John Stenzel, Coordinator
Computers in Composition Program
February 1997
The computer-assisted-instruction world has confronted instructors and administrators with difficult decisions and complex issues. In this report I examine a choice that affects the professional lives of instructors and students, and explore the conditions that make such a decision difficult and significant. Writing and thinking about these topics has led me to consider the "ecology" of this technological environment, the many interesting and unexpected ways in which pieces or species influence each other and interact. Early biologists did not fully understand the interconnections between living systems and the physical environment; just so, today's technology providers, administrators and salespeople often consider only one dimension at a time, or conceptualize a complex task in a limited way--and thus suggest solutions which bring their own problems. Natural environments once were thought to be simply collections of plants and animals; computer classrooms used to be considered simply computer-equipped rooms in which classes are taught. The interconnected and interdependent ingredients of the CAI classroom environment include computers themselves, the application software they run, the networking software and hardware that connect them, the server computer and switches that allocate network resources, and the physical configuration of the seats and screens themselves.
We have seen the results of limited thinking this past Fall, as writing instructors in our Macintosh classrooms were forced to change the way they teach because of a change in software, a change made not by the teachers but by the campus Information Technology Division, acting, it explained under pressure from market forces and its own upgrade policies. The result was disastrous: launching word processing software on even one machine could crash the classroom server, and many dozens of classes were curtailed or disrupted because of added demands on the local networks. All this occurred despite an equipment upgrade (faster PowerPC's had replaced old SE/30's in 241 Olson) and despite promises that the disruption would be manageable. In the wake of the problems that ensued, the Information Technology / Lab Management group (IT-Labs) has added new networking equipment, shifted software from server to individual hard drives, and worked many hours that might not have been necessary at all. An examination of this supposed upgrade can teach us much about the way information technology does and does not work for pedagogical purposes, and can suggest new ways to approach future technical decisions. Ultimately, the lessons learned from this experience and others have shed light on the larger issue of "upgrades" in institutional settings: any upgrade undertaken without careful analysis of time costs and lost productivity--analysis that acknowledges the economies or diseconomies of scale as well as training and transition costs--is a fiscally and pedagogically irresponsible act.
Only when personal computers became equipped with word-processing software did the computer cease to be a number-crunching tool of scientists and become a mass marketed phenomenon. Allowing comparatively painless manipulation and printing of text documents, the computer revolutionized the way writers plied their trade, and writing instructors quickly saw the potential benefits of computer assisted writing instruction. The English Department's Composition Program began integrating word processing and text manipulation into its classroom instruction in the early 1980's, since the computer classroom could in theory afford an outstanding venue for teaching writing and critical-thinking skills. Word processing remains the number one use of computers across campus, though recent years have seen the rapidly increasing popularity of electronic communication (e-mail) and the many-splendored roadside attractions of the Information Superhighway (the World Wide Web in particular).
Thus, the question of which word processing software to run in the UC Davis computer labs, which is where we attempt to teach our CAI sections, is fraught with significance. As Coordinator for the English Department's Computers in Composition Program, I was informed early in 1995-96 that IT-Labs intended to load the newly-released Microsoft Word 6.0 in the Macintosh labs and classrooms, which had previously been running version 5.0 or 5.1. The fact that we were trying to use a full-featured word processing package in classrooms at all, especially in a server-launched configuration, is evidence of a broader mismatch of tool to job; this is what we were handed institutionally, and it took far too long to question the validity of the configuration.
The ensuing struggle to delay the upgrade, and the aftermath of the upgrade itself, became an interesting case study in the ways centralized computing facilities respond to and (more frequently) dictate to their users, and the ways in which decisions made at one level and under one set of assumptions send effects rippling across a much broader area. To examine this seemingly innocuous situation is to peek into the institutional culture of technology providers, and to appreciate just how far the University has to go in reconciling the different needs and pressures of teaching and technology.
I'd had my doubts and dreads about the upgrade ever since before the first grim notices in the Macintosh user publications ("Pass the ketchup, this one's a dog," began the well-regarded Berkeley Macintosh User Group's review), since I was already dissatisfied with sluggish performance of MS-Word 5 in the lab classrooms. When I became Coordinator in the autumn of 1994-95, IT-Labs had been running Word 5 for at least a year, more or less successfully; at least we were accustomed to the spotty performance and no one apparently questioned the underlying assumptions of network architecture and software choice. Naively I thought that this program was the best one for us, that this represented some sort of optimization; like other writers, I only cared about getting words on a page, and being a longtime DOS user (WordPerfect 5.1 with its macros I found much to my liking), I faithfully followed the dictates of those in the know. My education was about to begin.
Some program-specific background is in order. The Macintosh classrooms, and with them the teachers who use them, had been getting swept along on the tide of word-processing upgrades for many years now, a process that began in our earliest Macintosh lab: the first Macs were stand-alone machines running MacWrite, which everyone thought was just fine for our purposes. In the next few years IT then moved to WriteNow, which people got used to, and then to MS-Word in various versions as Microsoft consolidated its dominance of the market. The switches, as best I can figure, were not dictated by pedagogical concerns, but were choices made by people in lab management (the specific department of IT had not yet been formed), either with or without notice given in advance of the upgrade. In some cases the new word-processing package had desirable features as the personal word processing software matured, or may have offered advantages of security (server back-up) or price, but the decisions were made from the supplier end--from IT and its internal, informal surveys of client queries, not from English instructors or anyone with significant teaching experience.
Though I taught in the labs I cannot recall being asked or informed about software preferences, and changeovers often brought temporary chaos, as when one word processor could not read files created by another. Often the change was made over the summer or break, and instructors did not learn of the switch until they double-clicked and saw the "Document cannot be opened, because the application that created it could not be found" error message. At that point a stopgap solution would always be introduced, and people were weaned from an old version onto a new one. When conversions were difficult, archives of older teaching materials were thereby rendered much less accessible, even when the converters did exist--a difficult-to-document but nonetheless significant pedagogical consequence not formally acknowledged by the promoters of a given change.
It's important to note here, that the consequences of such decisions vary from one unit to another, complicating the upgrade issue: because the nature of English composition requires the generation and modification of many thousands of word-processed files a year, the impact of a change is felt far more seriously than IT-Labs personnel might imagine, and seemingly small glitches actually have profound consequences. Such an observation underscores the need to include some changeover losses as part of the costing mechanism when assessing whether a particular technological solution is appropriate across a broad spectrum of users in a wide variety of settings.
Because some versions of the same program are transparent with each other, or, "backwards compatible," this particular problem had not manifested itself recently: Word 5 documents are completely readable by Word 4, and vice versa, so that instructors and students running old Macintoshes at home with System 6 and Word 4 were not inconvenienced by the switch to Word 5. More disruptively, though, as Microsoft adds new features and changes the way that files are actually stored electronically, versions of MS-Word have become less compatible with each other: a document created or modified by Word 6 is not readable by a computer running Word 5, unless that computer has had a conversion extension placed in the Word Commands folder. Word 6 has a "Save-as" command to save documents in Word 5 format, but this adds several steps, including a misleading reminder that the changes have not been saved, in what was otherwise a simple process. [Looming in the near future is yet another Microsoft upgrade, Office 97, which introduces a new file-storage configuration that will not be readable by Word 6!]
The issue of which word processor to run in campus computer labs became more contentious and more significant, as a move to version 6 would have significant direct and indirect impacts on the English Department, including greatly increased launch time, file incompatibility problems, instant equipment obsolescence, and much time wastage. To understand the consequences of the current upgrade we must examine the different settings in which we use the programs: in classrooms, in our offices, and at home. The next sections examine impacts in each of these areas; this process provides a powerful argument for changing the ways we think about computers, their instructional uses, and their real costs.
It should be noted that as labs the rooms work reasonably well, with certain caveats; as classrooms they leave much to be desired. Not surprisingly, the IT entity in charge of these rooms is Lab Management, and not classroom management: the networking and layout of the rooms have more to do with a lab setting than a classroom setting. [For a fuller discussion of design and layout issues, see my "CAI Classroom Layout and Design Report," (June 1996); http://wwwenglish.ucdavis.edu/compos/compcai/report.htm].
In the classroom the choice of word processor has other dimensions besides strict compatibility issues, most notably the logistics of launching the program itself. In the configuration we lived with for several years, simply readying the computers for writing took significant amounts of class time--often 10 minutes and upwards! For reasons having to do with site licensing and program security, IT-Labs could not or would not load each desktop hard drive with the full Word 5 software package, but had to store the program on a central file server elsewhere in the building. (The fact that over Christmas 1996-97 IT-Labs changed their policy and installed MS-Word 6 on individual hard drives is particularly galling for instructors who limped along in server-launched configurations for so many years, but this awareness is now informing current and future planning efforts.)
With the networking configuration then in place, each classroom server had the main applications residing on it, so the MS-Word program was downloaded to individual computers as users called for it, by launching the application or by double-clicking on a file icon. With a program of Word's size--around a megabyte for the main program, in the case of Word 5--this "server call" took a significant amount of time, depending on the amount of local network traffic. In some classrooms with poor network connections--with small-diameter information hoses transferring application software from the server "tank" to the individual computers--the simultaneous launching of Word 5 took 11 minutes or more, according to IT-Labs' own benchmark tests. The reasons for this poor performance stem from the networking arrangement itself, the wiring and electronic red-tape involved in moving many megabytes of program from the server to the individual desktops.
Computer users in recent years have noted a distressing phenomenon: as more and more computers get networked together to share resources, certain performance levels paradoxically have declined, a fact inexplicable without some understanding of network architecture. How can the same server-and-stations setup work reasonably well as a lab and miserably as a classroom? The difference in time or rhythm between lab use and classroom use illuminates one important facet of the problem: when a room full of desktop computers is used as a lab on a drop-in basis, individual server calls--requests from a desktop to download a program from the server to the individual computer--are staggered over time, spread out so that the server is only processing a few requests at a time. Students enter the lab, launch applications, work, and leave, without noticing delays in server performance. This type of interaction, can be described as an asynchronous mode--where actions occur not-at-the-same-time.
In a classroom setting, however, the server is called upon to do something quite different: at the beginning of the hour twenty-odd users attempt to log on and launch an application at nearly the same time, with an electronic traffic jam the result of this synchronous use. In the case of Olson 241 and 247, which share a single server, as many as 45 machines are involved, and only a couple of 10-megabit data lines. Launching MS-Word 5 on 25 machines in such circumstances took a fifth of a 50-minute class period, an unacceptable amount of time to waste. Because all the computers have a gatekeeper program that verifies login id's and passwords with the administrative computers on the campus backbone, this local network is also held hostage for many minutes as the confirmations are processed through a very small data stream shared amongst 20 to 45 users. If a teacher in one classroom was running a server-intensive program like Timbuktu and the other was trying to have students drop off files, the resulting gridlock sometimes led to lost papers and system failures: the server icon "disappeared," and machines had to be restarted to reestablish network connections.
Experienced instructors learned to cope with such handicaps by training their students to launch Word 5 immediately upon entering the classroom, in preparation for the work to come, but this procedural band-aid was not wholly satisfactory, and in fact did not work at all this Fall. Frequent network crashes and slow response virtually washed out Fall Quarter 1996 for CAI instructors, severely lowering instructor morale and IT-Labs' credibility. In a formal survey of computer classroom instructors done in 1994-95 under the old system, slow launch speed was the single most often-cited barrier to effective use of computer classroom time. Over the summer of 1996 an "upgrade" occurred that further increased these delays; no full-scale benchmark tests were undertaken to prove that the new system was usable, and subsequent experience showed that the new word processor was clearly not workable. I can't overemphasize how much toll such problems take on teachers' morale: veteran CAI teachers have made it clear that they do not want to teach in computer classrooms because the equipment seems to work even less reliably than it did two years ago!
So far at least, solutions are seen only in terms of buying faster servers or faster switches or faster computers, not in re-thinking the original assumptions about the ways programs run or the ways teachers work. Yet hard-technology solutions may not be the most cost-effective, since other considerations, such as "chatty" network protocols and the electronic red tape of the computers themselves may blunt the effect of theoretical improvements. On the other hand, reconfiguring the older Macintosh labs to meet the needs of classroom instruction may be the best solution after all, with rooms doing one job well--the teaching job--and not doing a whole set of jobs in a mediocre fashion.
Why Upgrade At All? Prelude to a Battle
Late in 1995, when IT-Labs was first contemplating its upgrade to Word 6, English instructors (and Spanish instructors as well) raised their voices in protest, having heard that the Word 6 program was far bigger and slower to run than Word 5, even on a fast single-user machine. The increased launch time would have slowed classrooms down so far that meaningful teaching would have been impossible, with no apparent gain in functionality. Responding to this outcry, Tim Leamy and Pete Blando of IT-Labs discovered that Word 6 could be subdivided, so that parts of it could be stored on desktop hard drives and part loaded from the server, thus reducing the wait time required when 20 users launched simultaneously. It was this reduction in Word 6 launch time that prompted them to initiate the upgrade over Christmas vacation 1995, since they presumed that this was our primary objection to the new version.
[I should add here that they either had not looked for or had not been able to discover a way to speed up Word 5, and thus were to some extent comparing apples and oranges with their newly revised launch-time figures for a compartmentalized Word 6. Interestingly, when confronted by the disastrous unworkability of Word 6 in Fall 1996, IT has loaded the word processor onto every hard drive, not the server, and reduced the launch time to a much more manageable figure! Of course, there was no comparison to Word 5.1 under these conditions, which is no comfort to instructors coped with interminable lags and crashes using Word 5 under the old server configuration.]
Perhaps surprising Tim and Pete, I immediately put on the brakes as soon as I heard of these plans. What I forcefully put to Pete in a meeting over the New Year's vacation was, first and foremost, that change comes at a price--not just dollar cost, but time cost and other subtle and not-so-subtle productivity considerations. I made it clear that this change to Word 6 would increase waiting time, instantly obsolete our instructors' machines, lower productivity, and add many new problems that would require my time and the time of other instructors to fix--time that was going to come out of our budgets and paychecks and not IT's. This uncompensated and unexamined time commitment is one of the key economic issues that IT as a division and as an industrial entity must analyze more fully. The consequences elicit several different kinds of questions, and ensuing sections of this paper deal with them in turn:
o How will we have to deal with MS Word 6?
o What reasons are there to install it at all?
o What should IT do to find an acceptable alternative?
o How should IT perceive and respond to upgrade pressures?
Many technically savvy readers may already foresee the consequences of this change, but my experience has shown that even for professionals in the field, the cause and effect chain needs closer examination, especially when considerations of scale are factored in. There may well be better programs for our classroom than Word 6--but the fact of the matter is that MS Word is an industry standard, and for the moment we are stuck with what IT has given or suggested to us. Whose responsibility is it to determine which software to use in our labs? Until now, the institution has rolled along, with a momentum of its own. Who has the resources to solicit demo copies, test demonstration set-ups, and weigh costs and benefits? An individual department, or the entity on campus best equipped and most experienced in these matters? I suggest that IT itself consider forming a software-adoption workgroup for core users (as distinct from the Center for Advanced Information Technology), and I hereby volunteer for membership.
When I asked my contacts at Lab Management what advantages the new version would offer us, the rationale I was given proved illuminating in several respects. Not surprisingly, what the technology suppliers thought we needed and what we needed were quite different. More subtly, their response indicated how IT suppliers get a skewed view of the demand for a package or a service, and make decisions that may not reflect the greatest cumulative good for the greatest number of their constituents. Here was the feature list that was would supposedly make us want to change versions:
o Word 6 has an as-you-go spell checker, which highlights typos as you make them--a feature that goes directly against the process-oriented writing principles we try to teach, wherein students must be trained not to obsess over surface "correctness" at the beginning of the writing process, but should feel comfortable drafting and then revising. I personally would disable this annoying and counterproductive feature as soon as I installed the software.
o Word 6 also offers more sophisticated integration of charts and graphs--but only a tiny proportion of our students deal with this at all, and no instructors, to my knowledge. Moreover, those who do are sophisticated enough to work around the challenges posed by the already feature-clogged Word 5, an argument that could apply to other facets of the upgrade controversy.
o Word 6 offers smoother integration of interlinked Excel 5.0 spreadsheets, so that the numbers and charts in a report generated in Word could reflect changes in the spreadsheet to which it was linked--another power-user feature applicable to a tiny proportion of our users.
o Word 6 for the Mac is compatible with Word 6 and 7 for Windows--an elegantly circular irrelevance for a pool of student and instructor users overwhelmingly Macintosh-oriented.
o Word 6 also has improved some sophisticated page-layout tools for multiple columns and wrap-arounds, and some Web-creation tools as well. On the other hand, there is apparently nothing in it that would convince an adept PageMaker or PageMill user to switch over.
So what does Word 6 feature that I might actually want? Its "multiple-undo's" function could be useful to writers, since at present only the most-recently-made deletion or correction can be undone. Returning to the downside, the menu bars and rulers of Word 6 are cluttered and just-different-enough-from-version-5 to cause users to have to unlearn and relearn, and there's just enough shift in the look and feel to give users a little more uneasy time on the learning curve. For example, command-A "selects all" text in Word 5, but in Word 6 a bug means that the ruler disappears as well! The grammar checker remains a teacher's nightmare, offering inadequate and misleading suggestions and running extremely slowly.
Above all, because the source code was originally written for the IBM world and never re-tooled to run efficiently on the Mac, MS-Word 6 for the Mac continues to be a slow, inefficient program as far as a huge number of everyday users are concerned. Even if new features were used, vagaries of Microsoft's source code mean that Word 6 adds significant delays when pulling down the File menu, for example, and interminable loading to run word counts and spell checks. Ultimately, the best argument IT can offer is that Word 6 is part of the upgraded Microsoft Office suite, that they were forced to upgrade Microsoft Office, and that the story ends there. But should it?
The move to Word 6 at Davis perfectly illustrates the ways that a single decision reverberates down to the desks of every instructor and student who uses the labs, with some unforeseen and certainly un-quantified results. Herewith, some cold hard numbers and some informed speculation. The paragraphs that follow illustrate a different kind of approach to a software upgrade, more in common with environmental analysis or economics than with the considerations that apparently drive the computer world.
The effects of the change are felt in several different phases of the writing process: in running the program in the labs, of course, but also in opening Word 6 files at home or in the office, in saving files, and in dealing with the effects of inattention to small details required by the upgrade.
First off, could we have solved the problem simply by installing the new version of Word 6 on every computer used by every instructor? As far as the English Department is concerned, the answer was a resounding "no," practically and financially speaking. Word 6 for the Macintosh is a powerful and large-scale program, requiring a computer with 8 megabytes (MB) of memory (RAM) and 20 MB of hard disk space, with the 68040 processor recommended. As of the Spring of '96 the English Department had about 50 Macintoshes on inventory, of which only about 15 had the RAM and the hard-drive space to run this program at all, and only a handful had 68040 or PowerPC speed. Though initial software costs would be minimal, since the University has the site license for the upgraded version, this only applies to computers on University property; instructors who had their own legitimate copies of Word 5 would have to pay the Bookstore's price for the upgrade and get the 16 floppy disks that the new program takes to install. [The Byzantine details of site licensing have never been clear to me, I confess.]
Now, let's assume for the purposes of this example that 100 instructors--lecturers and grad students--would be affected. That figure of 100 instructors covers a wide range of computer equipment, each adjusting to the upgrade in a slightly different way. For illustration, say (optimistically) that 25 instructors of the 100 have beefy enough machines to run Word 6, 50 are running Word 5 but could not run Word 6, and 25 are running Word 4 (in other words, are not having a problem with Word 5 but would have one with Word 6).
For those who can upgrade this means a dollar cost of 25 x N for the new version for appropriately equipped home computers [I am still searching for the exact figures to plug in here, but the method should be clear]. How much time would it take to execute the upgrade? Minimally and optimistically, a half an hour? That's 12 hours. Count on 24 hours just for the home computers, given the problems that can ensue with all upgrades. Some instructors do not have powerful enough Macintoshes at home, but would try to install the new program anyway. The resulting confusion would cost some people dearly, but for now let's just assume that 12 hours total might be spent on aborted installations. We are at 36 hours of lost time and we haven't even scratched the surface.
Now we have 25 instructors happily running Word 6 on their home computers. Let's not forget, though: launching Word 6 takes several minutes on a Centris 610, as opposed to 30 seconds for Word 5; let's suppose each instructor launches the program five times a week (conservatively), that's an added half-hour a week spent simply waiting for a word-processing program to load. Because of Word 6's exorbitant RAM-hogging characteristics, users would be less likely to leave Word running at all times--especially if they upgraded their Excel and Netscape versions as well. In aggregate, 5 hours a quarter times 25 instructors and we have 125 hours of wait time, just for the lucky ones who can follow IT's lead and move to the new version on their home machines.
For the 50 instructors who bought Mac Classics or II's or SE's years ago, Word 5 has run just fine, albeit slowly; for them, a 300K conversion extension must be installed in each and every computer's "Word Commands" folder, or they will not be able to read student Word 6 files sent over e-mail or by disk. This conversion extension is available from IT-Express, but each individual must procure a floppy with this conversion extension and properly install it at home. It should be noted that this work-around converter was not effectively announced and immediately promoted, despite its importance. For the sake of simplicity we won't count the trip to Shields to retrieve the converter, instead assuming that departmental technical staff will make the conversion available on their own time; we should, however, add perhaps another 5 minutes of installation if all goes well (and don't forget, we're talking about a group of people who probably don't know where their Word Commands folder is, or even whether they have one!), for about four hours of conversion time.
Now let's look at the daily lives of the people working with the conversion extension. These 50 people will have to go through another dialog box each and every time a document is saved, and will go through a conversion (how many seconds?) each time a Word 6 document is opened. Multiply this out by some conservative figure--say, five documents a day, five days a week, and you're talking 1250 operations a week, twelve thousand per quarter, times however many seconds the operation takes! The aggregate time taken by the upgrade is becoming significant, when integrated over the months and years of opening and closing files that constitute the instructors' work life--and that is completely un-compensated time, time subtracted without payment or other useful benefit as far as the basic word processor is concerned. Even at the conservative figure of 20 seconds per operation, this means roughly 65 hours per quarter. The numbers add up in a hurry.
The calculations given thus far do not account for the 25 instructors who are running Word 4 on older machines; they must remember to use Save-As-Word-4-or-Word-5 when they are in the labs, or their document will not be open-able at home. At least 25 instructors fall into this category; if they open 5 documents a day, 5 days a week, this is an extra 625x2=1250 extra steps (for an extra step is necessary at open and close) per week, or about 15,000 extra steps per quarter. Let's suppose each step takes 15 seconds on the older machines--that's about 60 hours a quarter spent on new conversions and translations. But what happens if one of these instructors forgets to Save-As, and brings a document home that cannot be read or opened? Even with a batting average of .950, this unfortunate situation would occur 750 times a quarter; the forgetful instructor would be unable to open a file until the next time s/he returned to an appropriately equipped machine.
So far I have shown a department-wide grand total of 250 hours of time per quarter added to routine word-processing chores in the wake of this upgrade--about a person-month of wait time per quarter, exclusive of the original cost of installation. To determine comprehensive dollar costs we must multiply this figure by some hourly rate; I leave the multiplication to someone braver than I am. As I hope this discussion has shown, a stepwise consideration of common operations reveals hitherto unaccounted-for costs in time and convenience accompanying upgrades such as this one.
I should add here that such numbers would be less depressing if there were any pedagogical benefit from the new program. Alas, the features offered have little or nothing to do with the day-to-day needs of the vast majority of users, who are more concerned with getting content on the page, with editing and polishing, than they are with the formatting niceties and power-user capacities that justify most upgrades now.
Such brute-force arithmetic, my dogged application of the "dismal science" of economics, is indeed mind-numbing and daunting, yet consider the alternative solution, the one most popular today and the one cheerily suggested by IT's "Recommended Solutions" handouts--that is, to buy new hardware to run a program that has made the previous computers obsolete. This is an intriguing topic, with serious economic and philosophical subtexts, but all over campus and all through industry, there seems to be a newer-is-better mindset that does not truly account for costs. If the 75 instructors unable to run Word 6 were to buy new computers at, say, $1300 or $1500 apiece--about the cost of a mid-line or low-end new Performa at Spring 1996 prices--that means an expenditure of about $100,000 for the instructors as a group, a fair piece of change spent mainly because of a single upgrade. Naturally there would be hidden costs and frustrating time spent on back-ups and installation, more demands placed on a department's computer-support infrastructure, since more of the upgrading would be done by owners rather than a single entity reaping the benefits of economies of scale. Granted, there are many reasons (such as networking and integrated packages) that would also drive new equipment purchases, but for word processing alone old machines are perfectly adequate.
Thus at the most basic nickels-and-dimes level, the upgrade to Word 6 has consequences unforeseen by most people. In particular, the aggregate costs, especially the hidden time costs, are rarely figured in to the decision-making progress, at least using the time-and-motion-study methodology that I have employed here. Certainly my numbers can be quibbled with individually, yet the overall concept is, I feel, worthy of closer attention as we plan for the electronic future in education. To ignore the banal consequences of such moves is as irresponsible as to ignore the environmental consequences of a dam or freeway--yet the mechanism for such review seems to have been neglected by the industry as a whole and by educational technology providers in particular.
All of this number-crunching would be rendered moot if the product to which we were switching offered any pedagogical advantages over its predecessor. Yet there is absolutely no evidence that Word 6 would increase any English teacher's productivity in the classroom or at home, with the possible exception of the half-dozen or so instructors who have bought Macs recently enough to have the new program bundled in.
When I wrote the original draft of this report I asked, "Would the program actually run as promised?" but I soon had to ask the more plaintive "Does the program actually run as promised?" Even as he was touting the program's strengths, IT's lab installations manager told me almost in passing that the upgrade would naturally bring "bugs that can only be discovered on a large scale trial." I could look forward to perhaps several weeks of crashed systems, mangled files, and panicky calls from instructors--with students and teachers, of course, the unwitting guinea pigs in this bizarre little experiment. The IT-Labs team would definitely be able to fix these problems, I could be assured, but we knew as professionals that problems were the price of progress, and this program represented progress.
In fact as it turns out we were both prescient: this Fall the classrooms were slower and more prone to crashes than they have been in years. Composition instructors, especially those in Surge IV and in upstairs Olson, complained of interminable launch times and frozen computers, and on at least a dozen occasions in the first two weeks of class entire systems locked up. Tests run by IT-Labs in October were apparently unable to get Word launched on all computers in a lab without crashing significant numbers, and toward the end of the quarter various reconfigurations were finally forced through. Despite being urged to use less bulky low-tech alternatives like SimpleText or Daedalus Write, instructors found that even having a handful of students forgetfully launch Word 6 by double-clicking on a file icon could bring the classroom network to its knees. Many instructors responded by abandoning the computers entirely.
In most cases of freeze-ups, all site attendants can do is power down the computers and start from scratch, resulting in a cumulative loss of dozens of hours of class time. There is no escaping the conclusion that this level of performance was not acceptable. We had gone from a set of classrooms in which most software worked last year to a set of classrooms where most things do not run as well. As of this writing (January 1997), we seem to have a workable pair of classrooms in upstairs Olson; the jury is still out on the trouble-plagued Surge IV lab.
It's important to recognize that such disruption would never be condoned in standard classes, and should not be condoned--or created!--in computer classrooms. As I have tried to point out, if someone entered a regular classroom and started a noisy chain saw, or suddenly tore up every student's work, there would be a huge outcry--yet the equivalent of these interruptions happened too frequently in recent months and years. For these and other reasons we must work toward a more sensible and pedagogically sound approach to classroom planning and management, to get the most out of the resources we have and to ensure better teaching and learning conditions for all.
All things considered, then, as a package the program has little to offer us, and as a set of present and future problems and headaches it seems to have been a bad decision to pursue. But as near as I can tell, all technology providers feel pressured to have the latest and greatest versions of everything; such policies are even codified in strategic plans like that of the 1992 IT Strategic Planning Committee. This pressure has its corollary in an unspoken rule of the computer industry: that upgrades offered must mean upgrades accepted, no matter what the costs in human terms, in dollar terms, in productivity and morale terms. The software industry's product cycle is so short that in case after case, new releases are adopted and endured, with bug fixes patched in only after hapless users have test-driven the programs. Windows 95 has offered a cautionary example of this, with the first large-scale defection away from upgrade mania: large numbers of frustrated users de-installed the new system, at least until more of the bugs were worked out.
At this point it is worthwhile to examine the ways in which IT gets its information and feedback about software choices. In the process I will suggest a way for information-technology providers to retool the ways they serve their constituencies--even how they define their constituencies and what they owe their users. I intend to apply this discipline in my own service on the 1997 Provost's Commission on the Future of Information Technology, in hopes that future generations of students and teachers will work with appropriate technology, not just the latest technology.
In staving off the mid-year upgrade I tried to argue that IT-Labs should not try to accommodate the most sophisticated user, but this line of reasoning fell on deaf ears. I used a homely analogy, observing that just because a person owns a $1000 ergonomic desk chair at home doesn't entitle him or her to demand such a chair from UCD in his office. A person who is used to the "look and feel" of a Lamborghini (or even a Lexus) can't expect to be issued one as a state car; we expect that worker to deal with the motor-pool Plymouth Reliant, adjust driving habits accordingly, and leave the Sharper Image driving gloves and slippers at home. The argument that "I can get from Davis to a meeting so much faster and more comfortably in a BMW, so I have to have that BMW" would probably not cut much ice, especially in this cost-conscious age, yet such lines of reasoning are not far off. Yet IT gets pressure from students and faculty alike to accommodate ever-increasing RAM and processor-speed demands, as a small minority of instructors insist that only with Netscape 3 and PowerPoint 5 will they be able to do their jobs, and that IT should stock its labs only with high-end machines.
On what sort of market research are these decisions based, and by what means is the demand manifested? It's worth a closer look at what constitutes "demand-pull" in the information technology market. When I have talked to IT-Lab Management administrators they assure me that they (and IT-Express) got calls from users asking about the software running in the labs, and that people were disappointed that Word 6 is not yet available. This may well be true, but who was it asking for the program? Most likely, people who have bought computers bundled with Word 6, and who were wondering whether they would have problems with compatibility. A follow-up question would clarify the most important point: just how many of these users were coming into IT labs with reports tied to spreadsheets, using the high-end functionality of Microsoft Office's Word and Excel? Remarkably few. Even if they were, they were probably the people most able to deal with the "limitations" of the earlier version--and why should we sacrifice hundreds of classroom hours, as we did this past Fall, for their sake? People didn't call up IT to confirm that Word 5 was still being used, but that doesn't mean that there was no demand for it. For the rank and file user, the prime criteria are stability and reliability, a predictable response and not the latest features. Telephone calls and complaints from high-end users should not be the basis for software choices, at least without more thoughtful reckoning of the kind I have described here.
Another reason cited for moving to version 6 was that IT-Training was using this version in its staff training classes--yet this is circular reasoning, and self-fulfilling. Because of administrative pressure, staff training was focusing on IBM users, I was assured, and IBM users wanted the de facto standard, Word 6 (Word 7, now, or perhaps 8). This in itself was supposed to convince me: we couldn't have different versions running, could we? This ignored the fact that at least two of the labs--241 and 247 Olson--are supposedly classrooms first and labs second, and that classroom use should take precedence over convenience for technical staff. This latter issue is currently being considered by IT-Labs and the Registrar--that a "one size fits all" philosophy of do-everything labs might be an outdated paradigm: just as we don't try to hold swim classes in the gym, we shouldn't try to use the same facility for teaching composition that we use to facilitate e-mail reading and Web surfing and spreadsheet construction.
It didn't (and still doesn't) seem to register that the vast majority of institutional users are using the basic word-processing features of Word, not the added 5% of functionality provided by Word 6. An informal canvass of English Department staff personnel revealed that linking spreadsheets was rarely if ever done, and quickly forgotten even if covered in the training. In fact, a common complaint about IT's training was that the instruction did not carry over well to the regular office, since the equipment and software were so different, the supporting materials so user-hostile, that the solutions offered bore so little relation to their jobs.
Thus, it seems that the self-fulfilling cycle of obsolescence is being fomented institutionally among staff training classes and in student labs, though I have yet to see any convincing case that upgrades in themselves lead to greater productivity. Are department MSO's and Technical Services Coordinators aware of the full consequences of upgrades, and trained to spot the fallacies of implicit faith in technological solutions? So far as I can tell, not aware enough--since they have so many other responsibilities and since the technical landscape is changing under their feet so rapidly. This is not a problem limited to the Davis campus, I hasten to add; however, a growing technological awareness in the industrialized world has challenged the underlying assumption to accept all technical change as good. It should be reasonable to believe that we should not accept upgrades unless they have clear pedagogical and productivity benefits across the spectrum that they are destined to cover..
The broader and more disturbing question arising from this whole discussion encompasses key issues of computer culture as it is manifested in what is becoming known as the Microsoft Age: does IT as a division actually have a mechanism for measuring the effects of changes in technology like this? When an upgrade is contemplated, does a group of people sit down to discuss how the change will affect the hundreds and thousands of users, directly and indirectly? Some of the decisions that have been taken in the past ten years would certainly suggest that no such unit exists. More profoundly: Is there a productivity criterion in these decisions? How is it measured? Or shall we continue to accept as an article of religious faith that new and improved software will naturally make us more efficient and productive workers? In case after case, this is not true. With every added feature, it seems, comes new bugs and conflicts and crashes.
Despite the widespread growth in the use of the World Wide Web for research and information retrieval, the number one application of computers on campus is still word processing. But I have yet to see the evidence that writers who use Word 6 are better or more efficient or more productive writers than those using Word 5--or even MacWrite, for that matter. The biggest convenience of a newer version is usually that it doesn't gag on files created by the new version! Ask most composition instructors and professional writers and you'll hear them say, "Give us cut and paste, some control over margins, headers and page numbering, maybe a few fonts and a spell check, and my life as a professional writer is just fine." Anything more is like swatting flies with a sledgehammer--and a mighty awkward and inefficient sledgehammer at that, if you consider some of the idiotic Friendly Thesaurus and Grammar-Chek "features" that purport to put the "up" in most upgrades.
I wouldn't be surprised if decisions are rarely thought through in this manner, either in IT or in the information industry at all, because to think of change in such terms requires an attention to cause and effect that is rare enough in the world in general. In my teaching of interdisciplinary writing I have worked many times with Environmental Studies courses, and am fascinated by the ways ecologists and environmental scientists discover links to links to links along food chains and across seemingly disparate geographical and evolutionary boundaries. Yet students, like bureaucrats and politicians, are rarely trained to spot these links, and are almost inevitably ill-equipped and disinclined to articulate the complex causal relationships among them.
Other entities are forced to think through consequences before making significant changes. When a campus unit considers building a lab or paving a road or constructing a parking lot, there is at least some analysis of costs and benefits before the construction begins. If the bike patrol suddenly decreed that a heavily-used bike circle would be "upgraded" to serve only bicycles with 700mm tires--with the rationalization that all the best racing and fastest bikes use 700mm tires--and that bikes with smaller diameter tires would no longer be able to travel around the circle without a special tire adaptor available from a single site on campus, there would naturally be a tremendous outcry. To extend an already strained analogy, would the campus community be pleased if the change to special 700-mm-only bike circles meant that the bike tire would no longer roll into individual driveways at home? I think not. Yet absurd as it is, such a scenario has been played out time and again, as the market-driven logic of the computer industry reflects itself in the policies and practices of the Information Technology Division.
What is needed is no less than a change in the way we view changes we make in our technological environment. Do we have a course in techno-ecology, or an office so designated, to explore the complex interconnections amongst hardware, software, operating systems, applications, and above all their users? If so, where is the evidence that such analysis takes place? In the field of International Agricultural Development, for example, the notion of "appropriate technology" has slowly taken root, after aid agencies and governments learned that the hard technology path, with increased dependence on fertilizers and machinery and pesticides, was usually not the best or most sustainable course of action for a region or an economy in the long run, despite the promise of increased potential yields and despite the seductions of shinier machinery or glamorous advertising. Instead, a more appropriate synthesis of technologies and behavior patterns may ultimately lead to greater productivity over a longer period, and that the keeping-up-with-the-Joneses argument is neither affordable nor even effective under real-world conditions with real-world limitations.
I have been Coordinator for Computer Aided Instruction for the English Department for two years, and I have on several occasions tried to ascertain just how IT makes some of the decisions it makes. As with other campus entities of its size, the answer is not simple. I recall a discussion with Tim Leamy where I was doggedly pursuing these lines of questions, and convincing him to delay the implementation at least to the summer, not over spring break, and Tim became somewhat exasperated: "So when are we going to change? When everyone else has gone to Word 6 or Word 7? How long do we wait?" At that moment I realized that there was another dynamic at work. If we didn't make this upgrade soon, the tone of his voice plainly suggested, UCD's labs would be left behind, years out of fashion, as obviously passˇ as bell-bottoms or disco or IBM Selectrics. To even suggest that the wheels of progress might be slowed for such self-inspection was heretical. Now don't get me wrong--I have a good professional and personal relationship with Tim, and a great respect for his ability to troubleshoot networks, yet on this one Tim was giving me this odd tolerant look, the same look that I could imagine giving a street person arguing for a flat earth theory.
I should note, however, that conversations with counterparts at other institutions assure me that this is not a problem peculiar to UC Davis' IT division, although I have used those initials quite frequently in this critique. Moreover, things at UC Davis are improving with new managerial blood and new sense of cooperation and collaboration. Thanks to meetings with the lab planning committee, English Department lecturers were able to give their ideas about configuring a new Olson basement computer room to be more effective for teaching. Complaints about the unworkability of 307 Surge IV resulted in a subtle rearrangement that profoundly improved teaching conditions. Recent reconfigurations of the upstairs Olson labs have made Word 6 run successfully--though this has meant installing the word processing package on every hard drive, adding a 3Com hub, and more wiring--and IT seems much more committed as an institution to serving its almost impossibly varied constituencies.
The solution to the immediate question of what word processing package to use is probably quite simple, if only I knew whom to ask. Is there a program out there that can, with a double-click of the mouse, open and save documents, crack all of Microsoft Word's secret codes, launches quickly, works reliably, has all the features we need, and is also inexpensive enough to convince IT-Labs to purchase? Is there someone in IT who has the answer to this question, a suggestion for the low-tech low-impact way out of the whole dilemma? Certainly the site license couldn't be more expensive than the thousands per quarter that I just determined as the price tag for moving to Word 6? Perhaps as a result of this report something good may emerge, some reader will give me the key to the maze, some brave soul will come forward with the magic software bullet that will slay the Word 6 dragon and solve all of the technical problems so that we can go back to doing what we set out to do: teach students how to write and think more effectively. Suggestions welcome, advice asked, credit given.
However, the deeper questions of technology and upgrade-mania remain, here at Davis and in schools and businesses everywhere. In my own professional consulting and in conversations with peers at other universities and in industry, I have heard these concerns echoed again and again--stories of upgrades that went bad, of reversions to previous systems, of lost time and stunted productivity. It is even more important to note that the same concerns will apply, only more so, for network bandwidth, switching capacity, and modem access, a grim reckoning that is at hand for everyone growing more dependent upon the Internet. At some point the enthusiasm for new software that requires new hardware must eventually face the limits of money and time; we must more adequately measure incremental increases in possible productivity improvements against the far more significant costs of change.
Approaching future upgrade decisions with a developing discipline of techno-ecology must be a priority for any institution that uses computers; for an educational institution the analysis must place a premium on the pedagogical impacts of proposed change. For too long we in the Composition Program (and other lab-classroom users) have succeeded in spite of, and not because of, decisions and policies that ignored pedagogical concerns. It is time for us to reaffirm our focus on improving teaching, analysis, and writing; if this means remaining behind the upgrade curve, let us make the most of it.
Begun Spring 1996, finished January -February 1997
E-mail comments or suggestions on this piece to the author.
Return to Composition Home Page