Work-Arounds, Make-Work, and Kludges

Human-Centered Computing

Work-arounds, Make-work, and Kludges

Philip Koopman, Carnegie Mellon University Robert R. Hoffman, Institute for Human and Machine Cognition

aradigms are often defined partly in terms of what Work-around We all have had to create and use a work-around at some they are not, or in terms of what they are reacting P point to get a user-hostile computer to do our bidding. Doc- against. The paradigm of human-centered computing is no umented evidence attests to the pervasiveness of work- 3 exception. In response to an essay in the Jan./Feb. 2002 arounds. Sometimes the work-around is inspired by a friend’s hint; sometimes it’s spelled out in a help page on Human-Centered Computing column “The State of Cog- the Web; sometimes it’s discovered by flailing around until nitive Systems Engineering,”1 we had a lengthy discussion something just happens to work, more or less. Although on the question, What is a user-hostile system? The fol- work-arounds might seem inherently ad hoc, in many situa- lowing quote is from that essay: tions they make the difference between success and failure. The Concise Oxford Dictionary definition of work- The road to user-hostile systems is paved with user-centered around is “a method for overcoming a problem or limi- intentions on the part of the designers. Even smart, clever, tation in a program or system.” (Neither the Oxford Un- well-intentioned people can build devices that are fragile and hostile, devices that force the human to adapt and build local abridged Dictionary nor the Oxford Dictionary of kludges and work-arounds. Worse still, even if one is aware of Computing lists “work-around,” but, ironically, the Con- this trap, one will still fall into it. cise Oxford Dictionary does. Go figure.) This definition doesn’t go far enough to pin down the concept so that it We decided that the terms kludge and work-around, might be more amenable for study. Does the “problem” and also the related concept of make-work,have yet to be being overcome mean a design defect or something as sim- clearly defined for the intelligent systems community. ple as a component failure due to wear-out? Does “limita- Human-centered systems are different from user-hostile tion” mean an unimplemented (but clearly imaginable and systems as well as from systems based on a designer- clearly useful) feature, an intentionally precluded action, centered approach.2 In this essay, we try to clarify the or an inability to use the tool to cope with an unanticipated senses of these three terms and suggest ways we might operating environment? And is the work-around method study work-arounds, make-work, and kludges as an inte- written down, foreseen by designers, or made up on the gral part of human-computer systems—rather than as fly? embarrassing necessities that are best swept under the Wikipedia, a Web-based encyclopedia (www.wikipedia. computing research rug. org), defines work-around as

A bypass of a recognized problem in a system. A workaround is typically a temporary fix that implies that a genuine solution to the problem is needed. Frequently workarounds are as cre- ative as true solutions, involving out-of-the-box thinking in their creation. A workaround is not a permanent solution. Typi- cally they are considered brittle in that they will not respond well to further pressure from a system beyond the original design. In implementing a workaround it is important to flag the change so as to later implement a proper solution. Placing pressure on a workaround may result in later failures in the system. For example, in computer programming workarounds Editors: Robert R. Hoffman, Patrick J. Hayes, and Kenneth M. Ford are often used to address a problem in a library, such as an Institute for Human and Machine Cognition, University of West Florida incorrect return value. When the library is changed, the [email protected] workaround may break the overall program functionality, since it may expect the older, wrong behavior from the library.

70 1094-7167/03/$17.00 © 2003 IEEE IEEE INTELLIGENT SYSTEMS Published by the IEEE Computer Society Whatis.com, a Web-based glossary (http:// A block can occur when you don’t know Sense 1: Completing tasks despite whatis.techtarget.com), defines work- whether a path to your goal even exists design flaws around as (that is, you’re working on a system with invisible and therefore unknown function- Work-around (1): A procedural change in A method, sometimes used temporarily, for alities). A block can also occur when a computer system use intended to compen- achieving a task or goal when the usual or known path is confusing, laborious, bro- sate for a design flaw, typically a software planned method isn’t working. In information technology, a work-around is often used to ken, or otherwise hostile. In this case, you behavior that is perceived to be a flaw. overcome hardware, programming, or com- might create a new path that you perceive munication problems. Once a problem is as either more user friendly (for example, Perhaps the most common experience of fixed, a work-around is usually abandoned. intuitive or transparent) or simpler and a work-around with computers is getting hence better able to expedite the work. In buggy programs to work well enough to This last statement seems to point to an empir- both cases, the method you create might or accomplish a task or achieve a goal. Soft- ical base that, as far as we know, doesn’t exist. might not actually be simpler than that ware companies often issue work-arounds to Be that as it may, all three definitions seem included in the existing functionality. help users cope with bugs. While you can to capture the usual intuitive notion and are We propose grouping four alternative argue that almost any nontrivial software all right as far as they go. However, they sug- senses of work-around, hinging on the will have defects, desktop office automation gest trying to unpack the different senses of nature of the block: software gets a lot of press for being plagued work-around might be useful. with bugs. An example that came up on a Les Gasser defined three alternative senses simple Web search was Microsoft Security of what he termed adaptations: fitting, aug- Bulletin MS02-027, which states in part, menting, and working around.4 Fitting jobs A block can occur when you don’t “Customers using IE should implement the and work schedules to accommodate com- work-around detailed in the FAQ.”5 This puting-resource limits is an adaptation you know whether a path to your goal particular work-around is a set of steps the can make to avoid a computing-system fail- user performs to reconfigure Internet ure due to overload. Augmenting involves even exists. A block can also Explorer. removing anomalies from the computing Some definitions (for example, in Whatis. environment to avoid triggering problems, occur when a known path is com) state that a work-around is a temporary such as scrubbing data and training users to compensation until a defect can be corrected. avoid making certain types of mistakes. confusing, laborious, broken, or This is the meaning we also find in Eric Ray- Working around involves users altering input mond’s Jargon File.6 MS02-027 is an exam- data or procedures to compensate for system otherwise hostile. ple. It involves disabling Gopher service until shortcomings, or using backup systems. a patch is released that eliminates a buffer When a work-around is not readily avail- overrun security vulnerability. Also, some able, people might even change their goal to work-arounds are only partial solutions. For something that they know the system can do. • Completing tasks despite design flaws in example, MS02-027 actually disables a ser- For example, if sending a file as an email a computational tool vice and leaves users on their own if they attachment doesn’t work, a potential work- • Completing tasks despite component must use Gopher for something, but it does around is putting it on an FTP or Web server failures improve the security of using IE for other and just sending a pointer to the intended • Extending functionality to complete a tasks. recipient. However, we look at this solution new task Sometimes work-arounds are put in simply as a work-around taking place at the • Intentionally evading designed limits in place not because the software itself is next level up from a person’s interaction with an effort to overcome them defective or is perceived to be defective but a computer. Gasser reported the common because the software requirements or oper- practice of using informal processes based This set is based on the notion of a system ating environment are inadequate. The on personal relationships to expedite organi- problem or limitation as expressed in the designer’s path through a system might be zational processes, with computer-based dictionary definitions given earlier and on so confusing, laborious, or otherwise hos- workflow activity backfilled after the actual Gasser’s emphasis on the intent of the per- tile that people create their own paths. task has been performed. Gasser gives the son exercising the work-around. Of course, These paths might or might not be “right,” example of a purchasing manager handwrit- you can use the mechanisms for executing but are at least perceived to be friendlier ing a requisition to avoid delay and then hav- a work-around for more than one of these (for example, intuitive or transparent), sim- ing the requisition entered into a computer purposes, and we’ll discuss them as we go. pler, or less time-consuming. after the fact for tracking purposes. (We’ve heard of a fifth meaning of “work- Generally you can think of a work- We define the general sense of work- around,” as a personality trait. Trainers might around in Sense 1 as a situation in which a around as follows: attribute this to their students, asserting that user tries to follow a set of well-defined some individuals seem prone to ignore what steps to accomplish a particular goal, but When a path to a goal is blocked, people use they’re told. We don’t take seriously the idea something goes wrong that blocks the user their knowledge to create and execute an that “working-around” is a personality trait, from accomplishing those steps. For exam- alternate path to that goal. on par with, say, sociability.) ple, the user might encounter some mind-

NOVEMBER/DECEMBER 2003 www.computer.org/intelligent 71 boggling complexity in the steps’ descrip- Some work-arounds are necessary tion scheme.9 This differs from the other tion or a software design flaw that prevents because the computer or software as origi- types of work-arounds in that the user is completing those steps in a particular set of nally designed simply doesn’t address the trying to do something that the system circumstances. For instance, Stephan Poel- problem or task at hand. This can be because designers specifically intended the user to mans describes work-arounds as deviations the task you’re trying to accomplish is not do, or intended to prevent the user from from a defined workflow system, whether “new” and there hasn’t been time to make doing, rather than accommodating to a to overcome problems or to improve user software changes. In some cases you can design defect or gap in design requirements. experience.7 foresee this, and you can design work- Some in the field of computer-aided arounds for existing software as part of a design define work-around in this way, as a Sense 2: Completing tasks business process change. But at other times, sort of subversion of a task: despite component failures this might not be practical. Working-around in Sense 3 is a common activity. An example In subversion [as a constraint negotiation Work-around (2): A procedural change to would be to use spreadsheet software to strategy], the user modifies the task approach to take advantage of known weaknesses in the using a computer system intended to compen- compose an essay outline. tool, overriding the spirit but not the mecha- sate for a hardware or component failure. In the area of task analysis, Michael nism by which the constraint is imposed (also Albers argued that goal-driven approaches known as a “workaround”).10 We can expect most computer systems of reasonable complexity to suffer compo- Indeed, in the context of CAD, a circuit nent failures, sometimes regularly. Work- design arrangement that exploits a loophole arounds in such cases involve people diag- Users are left on their own to in automated design-rule checkers will nosing a failure or recognizing a pattern of likely result in a chip design that doesn’t system misbehaviors and then executing an determine a work-around because work. To increase productivity in safety- alternate procedure to accomplish a goal. critical systems and other critical applica- The simplest work-around strategy in the tions, people might use work-arounds to face of a component failure is to have a the number of potential failure and similarly evade safety features, but they do backup system (computerized or manual) so at a cost of increased risk. Donald Day you can use in a pinch, even though it use extension scenarios exceeds conducted a study of work-arounds in this might have limited functionality.4 sense of subversion, based on results from Component failure work-arounds are the currently known techniques a questionnaire concerning the factors that similar to design defect work-arounds but affect flexibility when working under soft- differ in a few key ways. First, a completely for analysis and documentation. ware-imposed constraints.10 redundant backup system is often helpful Approximately 200 professionals in com- for component failures, but it might well puter-aided systems engineering completed have the same design defects as the primary the questionnaire. While most respondents system and thus might not help with design (that is, ad hoc activities) rather than a pre- reported that they didn’t feel especially defects. Having heterogeneously designed planned task analysis can be necessary when encumbered by their software tools, most systems is a way to provide at least some there are failures or exceptional operating also reported that they absolutely reserved work-around capability for both design and conditions.8 This reflects the fact that in the right to override or work around soft- component failures—for example, having many instances users are left on their own to ware-imposed constraints, assuming suffi- both a Windows and a Unix desktop com- determine a work-around because the great cient justification exists (for instance, to puter, or installing multiple Web browsers number of potential failure and use exten- adhere to overriding professional standards from competing vendors. sion scenarios exceeds currently known or to deal with an obvious software design One good thing about component failures techniques for analysis and documentation. fault). Those who felt “controlled” by their compared to design failures is that it might software tended to report lower levels of be possible to fix a component failure very Sense 4: Users intentionally misleading satisfaction with it and a higher likelihood quickly, whereas corrections of design their computers of being subversive. Conversely, those who problems can take a long time to create, and felt that they were less controlled by their work-arounds for them might require trial Work-around (4): A procedural deviation software tended to report greater levels of and error. So, in some cases a simple adap- to circumvent an intentionally designed-in satisfaction with the software and a greater tation strategy of rescheduling work hours limit or constraint on computer system likelihood of conforming to the constraints to wait for a repair might be possible operation. that the software imposed. instead of using a bona fide work-around. By implication, Day is saying that evad- Sometimes we use work-arounds to ing design features is a constructive activ- Sense 3: Extending functionality evade a computer system’s designed-in ity in some circumstances. Such instances behavior. This can be because we’re trying abound in the area of graphics software. Work-around (3): A new procedure that to use a computer to do something it wasn’t Alison Black studied how novice graphic uses a computer system in a way not origi- intended to do, such as holding down the designers get “sucked in” to the conven- nally envisioned to accomplish a task. shift key to defeat a music CD copy protec- tions and limitations of graphics software

72 www.computer.org/intelligent IEEE INTELLIGENT SYSTEMS that was created without benefit of a user Kludge kludge and a work-around is that a kludge needs analysis.11 She noted the work- Webster’s Ninth Collegiate Dictionary is a set of design artifacts or elements, in arounds that users created to cope with defines kludge as a system (especially com- one of these forms: awkward and inefficient commands in puters) made of components that are poorly graphics software (for example, the prob- matched or were originally intended for •A fix that is awkward or clumsy but is at lem of filling in enclosed objects that are some other use. This source indicates that least temporarily effective created piecemeal). the word has unknown origins but cites • An overall design that is of questionable Gasser interviewed approximately 60 1962 as the earliest recorded use. This elegance or downright ugly professionals who worked at several manu- might refer to a paper in which Jackson facturing firms to study how they adapt to Granholm,13 his tongue firmly in his cheek, On the other hand, a work-around is computing resources.4 He reported several suggested several rules of applied “kludge- generally considered to be a procedural instances in which users “gamed” their manship,” exemplified by a number of variation that a person creates and uses. computer systems by entering data known ways in which hackers could create a vari- Raymond’s Jargon File sees a kludge as an to be inaccurate, or otherwise not following ety of clever kludges. For example, when initial design approach and a work-around expected normal system use. But the prac- using magnetic tape as a storage medium, as a temporary bug fix.6 On the other hand, titioners were able to get their systems to the “kludgeman” would use both odd and most other sources, and our own experi- yield usable results, ones that sometimes ence, indicate that using “kludge” rather were more “accurate” than those derived than “work-around” is more appropriate by “following the rules.” An example of for temporary fixes. In any event, we feel this in everyday life is booking a confer- Kludge and work-around are that the distinction of procedure versus ence room for a critical meeting for a time designed artifact is useful, one that accords block starting a half hour early on the related in that both originated with most documented definitions. This expectation that any preceding meeting distinction seems largely consistent with will run later than scheduled. Here’s an various documented usages in which example from Gasser’s report: because an alternative solution kludge applies only to hardware and work- around applies to both procedures and Engineering analysts working [at] a large that is either more elegant or software. firm which designed and constructed chemi- Kludge and work-around are related in cal plants, learned which analyses could go wrong in the package of complex analytical more appropriate is for some that both originated because an alternative programs they used, and how to correct for solution that is either more elegant or more untenable results. For example, they rou- reason unavailable. appropriate is for some reason unavailable.13 tinely input temperature coefficients for pipes These concepts are also related in that they carrying hot fluids as though the pipes were intended to operate cold, causing the analysis are both, apparently, permanent residents of program to disregard certain heating stress the technology landscape: calculations. Over years of experience with even parity and as many widths as he could this program, the engineers had learned that find reels for. Hardware and software products are some- entering the “correct” information would We’ve seen the word spelled both “kluge” times the result of adding a new and basically lead to erroneous results and the pipes would incompatible design to an original design not work properly in the final design. They and “kludge” and have heard it pronounced rather than redesigning the product completely. worked around the technical problems of the both as “klooge”(as in “stooge”) and Users often have a different opinion than the program by “running the hot pipes cold” “kludge” (as in “fudge”), with the former designers do. To the extent that information 4 when using it. being the pronunciation most people prefer. technology products are combinations of elements originating from a variety of design Granholm traces the word to the German philosophies and constraints, almost any prod- The positive value of work-arounds is a “kluge” (pronounced “kloo-ga”), meaning uct is bound to contain some element of recurrent theme in all the empirical work smart, witty, or clever. We wouldn’t be sur- kludginess.14 we found. One more example is a study prised if the term was in fact an acronym by Gary Parish and William Sollfrey,12 for “knowledge and learning used to derive Make-work who examined work-arounds’ effects on good effects.” Nor would we be surprised if Make-work activities are repetitive, bor- unmanned spacecraft, including performing the term comes from the last name of an ing, time-consuming activities that some- reliability calculations that incorporated absent-minded German professor who was one must engage in to accomplish some- the expected benefits of work-arounds. notorious for making electromagnets from thing that could not be accomplished using They found that even the relatively primi- hairpins, old tractor batteries, and fence a shortcut, or that one should be able to tive satellites of that era benefited from wire. Raymond asserts that the term came easily accomplish but cannot. An example work-arounds, with 18 life-extending into use in the 1940s in reference to a is when a display of visualized scientific work-arounds recorded for 25 satellites device called the Kluge Paper Feeder, a data relies on screen sectoring to permit studied. Their definition of the term “fiendishly complex assortment of cams, the presentation of multiple data types or emphasized procedural and software belts, linkages … devilishly difficult to fields. Sectoring, an attempt to overcome changes and specifically excluded built-in repair … but oh, so clever.”6 screen real-estate limitations, places a great redundancy. The general difference between a burden on the user, requiring repetitive

NOVEMBER/DECEMBER 2003 www.computer.org/intelligent 73 point-click-and-drag minimization and social and organizational aspects of work- This clearly places WKM behavior in maximization to view individual data arounds. When and why do organizations the legitimate arena of “adaptive design,” fields. Another example is having to cut, choose to ignore subversive, work-around in which users help to finish the system paste, cut, then “paste as” to get a clean activities? Individuals aren’t the only ones design.16 Researchable topics abound for chart transfer from Excel to Corel Draw. who create work-arounds. As Day pointed studying the deliberate, systematic creation out, teams and project managers also create and deployment of WKMs. It might be Opportunities them. Why? When? Inspired by Parish and useful to explore ways to build compensa- Now that our semantic deboning of the Sollfrey,12 can we quantify the effects of tion mechanisms into software to make the triad has played itself out, you must be work-arounds on system dependability and user-plus-software system more resilient— asking just the sorts of questions that HCC usability? that is, ways to make work-arounds easier advocates would raise. We began this There’s no denying that WKMs will for users to create, document, and share. essay with a reminder that human-cen- exist as an enduring part of the computing For example, an online help system might tered systems differ from user-hostile sys- experience. But much software culture notice that a user is repeatedly getting a tems in that user-hostile systems are often today is based on the notion of trying to particular error message and might offer based on a designer-centered approach.2 achieve perfect software, which of course is advice on work-arounds relevant to that In this regard, Granholm was right on the an in-your-face manifestation of designer- error. Help systems today assume that the money when he pointed to some of the centered design. Perhaps it’s time to start user is making a mistake and offer canned reasons why kludges and work-arounds tutorial advice, but typically do not con- are necessary: sider that the real problem might be a software defect, limitation, or user hostility. It is sad, but true, that a kludge cannot be Much software culture today is Users must search the Web or other data- designed under just any old organizational bases to find work-arounds, if they’re pub- structure. One of the most helpful atmos- pheres in which a kludge may arrive at full based on the notion of trying to lished at all. flower is that of complete, massive, and Although foreseeing all possible failure iron-bound compartmentalization. It is a achieve perfect software, which modes might be impossible, it’s probably good idea if the I/O men, say, are not only useful to catch and represent likely failure not allowed to talk to the mainframe modes at design time. The only way to do designers, but also that they have, in fact, of course is an in-your-face never met them. Interface cross-talk should, that is to empirically study the “envisioned by all means, be done by edict and direc- world”17 and evaluate software for useful- tive, and not by memo and design note. manifestation of designer- ness and usability. Indeed, safety-critical After all, someone has to hew the line on systems use Failure Mode Effects and Crit- design philosophy, or people will go off in 18 all directions.13 centered design. icality Analysis to represent the effects of hardware component failures and subsys- His sarcasm rings true. Day made a similar tem functional failures. We can extend this point in a literal tone: to look at the ways that users are blocked treating WKMs as first-class citizens in and frustrated as they try to complete User dissatisfaction with some constraint terms of research topics, development sequences of steps. management techniques may be attributable efforts, and teaching. Gasser captured this In addition to possible practical applica- in part to a mismatch between builders’ and users’ views of appropriate and necessary point beautifully: tions, research along the lines we propose design practices.10 might lead to a scientific understanding of A fundamental assumption of design-oriented WKM behavior and a formal representa- Rather than simply bemoaning this state literature is that managers, designers, and tion of work-arounds and kludges.19 system proponents define what are “rational” of affairs, we could see it as an opportunity. actions in dealing with a computing system This involves adopting the goal of Day’s10 —rational procedures for using a system, and Gasser’s4,15 empirical studies: getting problems fixed, etc. From this view- point, workers in our study who worked Instead of studying how to eliminate prob- around formal systems were “escaping” from lems … we are attempting to describe and the constraints of the system, and acting explain the dynamics of computer use over “irrationally” with respect to system goals time. This leads us to focus on how circum- and managers’ expectations. But we have stances persist and evolve, rather than why discovered that, far from acting irrationally, WKMs are here to stay, perhaps even they exist in the first place.4 the informal practical actions of participants after computer scientists take notions of actually make systems more usable locally. Informal fitting, augmenting, and working human-centering to heart. Intelligent and Might not research on work-around, kludg- around are essential and locally rational parts human-centered systems do not blame or ing, and make-work (WKM) activities reveal of system use. Appropriate and rational ac- punish users for making mistakes. Perhaps ways we could achieve human-centering? tion is defined by the demands of the work it’s time for researchers and technologists situation and the institutional arrangements Might it not suggest new methods for study- surrounding computing, not by the ideologies to start including WKMs more explicitly in ing or evaluating human-centeredness? of managers or the presumed necessities of their agendas. IEEE Intelligent Systems Also wide open for research are the system structure.4 should be the home for that agenda.

74 www.computer.org/intelligent IEEE INTELLIGENT SYSTEMS System Performance and Procurement Re- Acknowledgments quirements: A Proposal, Rand Corp. report N-1260-AF, 1980. Philip Koopman Phil Koopman’s contribution to this article was is an associate pro- supported in part by the General Motors Collabora- 13. J. Granholm, “How to Design a Kludge,” fessor at Carnegie tive Laboratory at Carnegie Mellon University. Datamation,vol. 8, Feb. 1962, pp. 30–31. Mellon University. Robert Hoffman’s contribution was supported He is affiliated with through his participation in the Advanced Decision 14. L. Swanson and M.J. Warden, “Kludge,” the Department of Architectures Collaborative Technology Alliance, Whatis.com, 2003; http://whatis.techtarget. com. Electrical and Com- sponsored by the US Army Research Laboratory puter Engineering, under cooperative agreement DAAD19-01-2-0009. arses, “The Constraint Satisfaction the Institute for 15. F.Approach D to Design: A Psychological Inves- Software Research International, and the tigation,” Acta Psychologica,vol. 78, 1991, Institute for Complex Engineered Systems. References pp. 307–325. Contact him at the ECE Dept., Carnegie Mellon Univ., 5000 Forbes Ave., Pittsburgh, 1. R.R. Hoffman, G. Klein, and K.R. Laughery, 16. D.D. Woods and S.W.A. Dekker, “Anticipat- PA 15213; [email protected]. “The State of Cognitive Systems Engineer- ing the Effects of Technological Change: A ing,” IEEE Intelligent Systems,vol. 17, no. 1, New Era of Dynamics for Human Factors,” Robert R. Hoffman is a research scientist Jan./Feb. 2002, pp. 73–75. Theoretical Issues in Ergonomic Science,vol. at the Institute for Human and Machine 1, no. 3, 2000, pp. 272–282. Cognition. Contact him at the IHMC, 2. R.R. Hoffman et al., “A Rose by Any Other 40 Alcaniz St., Pensacola, FL 32501; Name … Would Probably Be Given an 17. S.W.A. Dekker, J.M. Nyce, and R.R. Hoff- [email protected]. Acronym,” IEEE Intelligent Systems,vol. 17, man, “From Contextual Inquiry to Designable no. 4, July/Aug. 2002, pp. 72–80. Futures: What Do We Need to Get There? IEEE Intelligent Systems,vol. 18, no. 2, 3. R. Kling and W. Scacchi, “Recurrent Dilem- Mar./Apr. 2003, pp. 74–77. mas of Computer Use in Complex Organiza- 19. C. Martin, Functional Fault Simulation for tions,” Proc. 1979 Nat’l Computer Conf., 18. C. McCollin, “Working around Failure,” Distributed Embedded Systems, master’s the- AFIPS Press, vol. 48, 1979, pp. 107–116. Manufacturing Engineer,vol. 78, no. 1, Feb. sis, Electrical and Computer Eng. Dept., 1999, pp. 37–40. Carnegie Mellon Univ., 2001. 4. L. Gasser, “The Integration of Computing and Routine Work,” ACM Trans. Office Informa- tion Systems,vol. 4, no. 3, July 1986, pp. 205–225. Addendum 5. “Microsoft Security Bulletin MS02-027,” 28 Feb. 2003, Microsoft, www.microsoft.com/ The September/October 2003 install- of fitting the human body with intelli- technet/security/bulletin/MS02-027.asp. ment of this column was titled "The gent technologies, as opposed to Borg Hypothesis." Jane Abbate, histo- merely adding cybernetic mechanisms 6. E. Raymond, The New Hacker’s Dictionary, MIT Press, 1991. rian of technology at the Chemical Her- such as artificial limbs or even artificial itage Foundation, points out to us that eyeballs. The notion that control theory 7. S. Poelmans, “Workarounds and Distributed the term cyborg was first introduced in ala Norbert Wiener (that is, types of Viscosity in a Workflow System: A Case 1960 by Manfred Clynes and Nathan feedback loops) is sufficient for intelli- Study,” ACM SIGGROUP Bull.,vol. 20, no. 3, Kline (“Cyborgs and Space,” Astronau- gence is at least arguable. The point of Dec. 1999, pp.11–12. tics, vol. 5, pp. 26–27, 74–76). They had our essay was to consider the integra- 8. M. Albers, “Goal-Driven Task Analysis: a similar mission to that described in tion of intelligent technologies into Improving Situation Awareness for Complex our essay: preparing humankind for humans and the implications this might Problem-Solving,” Proc. 16th Ann. Int’l Conf. space travel. have for human evolution. Computer Documentation,ACM Press, 1998, pp. 234–242. But this early vision of a cyborg fu- —Robert R. Hoffman ture did not include the crucial concept To be continued … 9. J.A. Halderman, “Analysis of the MediaMax CD3 Copy-Prevention System,” tech. report TR-679-03, Princeton Univ., Oct. 2003, http:// ncstrl.cs.princeton.edu/expand.php?id=TR- 679-03. 10. D. Day, “User Responses to Constraints in IEEEIEEE on all conferences Computerized Design Tools: An Extended sponsored by the Abstract,” Software Eng. Notes,vol. 21, no. Computer 5, Sept. 1996, pp. 47–50. Computer IEEE Computer Society. 11. A. Black, “Visible Planning on Paper and on SocietySociety Screen: The Impact of Working Medium on Not a member? Decision-Making by Novice Graphic Design- members ers,” Behavior and Information Technology, members Join online today! vol. 9, no. 4, 1990, pp. 283–296. savesave 25%25% www.computer.org/join/ 12. G. Parish and W. Sollfrey, Preliminary Analy- sis of the Effect of Work-Arounds on Space

NOVEMBER/DECEMBER 2003 www.computer.org/intelligent 75