`__________________
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`__________________
`IBG LLC and INTERACTIVE BROKERS LLC,
`
`Petitioner
`v.
`
`TRADING TECHNOLOGIES INTERNATIONAL, INC.
`Patent Owner
`_________________
`Case CBM2016-00009
`U.S. Patent No. 7,685,055
`_________________
`
`DECLARATION OF DAN R. OLSEN, JR.
`
`Page 1 of 124
`
`TRADING TECH EXHIBIT 2174
`IBG ET AL. v. TRADING TECH
`CBM2016-00009
`
`
`
`I. Qualifications
`
`1.
`
`1, Dan R. Olsen Jr., Ph.D., am a resident of Orem, Utah and have more
`
`than 35 years of experience in computer science and human-computer interaction
`
`(HCI). I hold a doctorate in Computing and Information fi'om the University of
`
`Pennsylvania. For 3 ‘/2 years I was an Assistant Professor of Computer Science at
`
`Arizona State University. I then served for 30 years on the faculty of Brigham
`
`Young University retiring as a full professor in 2015. During that time at BYU, I
`
`also served as the chair of the Department of Computer Science. I took leave from
`
`BYU in 1996 to become the founding director of the Human Computer Interaction
`
`Institute in the School of Computer Science at Carnegie Mellon University. I
`
`returned to BYU in 1998. I am currently the CEO of a software startup in
`
`educational technology (SparxTeq, Inc).
`
`2.
`
`During the course of my academic career, I authored over 70 papers in
`
`the field of computer science. The topics on which I have published papers are:
`
`I
`
`User Interface Management Systems
`
`Syntactic representations of user interfaces
`
`Multi-user interaction across networks
`
`Induction of interaction behavior from pictures
`
`Novel interaction techniques using speech and laser pointers
`
`Interactive machine learning
`
`Page 2 of 124
`
`
`
`Interactive robotics
`
`Interactive television
`
`3.
`
`I currently hold 4 patents in human-computer interaction. I have
`
`authored 3 textbooks on the techniques of software design for human-computer
`
`interaction.
`
`4.
`
`I have had extensive involvement. in professional societies, such as the
`
`Association for Computing Machinery (ACM), the premiere society in computing.
`
`I have served in many ofiices of ACM’s Special Interest Group on Computer
`
`Human Interaction (SIGCHI) and currently serve as its treasurer. I have been
`
`conference chair of CI-II, which is the premier conference in Computer Human
`
`Interaction. I was the founding editor of ACM’s Transactions on Computer Human
`
`Interaction. I was a co-founder and active leader for the conference on User
`
`Interface Software and Technology (UIST) for the past 29 years. I have also served
`
`at the governor’s request on the Utah Science, Technology and Research (USTAR)
`
`board, which oversees and fimds state economic development efforts in
`
`technology.
`
`5.
`
`I twice received best paper awards in intelligent user interfaces. In
`
`2004, I was appointed to the CHI Academy for international excellence in
`
`Computer Human Interaction research. In 2007, I was recognized as one of ACM’s
`
`Page 3 of 124
`
`
`
`Fellows for research in computer science and in 2012 received the CHI Lifetime
`
`Research Award, which is the highest award in Computer Human Interaction.
`
`II. Graphical User Interfaces and the ‘U55 Patent
`
`6.
`
`Attorneys for the Patent Holder have explained to me that U.S. Patent
`
`No. 7,685,055 (“the ‘055 patent”) has been challenged as a Covered Business
`
`Method (CBM) patent. I have been asked to review the nature of the invention in
`
`the ‘055 patent. As explained below, it is my opinion that the ‘055 patent claims a
`
`technological invention because the claimed invention provides a technical
`
`improvement to prior graphical user interfaces.
`
`III. Historical context
`
`7.
`
`In discussing how graphical user interfaces are a technology with
`
`specific technical problems, I would first like to refer to two very old patents
`
`involving earlier mechanical technologies. I will use these two patents to illustrate
`
`two key pieces of technical knowledge that are used widely in graphical user
`
`interface (GUI) technology.
`
`Page 4 of 124
`
`
`
`SPEIIHII nuwrna Poll VIHIGLIS
`151'! 3191-20. 1526
`
`2 mm-anm 1
`
`Figure l —« Speedometer in the Steering Wheel.
`
`IV.
`
`Simplification of perception — US patent 1,692,601
`
`8.
`
`In 1928, U.S. patent 1,692,601 was issued for an automobile
`
`speedometer that was mounted in the center of the steering wheel. This patent
`
`claims the ability to perceive an automobile’s speed. It was not for the sensor for
`
`measuring speed (which was well known at the time). It was not for the concept of
`
`displaying speed in a meter (which was also well known). The key claim was as
`
`follows:
`
`first, to provide a steering control means -for vehicles on which is
`
`mounted a speedometer whereby the speed of the vehicle may be
`
`readily determined by merely dropping the vision slightly to the
`
`middle portion of the steering wheel which is substantially in line with
`
`the line of vision of the driver and as close to the eyes of the driver as
`
`4
`
`Page 5 of 124
`
`
`
`possible, thus determining the speed of the vehicle with least danger
`
`to the driver and other occupants ...
`
`9.
`
`This patent’s key contribution was that it placed the speedometer in
`
`the center of the steering wheel where it was easier for the driver to perceive. This
`
`was not awarded for the esthetics or appearance of that speedometer placement. It
`
`was the arrangement of the components of the technology for easiest human
`
`perception that was the key to this patent. It will be shown in this report that the
`
`‘055 patent claims constructing a GUI to display information in particular locations
`
`to improve the interface between man and machine by improving the user’s
`
`perception of the relevant information, with bits and pixels comprising the GUI
`
`elements rather than cables, shafts and gears.
`
`V. Reduction of human effort - US 714,878
`
`10.
`
`In 1902, U.S. patent 714,878 was issued for a new steering
`
`mechanism. The claim was not for steering, which was well known, and not for the
`
`gears, shafts and motive power that were used. The essential claim of this patent is
`
`as follows:
`
`This invention relates to improvements in motor—vehicles or
`
`automobiles, and more particularly to the class of such vehicles
`
`wherein the front Wheels are both the driven and the steering wheels;
`
`and the invent—ion more especially pertains to the mechanisms and
`
`5
`
`Page 6 of 124
`
`
`
`controlling appliances whereby the motor may be made available for
`
`the propulsion of the vehicle through the front steering wheels,
`
`whereby the motor may be employed to swing the steering-Wheels to
`
`steer, whereby the motor may be simultaneously caused to both drive
`
`and steer, and whereby the motor may only drive the steering-wheels,
`
`the steering being operated manually.
`
`The improved mechanism is especially useful" on large and heavy
`
`motor wagons or trucks in which, especially at the time of starting the
`
`same, considerable power is necessary to change the relative position
`
`of the wheels under the body.
`
`11.
`
`By this time the steering of vehicles was well known. The specific
`
`technical problem that was addressed was that with very heavy vehicles the power
`
`required to turn the steering wheel-s was beyond the capacity of normal human
`
`beings. They were simply not strong enough. In this patent, a mechanism is
`
`described for using power from the motor to perform the task that a human could
`
`not do. As explained in this report, the ‘055 patent, rather than using a motor to
`
`reduce human effort, describes a way of constructing a GUI to increase speed,
`
`efficiency and usability. Bits and pixels comprising GUI components have
`
`replaced gears, motors and shafts but again human capacity to control and interface
`
`with a machine has still been enhanced in a novel way.
`
`Page 7 of 124
`
`
`
`VI. Graphical User Interface Technology
`
`12.
`
`One of the questions at issue in the Petitioner’s arguments is whether
`
`or not GUIs constitute a technology. User interface technology is the subject of
`
`study at institutions such as MIT-Media Lab, CMU—HCI Institute in their School of
`
`Computer Science, Stanford, UC-Berkeley, University of Washington and Georgia
`
`Tech. All of these highly technical institutions have strong research faculty and
`
`educational programs in human-computer interaction. The goals of HCI research
`
`are to invent new ways for people to interact with computers. This is not a new
`
`field of study, but rather a subset of man-machine interface design. Indeed, HCI
`
`has adopted a number of terms from its mechanical parent. For example, buttons,
`
`sliders, exist in both fields for study, and just as changes to these features may
`
`provide an improvement in a mechanical device, improving the equivalent features
`
`in a GUI allows a computer to function better or even in ways that were not
`
`previously conceived. For example, the various GUIS on the iPhone transform it
`
`into a phone, compass, calculator, and so on. Without these GUIs, the iPhone is a
`
`useless handheld computer. Indeed, the iPhone itself uses HCI design, for
`
`example, by using slide to unlock to access the phone to replace a mechanical lock.
`
`13.
`
`Two common measures of success in HCI research are speed and
`
`accuracy. Learnability is also a common success metric. The ease with which
`
`someone can master a user interface is very important. Although esthetics do come
`
`7
`
`Page 8 of 124
`
`
`
`into play when marketing some types ofproducts (e.g., retail products), HCI
`
`researchers generally ignore this aspect and focus on making a user interface more
`
`effective rather than just prettier.
`
`14. Merriam-Webster dictionary defines technology as “the practical
`
`application of knowledge, especially in a particular area.” At issue then is the
`
`knowledge, or scientific principles, that cause the i.nvention claimed in the ‘055
`
`patent to work better than previous solutions. In my opinion, at least two basic GUI
`
`principles cause users to perceive the claimed GUI as an improvement over prior
`
`GU15, namely: human Visual search, and optimizing human effort. This section of
`
`the report includes a light introduction to these principles so that they can be
`
`understood with respect to the claims of the ‘055 patent.
`
`VII. GUI Architecture
`
`15.
`
`In the case ofthe ‘055 patent, the claims identify the specific
`
`architecture/make-up, functionality, and structural components, including a
`
`.dynamic display aligned with a static price axis providing information about the
`
`particular commodity market in which trades are occurring. The ‘055 patent
`
`particularly describes a mechanism whereby the user can more easily control the
`
`display as it adapts to changes in the market. The design ofthis GUI, including its
`
`claimed adjusting and repositioning features are critical to the user’s ability to
`
`correctly perceive the state ofthe interaction information. There are many
`
`8
`
`Page 9 of 124
`
`
`
`possibilities for the design of the GUI architecture and they will vary widely in
`
`how rapidly and accurately the user can perceive information the user is trying to
`
`perceive. The claims of the ‘055 patent describe how to construct a GUI with a
`
`very specific and concrete arrangement and repositioning of the presentation of the
`
`market information aligned with a static price axis so as to facilitate the user’s
`
`(trader) perception of the market. With this static price axis comes a need to adapt
`
`to the changes of market pricing in a way that is effective for the user with very
`
`little effort.
`
`16.
`
`The ‘O55 patent claims provide a GUI that a user can see, feel and
`
`interact with no differently than a mechanical device. The ‘055 claims provide
`
`technical solutions to the technical problems of user’s perception of market data.
`
`For example, the ‘055 patent claims provide an innovative mechanism for adapting
`
`that display to market changes by providing the claimed adjusting and
`
`repositioning of the static price axis.
`
`17.
`
`Innovation in human control of processes has a long patent history.
`
`US Patent 3,018,661 issued in 1957 is for an aviation display. The goal of this
`
`display is as follows:
`
`It is an object of the present invention to provide an
`
`aircraft instrument constructed to facilitate the control of
`
`an aircraft simultaneously in pitch and roll by a human
`
`9
`
`Page 10 of 124
`
`
`
`pilot and which preferably is combined with means to
`
`display the pitch and roll attitude of the aircraft to give a
`
`readily appreciated indication of the actual attitude of the
`
`aircraft and the action which is required to attain the
`
`desired flight path.
`
`And the patent claimed providing that particular way of presenting information as
`
`follows:
`
`1. An aircraft instrument comprising means to define a
`
`viewing aperture, a first index supported for movement
`
`within the aperture, means within the aperture to define a
`
`datum position for the first index, driving means
`
`connected to the first index, means to control the driving
`
`means in accordance with the component of the normal
`
`plane absolute acceleration of the aircraft in direction of
`
`its Z axis so that the displacement of the first index from
`
`the said datum position is proportional to the s-aid
`
`component of the normal plane absolute acceleration of
`
`the aircraft, a further index supported for movement
`
`within the aperture and means to displace the fiirther
`
`index in relation to the said datum position in accordance
`
`10
`
`Page 11 of 124
`
`
`
`with a demanded value, both as to magnitude and
`
`direction of the normal plane absolute acceleration, said
`
`demanded value being that required if some desired flight
`
`path is to be achieved and a maneuver of the aircraft in
`
`bank and pitch to superpose the first index and further
`
`index thus resulting in the attainment of the demanded
`
`normal plane acceleration and the desired flight path,
`
`neglecting any components of the normal plane absolute
`
`acceleration in the direction of the transverse axis of the
`
`aircraft.
`
`is SJCMSWu‘we-weLunutfluN3.
`
`18.
`
`The innovation in this aviation patent relies upon the pilot’s ability to
`
`perceive his current flight state in a way that will more easily allow him to control
`
`11
`
`Page 12 of 124
`
`
`
`the plane. The ‘055 patent innovates in a similar way using the claimed structure,
`
`makeup and functionality of a GUI rather than gears and dials.
`
`VIII. Human factors
`
`19.
`
`This is not the place for a complete discussion of the human factors
`
`principles that impact the design of interfaces between man and machine.
`
`However, there are three that are instructive in this case. They are: short term
`
`memory, foveated perception and expressive leverage.
`
`A. Seven +/- two
`
`20. A great deal of what we do when we work depends upon our short-
`
`term memory. Short term memory consists of the knowledge we need right now
`
`for the task at hand. It has been shown that the number of concepts that can be held
`
`in short term memory is between 5 and 9, which is described as the “seven plus or
`
`minus two” rule. When the amount of information required for a task exceeds these
`
`limits we forget something to. make room for a new piece of information. This is
`
`why talking with someone will cause us to forget a phone number that We just
`
`looked up. The new information fiom talking pushes out the phone number we just
`
`saw. In commodity trading, driving a car, or piloting aircraft there are many pieces
`
`of information that must be considered rapidly and simultaneously to perform
`
`successfillly.
`
`Page 13 of 124
`
`
`
`B. Foveated perception
`
`21. When information is displayed on the screen, the speed and accuracy
`
`with which a user can interact is heavily influenced by their ability to find desired
`
`information on the screen. The visual Search for information is largely controlled
`
`by the anatomy of the eye and specifically the retina. Figure 3 shows the anatomy
`
`of the eye. Most of the retina is the periphery with a small spot near the center
`
`called the macula or the fovea. The periphery has a lot of sensors but they are quite
`
`spread out and can only sense gray, not color. This means that most of the image
`
`that we see at any one time is gray and quite blurry. The sensors at the fovea are
`
`densely packed so that we see in high resolution and they also can sense color.
`
`ii
`hflaculal
`Fovea
`High Resolution
`Color
`Slaw
`Sazcade Bfsec
`
`Figure 3 — Eye anatomy
`
`22. At first most people do not believe that their eye works this way
`
`because they think they see everything in high resolution and in color. In actuality
`
`that is your visual memory that is supplying the information as well as the fact that
`
`13
`
`Page 14 of 124
`
`
`
`your eye can move very rapidly. As soon as you think about wanting to see
`
`something, your eye moves to look at it and it appears in high resolution. This
`
`effect can be understood by a simple experiment. Pick a line of text in the middle
`
`of this paragraph. Hold your eyes still and without moving them, attempt to read
`
`the lines above and below. With double-spaced text you will not be able to see
`
`anything but a blur outside of the line you are looking at.
`
`23.
`
`Because only the fovea can pick up high resolution information, it
`
`relies upon the eye’s ability to move very rapidly (5 times per second) and on the
`
`periphery to identify important locations to look. However, the periphery is limited
`
`in its ability t_o identify where to look because of its low resolution (blurry images).
`
`Good interface design will organize information so that it is easy for the periphery
`
`to identify Where the eye should look for the desired information. For example, this
`
`is why Warning lights in a car are displayed around a car’s more frequently viewed
`
`speedometer. The claims of the ‘O55 patent describe a particular way of
`
`constructing a GUI with a specific set of Visual relationships to simplify visual
`
`search for the information.
`
`C. Expressive leverage
`
`24.
`
`The process of visual search is only part of the technical problem of
`
`creating an efficient interactive solution. We also need to minimize the human
`
`effort to interact with the GUI. One principle of such interactivity is called
`
`l4
`
`Page 15 of 124
`
`
`
`“expressive leverage” [OLSE 07]. Expressive leverage is the ratio between the
`
`amount of information to be expressed and the amount of human effort required in
`
`such expression. High expressive leverage creates very efficient user interfaces.
`
`Natural languages such as English are attractive due to their high expressive
`
`leverage.
`
`25.
`
`A very common way to measure human effort in an interaction is the
`
`keystroke-level model (KLM) [CARD 08]. This simply counts the number of key
`
`or button entries required to accomplish a task. This measure has many limitations
`
`but it will serve here as a simple measure of expressive leverage in this discussion.
`
`L__Jse Previous
`
`if
`George Nlendenhall
`
`Name:
`
`Street Address:
`
`1450 Sonoma Blvd
`
`I Fig creek
`
`City:
`State:
`
`Zip:
`
`Figure 4 — Form Filling
`
`26.
`
`In one application a user may be required to enter a shipping address
`
`using the form shown in figure 4. To enter the address shown in the figure requires
`
`49 key presses plus 5 mouse clicks for a KLM measure of 54. If the user enters this
`
`address many times, the designers can introduce the “Use Previous” button that
`
`15
`
`Page 16 of 124
`
`
`
`requi_res only 1 click (expressive leverage of 54/ 1) to accomplish the task. This is
`
`the technique used by Amazon in US Patent 5,960,411. These examples are
`
`modern instances of the same kind of innovation found in the power steering
`
`patent. Again bits and pixels have replaced shafts, wheels and gears but the
`
`concept of magnifying the power of humans to effect desired actions is the same.
`
`IX.
`
`The ‘055 patent analysis
`
`27.
`
`The preceding discussion has laid out some of the knowledge and
`
`principles found in the field of interface design. We have shown that improvements
`
`to interfaces have long been the subject of patentable technologies and provide
`
`specific benefits. In particular, the ‘055 patent addresses the technical problem of
`
`GU15 with a static price axis, such as the one in the ‘ 132 patent, Where the display
`
`of information can be adjusted and repositioned so that it is responsive to market
`
`changes while retaining the benefits of displaying bid/ask indicators relative to a
`
`static price axis. It is not the nature of commodity trading that is claimed in the
`
`‘055 patent but rather the technology of a new mechanism that improves usability,
`
`speed and efficiency.
`
`Page 17 of 124
`
`
`
`SVDUM FGBL DEGQ9
`
`actwas
`
`III
`
`ii
`
`Figure 5 ~— [Figure 3] from the ‘055 patent
`
`The claim 1] recites the following “adjusting” feature:
`
`in response to an input command received via an input
`
`device associated with the computing device, adjusting
`
`the first plurality price levels among a range of price
`
`levels to an adjusted plurality of price levels including
`
`the first plurality of price levels;
`
`' Claim 17 includes a similar feature. Ex. 1001, 36:14-25-
`
`17
`
`Page 18 of 124
`
`
`
`displaying a bid and ask display region on the graphical
`
`user interface, the bid and ask display region comprising
`
`a plurality of locations corresponding to the first plurality
`
`of price levels displayed along the static price axis,
`
`wherein each location corresponds to one of the first
`
`plurality of price levels, and wherein a number of the
`
`plurality of locations changes according to adjusting the
`
`first plurality of price levels;
`
`29.
`
`This adjusting allows the user to extend the visible range of the
`
`market the price axis can remain static (simplifying perception and interaction)
`
`relative to a wider range of market prices. This size is always a tradeoff with other
`
`information on the screen. This element empowers the user to make that tradeoff
`
`decision.
`
`30.
`
`The claim 12 recites the following “repositioning” feature:
`
`receiving the reposition command to reposition the static
`
`price axis when a designated price is within a designated
`
`number of price levels from the lowest value or the
`
`highest value along the static price axis; and
`
`responsive to receiving the reposition command,
`
`automatically repositioning the static price axis on the
`
`2 Claim 17 includes a similar feature. Ex. 1001, 36: 45-52.
`
`18
`
`Page 19 of 124
`
`
`
`graphical user interface such that a current inside market
`
`price is displayed at a new desired location.
`
`31.
`
`This repositioning provides the ability to keep relevant information on
`
`the interface. Unlike a speedometer example, market prices have no natural limit.
`
`In a rising market, for example, the price could rise beyond the limited range ofthe
`
`static price axis. When this happens, the static price axis needs to be reset so that
`
`the current market status is visible. These two elements engage three principles of
`
`GUI design. The first is simplification of visual search. These elements do not
`
`continuously move the price axis because that would make visual search
`
`confusing. Holding the price axis static until “a designated price is within a
`
`designated number of price levels” maintains the static presentation that is
`
`beneficial for visual search.
`
`32.
`
`Because the repositioning is performed automatically, the user does
`
`not need to think about maintaining the price axis in an appropriate range. This
`
`engages the “seven plus or minus two” rule. Because the user does not need to
`
`consider repositioning of the price axis, short term memory is conserved for tasks
`
`related to trading commodities rather than managing the user interface.
`
`33.
`
`Because the user need only rarely issue the reposition command that
`
`sets the parameters for repositioning, these elements provide expressive leverage.
`
`Page 20 of 124
`
`
`
`The parameters ofthe reposition command are reused continually in repositioning
`
`Without additional effort on the part of the user.
`
`34.
`
`The whole of the language of claim 1 is not about executing
`
`commodity trades. The claim provides the structure-, make-up, and functionality to
`
`address technical problems with the usability, speed and efiiciency of an interface
`
`with a static price axis. The inventors have applied GUI design knowledge to the
`
`particular area of commodities trading to achieve a practical solution of providing
`
`a more efficient interface for doing commodity trading for some users. Thus, they
`
`have applied technical knowledge to a technical problem.
`
`Page 21 of 124
`
`
`
`X. References
`
`[CARD 80] Card, Stuart K; Moran, Thomas P; Allen, Newell (1980). "The
`
`keystroke-level model for user performance time with interactive systems".
`
`Communications of the ACM 23 (7): 396-410.
`
`[OLSE 07] Olsen, D. R. "Evaluating User Interface Systems Research,"
`
`UIST 2007, ACM (2007)
`
`I declare that all statements made herein of my knowledge are true, and that
`
`all statements made on information and belief are believed to be true, and that
`
`these statements were made with the knowledge that willful false statements and
`
`the like so made are punishable by fine or imprisonment, or both, under Section
`
`1001 of Title 18 of the United States Code.
`
`Date: 72/ B Dan R. Olsen, J
`
`Page 22 of 124
`
`
`
`APPENDIX A
`[CARD 80]
`
`Page 23 of 124
`
`
`
`J I) Fole}
`Editor
`
`Graphics and
`Image Pl'OC£.‘I-Slilg
`The Ke stroke-Level
`Model or User
`Performance _T1me
`with Interacttve
`Systems
`
`Stuart K. Card and Thomas P. Moran
`Xerox Palo Alto Research C enter
`
`Allen Newcll
`
`Carnegie-Mellon University
`
`There are several aspects 0! user-computer
`performance that system designers should
`syslemati-ca.ll3' consider. This article proposu a simple
`model. the I{eystroke~Level Model. for predicting one
`aspect of performance: the time it takes an expert user
`to perform a given task on a given computer system.
`The mode! is based on counting keystrokes and other
`low-level operations.
`the user's menial
`preparations and the system's responses. Perfornxance
`iscodedintenrtsoftheseosterativuns undoperator
`fines summed to give predictions. Heuristic rules are
`given for predicting where mental preparations occur.
`When tested against an an to difierent systems. um
`model‘; preuiictim error is 21 percent for -mdivwal
`tasks. An examle is given to ilinstntehstflv the model
`canbeusedteprofiacepanmetricpredictiosnsandhow
`sensitititj‘ analysis can be used to redeem conclusions
`in the face at uncertain usumptions. Finally. the model
`is compared to several simpler versiaus. The potential
`role {or the !l'.eyst;rok¢-Level Model in system design is
`discussed.
`
`Key Words and Phrases: lamina-cornputer in-terlace.
`human-cuter interaction. user model. user
`performance. contrive psychology. ergonomics, human
`factors. systems design
`CR Categories: 3.36. 4.& 8.1
`
`
`grants-cl prorated Lhatlhe copmaxznmmadaordssmbuwd for (lift-‘.1
`cummcrual advantage. the ACNE oopyngtzi some and the mic or the
`pubhcauonmdmxtau:appea:.andncIuuc isywtn that rat:-pywnguhy
`permumun of the hnocuatmn {or Cmnputmg Mm:lx1tIcI'y To copy
`ottmwssc. ur to icpubhill. mquiflts a In and /or spa.-at's: pcnnnasm
`&I.t§hm'I' ptucnl aidmacx SK. Card and TJ’. Mann. Xmn
`Colqaotalxnt. Palo Alto Rarearch I_'cnl.tt'.J3J3Ca1H3l¢ Hill Road. Palo
`Alto. (in 943414‘. A Naval]. D1:
`of Computer Science. Car-
`m:=gio~Mellm: Unrva-rut}. Piruburgb. PA 15213.
`@ l¢H3 RCA-I [KIM -fi782fBi}/U7'lfl-0396 51]}, 75
`
`356
`
`I. Introduction
`
`'|"hc design and cvalualmn ul‘ lt1l8l“it(.‘ll\rt: uvmptltcr
`.-qrstcms should take Into .-tccnunt the tutu! pcrllnm-.mcc
`nfthc mrtihincd user-contpulcr .~'_v:ucin_ Such an account
`would rcllccl
`the pS_\r‘l.IlIulL'>g|Céll ch;iractcri.stu;s of u:-err.
`and llltttf mtcractiun with the task and the computer.
`This rarely occurs in am; systematic and ¢.IpllL'll way.
`The causes nI' Ihts failure unay lie prmly lI'| attitude.-s
`toward the possibility uf dcalirtg succa;-a:2t'ttll3.' with pray-
`chological factors. such as lhr: belief that intuttiun. sub-
`jcctivc experience. and unccdotc furrn tht: only txmiblc
`bases for dealing with them Whatever run); be true ul‘
`these: more global issues. om: flllljfli cause is the absence
`of good analysis tools for as.-ecssing mmhmcd u.-scr-cum-
`putcr performance.
`There exists quite a bit of research relevant to the
`area of u:.er—cou1pulcI' performance. but most of it
`15
`prtliminary in nature. Pcw et al. [14], in a review oI”40
`potentially relevant human-system performance models.
`conclude “that integrative models of human performance
`compatible with the requirements for representing com-
`mand and control system performance do not exist at the
`present time." Ramsey and Atwood [IS]. after reviewing
`the human factors literature pertinent to computer sys-
`lcms.I:oncludctl1al while there exists enough material to
`develop a qualitative “human factors design guide."
`there is insufficient material for a “quantitative reference
`handbook."
`
`This paper presents one spccific quantitative analysis
`tool: a simple model for the time it
`takes a user to
`perform a task with a given method on an interactive
`computer system. This model appears to us to be simple
`enough, accurate enough. and Flexible enough to be
`applied in practical design and evaluation situations.
`The model addresses only a single aspect or perform-
`ance. To put this aspect into pcrspective. note that there
`are many different dimensions to the pcrformancc of a
`user-computer system:
`
`— Tltfle. How long does it take a user to accomplish a
`given set oftasks using the system‘?
`-Emm. How many errors does a user make and how
`serious are they?
`—Learning. How long does It take a novice user to
`learn how to use the systern In do a given set of
`tasks?
`
`—Fum-rfonatiry. What range of tasks can a user do in
`practice with the system?
`——Reca!l. How easy is it for a uscr to recall how to use
`the system on a task that he has not done for some
`time?
`
`The author: of Ihn repatl an luau: :11 alphabetical order. A.
`Hamel! in a communal to Xerox PARC, Thu paper at in revised vmum
`at 131. For 0 View of the [sugar muurclt prugrum of which the study
`ducrlbcd in lhii paper 1; a pm. in [3].
`
`July I930
`Volume 23
`Number 1'
`
`Page 24 of 124
`
`
`
`rult'(‘Ila‘-'t.lIlm*l. How many things does a user have
`t
`to keep in mind while using the system‘?
`F"nti'gtre. How tired do users get when they use the
`~.v-‘tom for extended periods’?
`
`-"it-n‘-ptztiiiltti‘. How do ll.‘i§3l'.‘\'
`the s_y\lt'trI"
`
`iaubjcctivcly evaluate
`
`Next. note than there is no .u'ngl¢- kind ofu.ter. Uvert.
`varv along many dimension.-3:
`
`Flicir ('1 rm: ofknmw'edgc of the ditlercnt tasks.
`Tlteir knowledge cy'orher .ty.tteni.s'. which may have
`positive or negative effects. on the performance in
`the S)‘.‘ilCI‘tl of interest.
`Their motor .s‘lrt'!l‘.-3 on vtmott.-t
`typing speed}.
`Their general techr:£r:alabi1ir_y in using systems {e.g.,
`programmers vs. nonprogramrners).
`Their €,\'pe’rft?nC€ with the system. i.c,, whether they
`are rruw't.‘e tuerx, who know little about the system;
`r-a.s-not u.tt3'r.r. who know a moderate amount about
`
`input devices (e.g..
`
`the system and use it at irregular intervals: or expert
`tt.rer.t', who know the system intimately and use it
`frequently.
`
`Finally, note that there is no single kind oftark. This
`is especially true in interactive systems, which are ex-
`pressly built around a command language to permit a
`wide diversity of tasks to be accomplished. The number
`of qualitatively different tasks perforruablc by a modern
`text editor, for instance. runs to the hundreds.
`
`All aspects of performance, all types of users, and all
`kinds of tasks are important. However, no uniform ap-
`proach to modeling the entire range of factors in a simple
`way appears possible at this time. Thus, of necessity. the
`model to be presented is specific to one aspect of the
`total uscr—cotnputer system: How long it taker expert
`users to perform routine tasks.
`
`The model we present here is simple, yet effective.
`The central idea behind the model is that the time for an
`
`expert to do a task on an interactive system is determined
`by the time it takes to do the keystrokes. Therefore, just
`write down the method for the task. count the number
`
`of keystrokes required, and multiply by the time per
`keystroke to get the total time. This idea is a little too
`simplistic. Operations other than keystrokes must be
`added to the model. Since these other operations are at
`about the same level (time grain) as keystrokes. we duh
`it the "Keystroke-Level Model." (The only other sintilar
`proposal we know of is that of Embley et al. [9], which
`we discuss in Section 6.l.)
`
`The structure of this paper is as follows: Section 2
`forrnulatcs the time prediction problem more precisely.
`Section 3 lays out the Keystroke-Level Model. Section 4
`provides some empirical validation for the model. Sec-
`tion S illustrates how the model can be applied in prac-
`lice. And Section 6 analyzes some simpler versions of the
`model.
`
`197
`
`2. The Time Prediction Problem
`
`The prediction problem that we will address is as
`follows:
`
`Given: A tzuzit [possibly involving several sttbtasks):
`the co