Human Performance Modeling

From CS2610 Fall 2016
Jump to: navigation, search

Slides

Slides

Readings

Optional

Reading Critiques

Steven Faurie 16:34:40 9/5/2016

I accidentally submitted these to the 9/6 critique link before. Steve Faurie Beating Fitt’s Law This paper explored several techniques to make pointing at on screen objects more efficient. The three sections described decreasing the distance needed to travel to an object, increasing the size of the object a user wishes to click on, or increasing both. Several of the techniques described have been implemented to varying degrees in different programs and operating systems. With decreasing distance seeming to be the most rare, outside of right clicking to open a menu that will display actions associated with the object that was right clicked on. I think I would find several of the techniques annoying as a user. Especially increasing the stickiness of clickable objects, or displaying a big ugly cursor described in the section about area cursors. The section about C-D gain was interesting and the description of increasing errors with increasing mouse speed made a lot of sense for an average user, but I know many of my co-workers that play a lot of FPS type video games like a very fast moving mouse cursor. Because they’re experts with the device their error rates are probably comparable to a typical user with a more traditional cursor to mouse movement ratio. Many modern higher end computer mice even feature buttons on them that let you adjust the C-D gain on the fly. Which is something that might not have existed at the time this article was written. This allows users to adjust their mouse speed easily for the task they’re currently performing. Beyond Fitt’s Law: This paper developed a quantitative way to measure tasks that required maintaining velocity and steering a pointing device. The author’s developed a corollary to Fitt’s law that described these types of motions. They ended up showing that as the difficulty of a task increased the time it took to do the task increased linearly. The paper gives another tool to researchers interested in HCI. By using the equations developed by the paper researches can analyze actions like drawing a letter with a mouse or on a tablet. More tools mean more precise quantitative analyses regarding following paths can be made, and the effectiveness of different designs relating to these types of task can be compared more precisely.

Haoran Zhang 18:26:58 9/6/2016

Beating Fitts’s law: virtual enhancements for pointing facilitation Authors survey vary research in artificially facilitating pointing at targets in GUI. They said, the real world was restricted by Fitts’s law due to physical laws, but in virtual world, we may beat the Fitts’s law by changing the distance of operation or larger the size of target. However, there are technologies to let users easy to reach the target, but some of them are not scalable when we apply those technologies in other situation. For example, technology helps to select a single target that rise the difficulty to select a single target in multiple target. There are three main categories ways to optimize the experience of point target, primarily attempt to decrease D, primarily attempt to increase W, and decrease D and increase W at the same time. Since there are too many ways to archive our target, such as temporarily bringing potential targets to the cursor or design a larger cursor, or even resize the target. But those methods may introduce new problem. For example, if you want to use a larger cursor to point one target in multiple selections, the area cursor may overlap on other selections. Also, if you move potential targets to cursor, it may bring useless targets to cursor. Another way is increase C-D gain, such as minimize unimportant target and maximize important target. With more and more users use tablets, so it is may be more important to increase C-D gain, because, in tablets, there is no cursors anymore, and the system doesn’t know where is our finger. Thus, some method may useless due to lack of cursors information. Anyway, the best way to make a judgment to a technology, is not the measurement, such as Fitts’s law, it should be the end user themselves. Beyond Fitts’s Law: Models for Trajectory-Based HCI Tasks In this paper, authors did few experiments on trajectory-based HCI tasks, they are goal passing, then increasing constraints, then narrowing tunnel, and spiral tunnel. The results showed that, Fitts’s law is one of the robust in HCI area, and then they propose more robust regularities in trajectory-based HCI tasks, based on Fitts’s law. In this paper, authors used few experiments to demonstrate the logarithmic relationship between moving time and width of target. In addition, it proved that, the relation between them is liner, can all correlations are greater than 0.96. That is mean that, the Fitts’s law is robust in this task. Also, these experiments are useful when we design a new interface. For example, the menu design, sometime it involves steering tasks, both vertical and horizontal. Thus, if we can follow the rules proposed by authors, we may design a more user-friendly menu or applications.

Zhenjiang Fan 18:13:43 9/7/2016

‘‘Beating’’ Fitts’ law: virtual enhancements for pointing facilitation ::: The survey gives us all we need about the pointing performance topics: the definitions, basics, background, past and current researches, different techniques toward different circumstances, weaknesses and strength of every technique, comparison of different techniques, author’s view on the subject, as well as future potential research directions. It is obviously a great survey on the topic of enhancements for pointing facilitation. The word “beating” in its title, if you really think about it, is not a very proper word to use given the subject and content of the survey. Because if you can beat the Fitts’ Law, that means you have some techniques or new laws that are better or more accurate than the Fitts’s Law. Apparently, all the techniques mentioned in the survey are theoretically based on the Fitts’ Law and the author does not come up with a new law. So, I think, the word here should be replaced by “exploiting”, “perfecting” or ”utilizing”. It is a great beginning for the survey itself as well as its audiences that the survey introduces the whole basic related topics before starting going to details, such as graphic user interface, relationship between graphic interface and physical interface, origin of the Fitts’ Law, pointing performance, the reasons and theories why improving pointing performance is important. When the author goes on explaining very technique, it provides us different works by other researchers and tries to give us all the statistical and theoretical information related to that technique. The survey also provides lots of useful figures and graphics that help us to understand these techniques. The survey is well organized too. There three big direction of how to improve pointing performance: facilitating pointing by primarily reducing the variable D in the Fitts’ Law, facilitating pointing by primarily increasing the variable W and facilitating pointing by both decreasing D and increasing W. During each segment of explanation of each direction, the author goes deep into every detail of that technique. The most intriguing part of the survey is its conclusion. It not only summaries all the techniques and compare each one to another, but also give us some senses where we should go in the future.

Zhenjiang Fan 21:31:40 9/7/2016

Beyond Fitts' law: models for trajectory-based HCI tasks::: It has been verified by lots of conducted studies and researches that the Fitts’ Law can be applied to pointing tasks(one of human-computer interaction researches). The paper want to explore the possible existence of a regular quantitative model law that can be applied to trajectory-based tasks. The paper first introduces the Fitts’ Law and then state the reason why the Fitts’ Law is not an adequate model for trajectory-based tasks. Then the paper lays out a strategy to seek the possible existence of a regular quantitative model law that can be applied to trajectory-based tasks. Then it conducts some experiments to get some results. These different experiments cover an arrange of trajectory tasks in terms of goals’ number, constraints level and complexity from two goals task to N goals task, to N goals task with a narrowing tunnel, to N goals task with a spiral tunnel. And it comes up with its steering law according to these experiments. But I do have some questions about the ways the paper conducts its research. First of all, the research environment is limited to just one device, also limited by so many constraints. And I do have doubts about the way the paper uses the variable W in the Fitts’ Law. In Fitts’ Law, the variable W represents the target’s width, but here in the paper’s research, the variable W denotes the width of the task routine(users’ movement routine), even though the paper say W is the width of the goals. Apparently this is not the case the paper states. I also have some doubt about the way the paper conducts all its formulas in linear fashion. Because I don’t think users’ movements in trajectory tasks are represented in linear fashion, because users’ movements behave in two or multidimensional way, physically or virtually in the screen, in an irregular way. I think the results of these experiments are great though.

Tazin Afrin 22:26:06 9/7/2016

Critique of “Beyond Fitt’s Law: Models for Trajectory Based HCI Tasks”: In this paper, the authors ran 4 experiments to investigate possible regularities in trajectory-based tasks and presents a simple steering law for such tasks. In modern computer interfaces trajectory based tasks are very important for example, navigating through menus, 3D drawing etc. which cannot be adequately modeled using Fitt’s law. The authors experiment some goal passing and tunnel passing tasks on a computer. Then to increase the difficulty they also experiment drawing a trajectory along a curve and a spiral shape. Although goal passing task is similar to Fitt’s tapping task, tunnel passing task is not, but it can be modeled using the proposed steering task by decomposition. After experimenting slong the curve and spiral, the author came with a local law that, at any point the speed of the steering movement is proportional to the width of the path. Overall, I think this study has a greater impact on HCI to compare designs and performances of different devices, that Fitt’s law was not able to do. However, I find the experiments are dependent on a lot of things, like the starting point or direction. The authors did not present any results for what could happen if started from the opposite direction. Critique of “Beating Fitt’s law : virtual enhancements for pointing facilitation”: In this paper, the author discusses various proposed models for artificially enhancing the pointing experience in graphical user interfaces by either reducing target distance or increasing target width or both. While the methods are compared using Fitt’s law, none of the models provide a wide range of usability in practice. The onscreen graphical elements are already laid out in an optimal fashion, yet the icons and menus can be arranged in a way so that it reduces the distance. Another way to reduce the distance is to temporarily bringing the targets to the cursor, which is kind of annoying for the user and could possibly interfere with other targets. ‘Object pointing’ is an interesting solution where empty spaces are skipped while moving the cursor from one target to another. An example of increasing the width of cursor or target is the MacOSX ‘dock’, where targets expands when moving cursor over it. But the most widely accepted method is the C-D gain, where distance is reduced depending on the movement and speed of the cursor. The survey did a great job by comparing the different approach on the same platform using Fitt’s law. But I agree with the author when he asked if all further enhancement on C-D mapping is even worthwhile with the increase of tablets and smartphones. But this survey did not look at any model for touchscreen performance measurements. With the increasing use of touchscreen on tablets and laptops it could be more interesting to see this kind of survey comparisons for touchscreens.

Zhenjiang Fan 23:05:51 9/7/2016

An error model for pointing based on Fitts' law::: Just like the paper says that error prediction does play a very important role in the design work of user interface. Maybe that is why the author has been spending so much time on this topic. The way the paper derives its error prediction equation is pretty straightforward mathematically given the fact that the Fitts’s Law dose mathematically imply such an error prediction equation. The paper states that the effect of target size W on error is greater than that of target distance A, even though the Fitts’s Law indicating that W and A contribute equally to error rate. But the paper does not give us any reasons for this conclusion. But as we can imagine ourselves, the size of the target must play more important role in the occurrence of pointing error. The bigger the target size, the more likely a pointing hit. What makes the paper’s work less intriguing is that fact that its equation is built on several assumptions and conditions. One of assumptions, I think, is very discouraging, that is the constants a and b in the Fitts’ Law have to change. The author goes over lots other works, either support its own work or the works the author thinks should be improved. Given the unique characteristics of the work, I do think, their experimental environment and results are adequate to support its conclusion, but some of the conducted results are not that encouraging. I do think the paper fills the void of error prediction model in the pointing interface area.

nannan wen 23:07:08 9/7/2016

Beyond Fitts’ Law: Models for Trajectory-Based HCI Tasks by Johnny Accot and Shumin Zhai review: This paper presents the motivation, analysis, a series of four experiments, and the applications of the steering law. What Fitts’ law indicated is a tradeoff in human movement: the faster we move, the less precise our movements are. But Fitts’ law only address one type of movement, but nowadays there are different kinds of input devices that are not sufficient for today’s practical needs. In the paper, they analyzed a handwriting processes, regardless the script size, large or small. but , based on their experiment, the characters written larger script size were less precise than the characters in smaller size, so that the relative speed-accuracy ratio. Their study also shows that due to the fact that the time humans need to process the visual feedback information when moving has a lower bound. In this paper, they also did some experiment in apparatus, the first experiment is goal passing, it includes procedure and design and some results that they find. From their experiment, it shows that a steering task with constraints on both end follows the same logarithmic law as Fitts’ tapping task. This serves as a stepping stone towards formulating relationships between movement time and continuous constraint in steering tasks. Then they did experiment 2, which is an increasing constraints. Including their procedure and design, and results. After that they did experiment 3, which is narrowing tunnel. In this experiment, their goal is to test if their method could be applied to linear trajectories but with a non-constant path width. They recruited ten subjects to participated in this experiment. The design and procedure of the experiment was the same as for experiment 2. They find some useful results too from this experiment. What they find from their experiment is that Fitts’ law is one of the very few robust and quantitative laws that can be applied to human-computer interaction research and design. “Beating” Fitts’ law: virtual enhancements for pointing facilitation, by Ravin Balakrishnan review: In this paper, they surveyed some research into new techniques for artificially facilitating pointing at targets in graphical user interfaces. Gradually graphical user interfaces have superceded the command line interface. For the background part, the author states that in order to gain insight into the directions that can be taken in designing virtual techniques to facilitate pointing, it is helpful to understand the possible underlying motor control models that are likely explanations for Fitts’ law. Then the author talked about facilitating pointing by primarily reducing D, first they designed widgets that minimize D. Then they removed empty space between the cursor and targets. Fitts’ Law is a model of human movement, as the calculations would be inaccurate. This is a good survey paper, it presents all recent researches about target pointing techniques and also gives out concerns and future problems.

Keren Ye 23:37:50 9/7/2016

‘‘Beating’’ Fitts’ law: virtual enhancements for pointing facilitation This is a survey focus on new techniques for artificially facilitating pointing at targets in graphical user interfaces. Several factors in Fitts’ law are tested by the authors, namely (1) reduce the distance D, (2) increase the target’s width W, and (3) decrease D and increase W at the same time. The paper discussed three approaches that facilitate pointing by reducing D. (1) New widgets such as pie menus are not easy to apply without drastic redesign of the overall interface layout yet hot keys and dedicated scrolling wheels succeed in some situation. (2) Temporarily bringing potential targets to the cursor suffer the problem of clutter when the desktop is dense. (3) Approach that remove empty space such as object pointing has the several problems, especially it annoy the user because of visual discontinuities. The authors discuss two approaches that facilitate pointing by increasing W. (1) While the area cursors obey the Fitts’ law, it have two disadvantages: they may obscure underlying data and they may incorrectly select target from several targets closely grouped together. Some methods are proposed to tackle with area cursors. (2) Expanding targets follows the Fitts’ law where there was a single isolated target to be selected, however, expanding only one target can affect its neighbours. To both decrease D and increase W, the authors mention the approach that dynamically changing the control-display gain. They states that this works well where there are a few intervening distracter targets while it is not true for most real user interfaces. In sum, the paper reviews several researches for artificially facilitating pointing at targets in graphical user interfaces. It suggests that while the techniques developed to date are promising, many of them do not scale well to the common situation where multiple targets are located in close proximity. Beyond Fitts’ Law: Models for Trajectory-Based HCI Tasks An important idea proposed by the paper is that Fitts’ law suits only for pointing tasks. For trajectory-based interactions, they explore the possible existence of robust regularities and find the steering law exist. The authors proposed several experimental steps to derive and validate quantitative relationships between completion time and movement constraints, as concluded by the authors: “(1) The logarithmic relationship between movement time and tangential width of target in a tapping task also exists between movement time and normal width of the target in a “goal passing” task. (2) A thought experiment of placing infinite numbers of goals along a movement trajectory lead us to hypothesize that there is a simple linear relationship between movement time and the “tunnel” width in steering tasks. (3) The relationship is confirmed in three types of “tunnels’: straight, narrowing, and spiral. (4) The authors then generalize the relationships in both integral and local forms. The integral form states that the steering time is linearly related to the index of difficulty, which is defined as the integral of the inverse of the width along the path; the local form states that the speed of movement is linearly related to the normal constraint.”

Xiaozhong Zhang 1:04:30 9/8/2016

Beating Fitts' law: Virtual enhancements for pointing facilitation The paper surveyed some attempts to improve user experience of pointing and selection. It first cited the Fitt's Law which calculates the movement time based on object distance and dimensions. Then the underlying motor model was described, which include a first pulse towards the target and a second controlled stop to hit the target. The connection between the two was stated as the that the distance in the Fitt's Law is associated with the pulse phase while the object dimensions are related to the controlled hitting phase. Then the paper explained three methods to improve the pointing performance, namely reducing the distance, increasing the width and both at the same time. In the conclusion part, the paper mentioned that in many of the techniques surveyed, the techniques tend to produce significantly improved pointing performance when selecting isolated targets, difficulties arise when they are used for selecting one of multiple targets that are spatially close together. It also concluded that the choice of input device can significantly alter the effectiveness of a particular pointing enhancement technique. And at last, the paper suggested the surveyed research should adopt user acceptability as another evaluation criterion besides the quantitive metrics. Beyond Fitts' Law: Models for Trajectory-Based HCI Tasks The paper argued that Fitts’ law cannot successfully model trajectory-based interactions and proposed a new method called steering law. The paper then designed four experiments with increasing tapping constraints, so that the resulted tapping experiment become more and more similar to a trajectory-based experiment. The first one is goal passing, in which two goals were set for the user to pass. The result shows that the steering task observes the same logarithmic law as Fitts’. Then the number of goals were extended to infinity and the problem was transferred to a tunnel traveling problem. The paper then defined a third experiment by adding the width constraint to the tunnel. Then, the author used a new concept by integrating the inverse of the path width along the trajectory and defined a global law. To test their method for complex paths, they studied a fourth experiment called the spiral tunnel and designed with different path width and number of curves, and derived a local law for instantaneous speed. As the paper also mentioned, the regularities presented in this study can enrich the small repertoire of quantitative tools in HCI research and design.

Debarun Das 1:45:37 9/8/2016

“‘‘Beating’’ Fitts’ law: virtual enhancements for pointing facilitation” by Ravin Balakrishnan: This paper aims to do a survey of the different research works that have been done to beat the Fitts’ Law in creating virtual enhancements to improve pointing performance. It begins by discussing the different background concepts. These include the concepts of iterative corrections model, impulse variability model and the optimized initial impulse model. The first model makes use of feedbacks to make sub movements towards the target. The second model depends on the initial impulsive movement towards the target. The optimized initial impulse model is a hybrid of the first two models. This is followed by discussion of research work that has been done on three areas. The first research area is that of reducing ‘D’ in the Fitts’ law. Prominent works in this area include pop-up menu (both pie chart and linear menu) that appear right beside the cursor and the drag-and-pop feature (where virtual proxies of the icons are brought closer to the cursor to reduce ‘D’). The second area of work includes those that facilitate pointing by increasing ‘W’. This includes the use of area cursor and point cursor; and expansion of target icons. However, the drawback of expansion of target icon is that it does not perform well in case of multiple icons grouped together. Finally, works where virtual enhancement is facilitated by both decreasing ‘D’ and increasing ‘W’ are discussed. However, all the works discussed do not work well in all situations of a typical graphical user interface. This paper finally discusses the shortcomings of each of the methods and indicates further work that has to be done to make effective user interface to facilitate virtual enhancement of pointing facilitation. ……………………………………………………………………………. “Beyond Fitts' law: models for trajectory-based HCI tasks” by Johnny Accot and Shumin Zhai: This paper aims to measure the movement time for trajectory based interactions (like “drawing curves”). This is needed as Fitts law successfully estimates the movement time for only pointing interactions. It discusses about a “steering law” which is established by the help of four experiments to establish relation between movement time and the constraints involved in each task. The experiments increase in complexity to define the final model based on the “steering law”. The paper effectively accomplishes generalization of Fitts’ law for trajectory based movements. This paper hence serves an important one in the field of HCI as it extends the Fitts’ law to define movement times in trajectory based interactions too.

Anuradha Kulkarni 7:04:01 9/8/2016

“Beating” Fitts law: virtual enhancements for pointing facilitation : This paper elucidates techniques to improve the pointing tasks in interface design. The paper discusses about the developments to supersede the limitations of the Fitt’s Law, which states that pointing time is limited by the width of the target and the distance to the target. Even though the paper discusses many methods to improve the performance keeping in mind to beat the Fitts' Law but methods discussed in the paper all base their theory on Fitts' Law. This shows how significant and important Fitts' Law has been. This paper employs three techniques: one by decreasing the distance to the target, second by increasing the width of the target and third aims at the combination of decreasing the distance and increasing the width. The main contribution of this paper is to evaluate all these three kinds of optimizations. It was observed that on decreasing the distance to the target, the near targets are more difficult to select. On increasing the width of the target, the icons are huge and leads to bad performance. For each of these methods we see that when there are too many icons there is no much performance improvement rather in most of the cases the methods discussed depletes the performance. This paper is interesting and the methods were explained with an example which aided in better understanding. Beyond Fitt’s law: Models for Trajectory-Based HCI Tasks: This paper elucidates the mathematical model for predicting movement time in trajectory based tasks. i.e. steering. The final model that is derived is an extension of the Fitt’s Law. The paper discusses four experiments that were employed in order to derive with the steering law/equation. The four experiments are explained in detail. The experiment settings, design, tasks and results were clearly explained. The order of the experiments is properly placed as it starts with simple experiments and then goes to narrowing tunnel and spiral tunnel. The paper is interesting as the final model is a natural derivation of the Fitt’s Law where it is a linear function of the index of difficulty and the index difficulty is still defined based on the width and amplitude of the task. This paper is also significant for two reasons. One that it shows the importance of Fitt’s Law and second that it provides a strong correlation between the past technology and the recent technology.

Anuradha Kulkarni 7:04:51 9/8/2016

“Beating” Fitts law: virtual enhancements for pointing facilitation : This paper elucidates techniques to improve the pointing tasks in interface design. The paper discusses about the developments to supersede the limitations of the Fitt’s Law, which states that pointing time is limited by the width of the target and the distance to the target. Even though the paper discusses many methods to improve the performance keeping in mind to beat the Fitts' Law but methods discussed in the paper all base their theory on Fitts' Law. This shows how significant and important Fitts' Law has been. This paper employs three techniques: one by decreasing the distance to the target, second by increasing the width of the target and third aims at the combination of decreasing the distance and increasing the width. The main contribution of this paper is to evaluate all these three kinds of optimizations. It was observed that on decreasing the distance to the target, the near targets are more difficult to select. On increasing the width of the target, the icons are huge and leads to bad performance. For each of these methods we see that when there are too many icons there is no much performance improvement rather in most of the cases the methods discussed depletes the performance. This paper is interesting and the methods were explained with an example which aided in better understanding. Beyond Fitt’s law: Models for Trajectory-Based HCI Tasks: This paper elucidates the mathematical model for predicting movement time in trajectory based tasks. i.e. steering. The final model that is derived is an extension of the Fitt’s Law. The paper discusses four experiments that were employed in order to derive with the steering law/equation. The four experiments are explained in detail. The experiment settings, design, tasks and results were clearly explained. The order of the experiments is properly placed as it starts with simple experiments and then goes to narrowing tunnel and spiral tunnel. The paper is interesting as the final model is a natural derivation of the Fitt’s Law where it is a linear function of the index of difficulty and the index difficulty is still defined based on the width and amplitude of the task. This paper is also significant for two reasons. One that it shows the importance of Fitt’s Law and second that it provides a strong correlation between the past technology and the recent technology.

Zuha Agha 8:40:05 9/8/2016

Beating Fitt’s Law: The paper studies the possibility of beating the constraints of Fitt’s law for pointing devices by artificially reducing the distance and width of the target being pointed in the virtual world of graphical user interfaces. It shows how changes in the distance to the target, target width and speed of movement affects the time taken and the certainty of hitting the target. The simplest way to facilitate pointing is to reduce the distance of the cursor to the target by introducing linear or pie menus, however such menus only allow one of the many options to be selected. Another option is to temporarily reduce the distance between target and cursor by drag-and-drop but such techniques only work effectively for sparse interfaces. Alternatively, object pointing can be used that allows bypassing the blank spaces between the objects to select between cursor and object targets only such as jumping from one option to another using the tab key. Another way of facilitating pointing is by increasing the width of the target by either having large area cursors that are semi-transparent with hotspots or by dynamically changing the size of the target. Lastly, a hybrid approach is suggested that can facilitate pointing by decreasing the distance and increasing the width. In order to achieve that, one way could be to dynamically change the C-D (control-display) gain, which is the ratio of movement of cursor to movement of the controlled object. Higher C-D gain leads to difficulty in precise movement but smaller C-D gain leads to coarser movement. Target-aware gain adaptation helps improve performance for isolated objects, but again leads to problems for grouped objects. Overall, the study shows that while some techniques work better than others under certain pointing scenarios, none work well universally. Usually the challenge is dealing with multiple targets in close proximity. I think the paper provides good insight into the mechanics of pointing and its physical and spatial constraints. But as the paper is more than a decade old, a lot of advances have been made into this field with devices such as smart watches and google glass with creative interfaces that operate effectively despite the constraints of Fitt’s law. Beyond Fitt’s Law: The paper studies trajectory-based interface interactions and develops a steering law after analyzing the motion of such interactions. The paper extends the essence of Fitt’s law to trajectory-based interactions such as drawing curves. The paper designs a set of experiment that requires subjects to draw a line from one end of the tunnel to the other and records their movement time. Variations of the experiment conducted include decreasing the amplitude of the tunnel logarithmically which showed that there is a linear relationship between movement time and tunnel width in steering tasks. Other variations that makes the tunnel narrower along the path and use a spiraling tunnel. All experiments show that the steering time is linearly related to the index of difficulty, and error rate increases as difficulty increases. In my opinion the paper presents an interesting mathematical and experimental prototype for trajectory based models but the trajectory prototypes discussed are too limited in nature and may not cover the broad spectrum of trajectories.

Zuha Agha 8:40:12 9/8/2016

Beating Fitt’s Law: The paper studies the possibility of beating the constraints of Fitt’s law for pointing devices by artificially reducing the distance and width of the target being pointed in the virtual world of graphical user interfaces. It shows how changes in the distance to the target, target width and speed of movement affects the time taken and the certainty of hitting the target. The simplest way to facilitate pointing is to reduce the distance of the cursor to the target by introducing linear or pie menus, however such menus only allow one of the many options to be selected. Another option is to temporarily reduce the distance between target and cursor by drag-and-drop but such techniques only work effectively for sparse interfaces. Alternatively, object pointing can be used that allows bypassing the blank spaces between the objects to select between cursor and object targets only such as jumping from one option to another using the tab key. Another way of facilitating pointing is by increasing the width of the target by either having large area cursors that are semi-transparent with hotspots or by dynamically changing the size of the target. Lastly, a hybrid approach is suggested that can facilitate pointing by decreasing the distance and increasing the width. In order to achieve that, one way could be to dynamically change the C-D (control-display) gain, which is the ratio of movement of cursor to movement of the controlled object. Higher C-D gain leads to difficulty in precise movement but smaller C-D gain leads to coarser movement. Target-aware gain adaptation helps improve performance for isolated objects, but again leads to problems for grouped objects. Overall, the study shows that while some techniques work better than others under certain pointing scenarios, none work well universally. Usually the challenge is dealing with multiple targets in close proximity. I think the paper provides good insight into the mechanics of pointing and its physical and spatial constraints. But as the paper is more than a decade old, a lot of advances have been made into this field with devices such as smart watches and google glass with creative interfaces that operate effectively despite the constraints of Fitt’s law. Beyond Fitt’s Law: The paper studies trajectory-based interface interactions and develops a steering law after analyzing the motion of such interactions. The paper extends the essence of Fitt’s law to trajectory-based interactions such as drawing curves. The paper designs a set of experiment that requires subjects to draw a line from one end of the tunnel to the other and records their movement time. Variations of the experiment conducted include decreasing the amplitude of the tunnel logarithmically which showed that there is a linear relationship between movement time and tunnel width in steering tasks. Other variations that makes the tunnel narrower along the path and use a spiraling tunnel. All experiments show that the steering time is linearly related to the index of difficulty, and error rate increases as difficulty increases. In my opinion the paper presents an interesting mathematical and experimental prototype for trajectory based models but the trajectory prototypes discussed are too limited in nature and may not cover the broad spectrum of trajectories.