administration phases are logical administrative followup reactions to the Contagion phase. Nolan notes that very few companies have exhibited the Maturity stage. The implementation plan should consider the ethical effects on the university as the technology gets used by wider and wider audiences. Deciding where to first employ a new technology, and which people will get to use it first are never easy decisions. Unlike many businesses, universities have natural cut-over points in the December-January and summer timeframes. The fifth issue is, "can we justify spending extra money this year, as opposed to next year, or the year after, to avail ourselves of this new technology?" This question is often unanswered as the technology's appeal overwhelms a group. If funding is scarce, we might have to take money away from existing needs to fund a new computer technology. If we wait a year of so, we may fare better, as the new technology will be cheaper. The sixth ethical issue is, "what are the university attitudes toward change?" Postman points out that technologists relish the vitality change brings to an organization. But many classroom teachers resist computer changes, arguing that their lectures, simulation games, statistical analyses, etc., are working correctly on the existing system. They see no need to learn new computer technology just to do the same job they have been successfully doing. When a change to new computer hardware and operating system combination is made, old programs may not even run on the new system. Administrators usually don't know all the uses faculty, staff, and students are making of the existing computer system. Changing to a new hardware, operating system may have unclear effects, as well as cause many valuable classroom experiences to become ineffective. There are people who only want to learn as little about computers as is necessary to do their jobs. Such people will usually resist implementing any new computer technology. The seventh ethical issue is, "how long will this new technology last?" In the early days of microcomputers, vendors touted their wares as having long lives, etc. We now know that most new computer technologies only have a three to five year longevity. Computer vendors now realize that their real money comes from selling you the same products again and again. Good examples come from wordprocessors and spreadsheets. One author has produced over 10,000 pages of text in the last fifteen years. He is currently on his sixth wordprocessor. Venerable ole Wordperfect 5.1 lasted from 1990 to 1995, but 1996 has brought Wordperfect 6.1 for Windows to his desktop. Serious computer users have gone through generations of wordprocessors, spreadsheets, telecommunications programs, and database programs in the last ten years. How long a new computer technology may last may not be known yet, but usually a few guidelines are available. The history of the company bringing out the new technology is an important consideration. The old data processing adage is "no one ever lost a job advocating IBM as a vendor." Another guideline is to determine where this new technology fits with other technologies. Is the new technology a niche product that will be obviated or subsumed into the next generation? Many good early software add-ons became part of next major releases of the main products. Consider HAL, Always, and other Lotus spreadsheet add-ons, which are now incorporated into current Lotus, Quattro Pro, and Excel spreadsheets. Some computer technologies are so persuasive that users adopt without considering the implications. A good example is the Postscript laser printing language, introduced in the early 1980's. Postscript allowed 35 scalable fonts and produced output that was clearly superior to its competitors. Those adopting Postscript did not worry about how long Postscript would last. To paraphrase their phrase, "what they saw is what they got." Section 4: Examples: Next Workstations and Win95 In this section, we present two case studies to illustrate the seven ethical concerns. The first involves a decision concerning the Next, Inc. workstations that was made in the mid 1980's. The second considers a currently popular problem, whether to adopt Windows 95. A perspective on the early history of Apple Computer is provided by John Sculley in Odyssey [10]. Steve Wozniak and Steve Jobs founded Apple Computer in 1978, and Apple's initial success was due to its Apple 2 systems, introduced in 1979. Eventually, its working environment changed and Wozniak left. Under Jobs leadership, in 1984, Apple launched the bold new system, which we now call a MacIntosh. Eventually, Apple's success waned and Jobs was relieved of his CEO duties. Jobs left Apple and founded Next, Inc. His ambitious goal for Next was to build workstations that were significant improvements to the MAC. In The Mythical Man Month, Fred Brooks [1] presents an excellent, short essay called "The Second System Effect". Second system builders try to incorporate all they know into a new system design. Brooks characterizes second systems as having poor foundations, cost overruns, being top heavy and arriving late. Brooks' advice is to be leery of second system designers. In retrospect, Steve Jobs' vision for Next workstations was second system. Jobs may not have known of second systems effects, but as an entrepreneur, risked his considerable fortune on his vision. He propelled Next into a race to build truly innovative workstations in a very short time. The very first Next workstations were based on four concepts: 1) the Mach operating system, 2) Display Postscript, 3) the NextStep object oriented programming environment, and 4) the Motorola 680xx series microprocessors. Three of these four concepts were quite new, but Jobs' charisma convinced many that Next workstations were superior to PCs and MACs, and would do well in the marketplace. In short, it seemed that the Next workstations were ideally suited to university communities. Those considering purchasing Next workstations knew Jobs to be a friend of education. Apple Computer had long proposed "an Apple for every teacher" and was committed to education. As the Next team brought purchase proposals to campus, they offered very generous matching grants for purchase of large quantities of workstations. In short, there was implicit university trust in Next's commitment to education, which was bolstered by the excellent grant program and Next's charismatic leader. Many Next adopters failed to consider the important ethical questions. Next workstations required considerable CPU power to run effectively. Motorola's 680xx chip performance eventually fell behind the Intel counterparts, which were getting faster and cheaper. Motorola quit upgrading their 680xx line, deciding to move toward the Power PC chip. The Motorola decision meant that Next would have to adopt a completely new microprocessor. Next considered using Intel microprocessors, but realized that the Intel field was already crowded. In the end, Next quit producing workstations and dealt a death blow to its users. Shortly thereafter, whole fleets of these workstations were taken out of active university service. The Next scenario underscores the issue, "how long will the technology last?" The Next purchase quickly proved to be a boondoggle. The Next workstation adopters worried less about Next's future. They were more concerned with what seemed to be the best of present opportunities. Universities receiving Next workstation grants for campus-wide Next workstations were quite proud of their decision. In the end, the Next scenario occurred because its technology and leadership offered such promise that the significant ethical issues raised in Section 3 were mostly ignored. The second ethical case study is a concerns the current problem of adopting MicroSoft Windows 95 (Win95). First, we will consider the cost to move to Win95 vs. the perceived benefit. The following quote is from a Datamation article entitled "Plan, Plan, Plan, ... Then Automate", by Vance McCarthy [4] One company with 7,000 PC's estimated that at 90 minutes per install and debug, it would take more than 60 man months--at a cost of more than $3.5 million--just to install Win95". In a companion article Datamation article, Emily Kay [2] suggests that outsourcing Win95 installation cost estimates range from $150 to $200 per hour for Tiger Team installations to $150 to $800 per machine. Everyone agrees that there will be significant costs for training and other software installation. Even the $800 per workstation estimate may be too low. To quantify the costs of upgrading to Win95, the authors' home university, Western Illinois University (WIU) will be used. WIU has 11,000 students and approximately 2,000 desktop workstations, of which less than 50% are Win95 ready. If we adopt a plan to upgrade workstations every 4 years, 500 workstations are needed each year. We assume that a Pentium 100 workstation costs $1,900 and that the cost for installation and upgrade of existing software is another $600. Thus a the cost is $2,500 per workstation. Thus, if WIU acted immediately, it would cost $1,250,000 this year and every year to keep upgrading workstations. In comparison, to give the faculty a 4.5% raise per year costs approximately $1 million. Can such a purchase be justified by a small coterie of technologists arguing that Win95 is the way of the future? Certainly, any academic knows how a democratic vote on a 4.5% raise for everyone or a new Pentium 100 Win95 workstation for a few would turn out! What about the stability of MicroSoft, the company bringing out Win95? MS-DOS has had 15 years success and Windows 3.x has been in the marketplace for 4 years. It has been widely reported that Win 3.x is available on 70 million workstations. MicroSoft has expended enormous effort to roll out Win95 as the chosen successor to Win3.x. MicroSoft's founder is not just a charismatic leader, but some say the richest man in America! Considering these options, adopting Win95 is much safer decision than the Next workstation scenario. Despite claims that MicroSoft has already sold 10 million copies of Win95, many universities and private corporations have not yet adopted Win95. Many question the reliability of Win95. Four reasons to bolster public confidence in Win95's reliability are: 1) it is being delivered by MicroSoft, 2) it uses plug and play and all available memory, and thus simplifies configuration nightmares currently plaguing PC/DOS and Windows owners, 3) it offers a more integrated and consistent graphics user interface, and 4) it offers real multi-tasking for applications. To realize the full multi-tasking benefit of Win95, a desktop must run only Win95 software. This benefit allows software vendors to sell us another version of the same software we are now using. There have been many articles in the glossy, advertising driven magazines discussing Win95 internals. In "Windows vs. OS/2 Warp: Shootout at the OS Corral" Randall Kennedy [3] discusses the fundamentals of the Win95 tasking engine. Win 3.x offer only what MicroSoft calls "cooperative multi-tasking" and not true preemptive multi-tasking, such as that provided by on Unix systems. Win 3.x cooperative multi-tasking means that a Win 3.x application must voluntarily surrender the CPU before the Win 3.x tasking engine can allow another process to run. In building Win 3.x's successor, MicroSoft opted for Win95 to be compatible with Win 3.x and DOS. Thus, Win95's multi-tasking only works correctly for Win95 applications. Under Win95, if a user has three applications: a Win95 app, a Win3.x app, and a DOS app, all running simultaneously in three different windows, the DOS application can dominate, and its failure can freeze the workstation, and cause the data from the other windows to be lost. Thus Win95 will be reliable for Win95 applications only. If a user wants to run some Win95 applications and some old DOS applications simultaneously, there is still no security against the DOS application killing the machine. Some might ask "how much desktop change is necessary?" The typical university workstation has changed approximately every three of four years from 1980 to the present. There were Apple 2's and IBM PC's in the early 80's, then PC AT's and MACs, then Intel 386 based desktops and MAC SE's. In the early 90's, we got Intel 486 machines running Win 3.x and MAC Quattras. Presently, we want Pentium workstations or Power MACs. Similarly, those running Unix workstations have seen a succession of increasingly powerful workstations from Sun, Digital, HP, etc. As time marches on, we get considerably more computer power for the dollar. Since some change occurs every three or four years, is the move Win95 going to be the next logical change? Another aspect of the Win95 reliability issue stems from the current view "the network is the computer". In the past, we had large campus computer systems, but today, many feel that it is more important to be on the Internet and World Wide Web. To connect a workstation to the Internet requires generating the 32-bit Internet address. University workstations often achieve their Internet addresses by connecting to the campus LAN and running a dynamic procedure called "bootp". Microsoft chose not to deliver bootp capability with Win95; instead it delivers "DHCP" capability. Without bootp capability, Win95 causes real networking problems. To use Win95 terms, it is hard to "start it up" on the Internet. What about the Win95 implementation method and timeline? Many trade-press articles have proudly proclaimed that "DOS is dead". Most universities where lots of DOS applications are running realize that something newer and better should be used. To get a Win95 application certified as Win95 ready, a software developer must also ensure that it runs on the MicroSoft Windows NT operating system. Many suggest that Win95 is a stopgap measure as part of Microsoft's strategy to move all desktops to Win NT. Determining a cost effective Win95 adoption timeline is important. Some universities are following the implementation and timeline plan WIU is currently using. Over 80% of WIU workstations are DOS based, connected via Novell Netware servers. Of the remaining 20%, there are more MACs than Unix workstations. The WIU timeline uses a wait and see attitude where a coterie of power users are pioneering Win95 campus use. In the meantime, when a DOS based workstation is upgraded, the upgrade will run Win95. It looks like a wholesale commitment to Win95 will begin in Fall 1996 or Spring 1997. This approach can effectively handles well known Win95 problem areas. Section 5: The "Riding the Ethical Wave Checklist" We have attempted to show some issues involved in a decision to "Ride the Wave of Computer Technology". The following twenty-three item checklist could serve to begin such a process. Different university communities might add or subtract items from the list, but each category should be considered. Part 1: Advantages and Disadvantages 1. Advantage #1 2. Advantage #2 3. Advantage #3 1. Disadvantage #1 2. Disadvantage #2 3. Disadvantage #3 Part 2: Initial Costs 1. Initial Cost, if Implemented This Year 2. Cost if Implemented Next Year 3. Ease of Implementation if Implemented Next Year 4. Value three Years Later Part 3:Training Issues 1. Size of Training and Implementation Staff 2. Days of Training required for average Trainer and Implementer 3. Training/Implementation Staff attitudes toward retraining 4. Approximate Cost of training the Trainers/Implementers 5. Estimated number of faculty needing training 6. Days of Training required for average faculty 7. Faculty attitudes toward training 8. Approximate Cost of Faculty training 9. Estimated number of Administrative staff needing training 10. Days of Training required for average staff member 11. Staff attitudes toward training 12. Approximate Cost of Staff training 13. Approximate Cost of Training Everyone Section 6: Benefits of Riding The Wave During the last 20 years, employers have provided many good jobs for university graduates with good computer backgrounds. During the last 10 years, the author's own Computer Science department's job placement statistics show that over 99% of graduates who do not pursue graduate school got jobs in the computer industry. Many non-computer departments have added computer use courses within their discipline. We have seen new courses such as business simulation games, spreadsheets for finance majors, computerized accounting, using SPSSX or SAS for social science research, using the Internet for research, etc. We all understand that a graduate who is totally lacking in computer skills will be less employable. Computer users are never happy with the status quo. We delight in having a machine that is faster than our previous machine, or one that has more innate capability. Many computer users are not satisfied with three year old computer technology. As Postman has pointed out, a segment of society will always want to ride the latest wave of computer technology. Such people interject this fact whenever they can: into advertising brochures, recruiting speeches, lists of accomplishments, etc. For many universities, providing quality education leading to good jobs is an important goal. Urban universities often have classes that provide hands-on experiences with specific types of computer systems and application software used by area businesses. In such cases, using three to five year old technology will not accomplish the stated job-related university goal. Students receiving training on three to five year old computing equipment might have less success getting good jobs. Each university has competitors. If a competing university has better computer technology, this serves as rationale to improve our technology. Thus keeping up with the competition, when it is deemed necessary, is another benefit of riding the computer technology wave. Despite the naysayers who shun adopting any computer change, a new computer technology always offers a hope that familiar things can be done better. For example, a graphical user interface, such as those used on MACINTOSH, MS Windows, or X-Windows systems, is easier to learn and remember. Using outdated computer equipment does not create a stimulating learning environment for teachers or students. Better computer technology improves campus morale, which helps attract and keep faculty and more students. Section 7. Summary We have discussed ethical issues that accompany a decision to adopt the latest computer technology. Such technology offers rewards and risks, benefits and pitfalls, an opportunity for success and many paths to failure. Anyone facing such a decision should build a checklist similar to the one presented in Section 4 and use it to facilitate decision making. In our modern world, the people cost should be studied carefully. Finally, the time element should be analyzed in two distinct ways. First, what are the incremental costs of the technology if adopted now, a year from now, or two years from now? Second, how long will the new technology last. After considering these options, a group should be better able to make a decision that is both technically correct and ethically pleasing. Bibliography [1] Brooks, Fred, The Mythical Man Month, Addison Wesley Publishers, Reading, MA, Jan. 1992. [2] Kay, Emily, "SHOULD YOU Outsource your MOVE to Windows 95?", Datamation, Nov 1, 1995. [3] Kennedy, Randall, "Windows 95 vs. OS/2 Warp: Shootout at the OS Corral" IBM Software Quarterly, October, 1995. [4] McCarthy, Vance, "WIN95: Plan, Plan, Plan... Then Automate" Datamation, Nov 1, 1995. [5] McLeod, Raymond, Management Information Systems Third Edition, SRA Publishers, Chicago, IL, 1986. [6] Nolan, Richard, "Managing the Crises in Data Processing," Harvard Business Review 57 (March-April 1979). [7] Postman, Neil Technopoly: The Surrender of Culture to Technology Vantage Books, Random House Publishers, Nov. 1993. [8] Scott, Thomas and D. Beveridge, "Defining University Computer Priorities", Proceedings, Computers on Campus National Conference, Myrtle Beach, SC, Nov. 1993. [9] Scott, Thomas and J. Calhoun, "Estimating the Marginal Cost of Adding LAN Workstations" Proceedings, Symposium on Applied Computing, Kansas City, MO, April, 1992. [10] Sculley, John, Odyssey: Pepsi to Apple, Harper and Row, New York, NY, 1987.