With the advent and ubiquity of graphical user interfaces and the desktop metaphor, window switching has become one of the most common operations performed on a computer. Window switching is also one of the most common tasks of any window manager, occurring several hundred times a day.
Problem Statement And Research Goal
Successful completion of the goal will provide a basic theory that can guide designers in developing new switching techniques. Successful completion of goals will result in new switching techniques that empirically and subjectively outperform currently available alternatives.
Research Approach
It will also provide insight into the limits of user knowledge of current switching techniques. Using the knowledge gathered in the previous objectives, analyze, design and evaluate new communication techniques based on design objective.
Research Contributions
Empirical evaluations showed that stack scanning was faster than other techniques when the number of windows is large and visual similarity between windows is important. Empirical evaluations showed that WindowsTagging is faster than the Expos technique and was highly preferred by the participants.
Structure of the Dissertation
Two experiments were designed and implemented to compare the performance of Push-and-Drag switching with other window switching techniques (Direct Show, Taskbar, Alt+Tab) under different scenarios. Results showed that stack scanning was faster than other techniques when the number of windows is high and the visual similarity between windows is important.
Window Switching Techniques
Temporal Approach
- Alt+Tab
- RelAltTab
This list order was used to replace the order of windows in the Alt+Tab user interface. When you return to the starting point, place the window in the Alt+Tab list.
Spatial Approach
- Taskbar/Dock
- Expos´e
- EyeExpos´e
- Taskpos´e
- SCOTZ
- FST
Visual TaskTip (Figure 2.8) was developed to improve the recognition of buttons on the taskbar, when the mouse is hovered over the taskbar, it displays a thumbnail image of the opened window1. Keith Humm presented Spatally Consistent Thumbnails Zones (SCOTZ, see Figure 2.11), which is an example of a spatial approach and uses the stable zones to display window thumbnails Humm (2007).
Hybrid Approach
Group Switching Techniques
Explicit Window Grouping Techniques
- Virtual Desktop Managers
- GroupBar
- Elastic Windows
- Scalable Fabric
- SCWM
Virtual desktop managers (VDM) have been one of the most popular solutions to explicitly define groups, but at the cost of strict separation between them, making it difficult to switch between windows from multiple groups. However, users must use the size of the focus area and periphery in order to operate (focus area) and recognize (periphery) windows.
Implicit Window Grouping Techniques
- WindowScape
- Stack leafing
- SWISH
A disadvantage these systems face is that they require windows to be in a single group, forcing users to decide ahead of time where a new window belongs. These systems do not allow windows to reside in multiple groups at the same time.
Tabs Switching Techniques
Desktop Organization and Switching Techniques
Spatial Memory And Visual Search
Psychology of Visual Search and Memory
Their evaluation shows that semantic labels provide stronger retrieval cues than spatial organization alone, but indicate that combinations of semantic and spatial organization improve performance.
Spatial Memory and User Interfaces
There is some evidence that people often prefer to use visual skills in organizing their work Lewis et al. Document thumbnails were no more effective than thumbnail icons in the pull test Maarten et al.
Visual Search
The role of simulated location and context has been investigated in the Infocockpit, a multimonitor and multimodal interface testbed Tan et al. Subjects recalled word pairs displayed at different points on a three-screen display, and in one condition a “context” was created by projecting a photograph (e.g., of a museum interior) as a panoramic background in the testing room.
Finding obscured Windows
Switching Techniques Time Model
Hick-Hyman And Fitt’s Laws
Card et al. (1983), +1 is "because there is uncertainty about whether or not to answer, as well as about which answer to give". Essentially, Fitts' Law states that objects that are smaller or farther away will take longer to acquire than those that are larger or closer.
GOMS/KLM
The target acquisition phase is well described by Fitts Fitts' (1954) law and has been extensively verified (for example, Card et al. 1991)) and is widely used in HCI research, mainly because many expert tasks require selection of objects low level of screen graphics or physical keys. The GOMS and KLM models are limited for two reasons: they are limited to expert performance of routine tasks and use an average time of 1.10 seconds to point the mouse to an object on the screen, which is crude compared to precision accessible through Fitts' law.
Log-based Empirical Methods
If there is more than one method that can be used to achieve a goal, a selection rule is invoked to determine which method to choose, depending on the context Cardet al. The original KLM had six classes of operators: K to press a key, P to point to a place on the screen with the mouse, H to move hands to initial position on the keyboard, M to mentally prepare to perform an action , and R for system response where the user waits for the system.
Tags
Compared to the content-based image search of Datta et al. 2008), tag-based search more easily copes with semantic queries. Some researchers and applications have experimented with combining structured hierarchy and “flat” markup to aid information retrieval Heymann & Garcia-Molina (2006).
Summary
- Issues Associated With Log-based Empirical Methods on Window
- Issues Associated With Switching Techniques Evaluations
- Leveraging Natural Human Capabilities
- The Combination of Implicit and Explicit Grouping Techniques Are
- Window Tagging
Combining our analysis of related work with switching techniques, we believe that such automation is critical (implicit clustering technique), since users are usually not inclined to organize their personal information before Bernstein et al. 2008a) and New switching techniques should allow users to control groups manually (explicit grouping technique), thus combining implicit. We then detail the design and implementation of the longitudinal study to collect user data.
Experiment Objectives
Quantitative Goals
Qualitative Goals
Longitudinal Study
Definition
Windows can overlap, two windows are said to overlap or intersect if the windows share at least one common pixel. If two windows intersect or overlap, the covered windows have some valuable information that is invisible to the user.
Participants And Apparatus
Two windows are adjacent if a section of each of their edges touches each other (so there are no pixels between the two windows), but the two windows do not intersect. An open window can be invisible to the user for two main reasons: (1) the user has hidden the window, for example by minimizing it, and (2) the window is closed because another window or set of windows higher in the z-order hides it.
Design
- WindowsOSLog
- Log Information
- Information Incompleteness
Window styles specify a set of window attributes, including whether the window is (1) visible, (2) popup (typically used for dialog boxes), (3) toolbar, and (4) always on top. The monitor information specifies which monitor the window is located on (a window can be located on more than one monitor, but this is rare).
Results And Analysis
Number of windows on the desktop
For single monitor user, the difference between the average number of windows opened is small with display resolution from 1024 x 768 to 1280 x 1024, but there was a difference for 1680 x 1050, for which users had more windows open at once (the average number of windows opened 10) (Table 3.4). Users also kept more windows on primary monitor than on secondary monitor for the dual monitor users (Table 3.6).
Window Event
For dual monitor users, you could coordinate windows between two displays, so the window move event should be more frequent than for single monitor users. We also observed that single-monitor users rarely moved windows and that there was no significant difference for the maximize/minimize event between single-monitor and dual-monitor users.
Window Switching Techniques
- Window Switching Techniques Used to Switch Windows . 50
- Types of Switching
- Tabbed Windows vs. Group Windows
Alt+Tab to switch between windows, they may neglect/forget the click action, and then get a wrong window. When users use one technique and get a wrong window, they usually don't change the technique, especially when using Alt+Tab and Taskbar.
Window Visibility
- Number of Visible Windows
- Visible Windows vs. Window Switching Techniques
On the one hand, users can see information from them, and on the other hand, when users want to switch to it, it is easy to use the Direct Pointing technique to complete this action. In this situation, users do not need to switch back and forth between two different views of the same window set (for example, if users use Expos´e to switch windows, users need to switch back and forth between windows and thumbnails) (a previous search showed that that was so.
Spatial Memory
Position can become an important factor to switch when windows are visible. From this study, we can extract and classify user memory factors used for window switching are: content (some semantic keywords including title, window content) > application type > position > recency order.
Windows Layout
- The Distribution of Window Size
- Windows Group
For the dual monitor system, groups are defined by each monitor (each monitor is like an independent unit), the total number of groups is the sum of the number of groups of each monitor. Table 3.11 shows the distribution of the number of clusters in the primary monitor and the secondary monitor for the dual monitor for each overlap threshold condition.
TDI and MDI Applications
- The Number of Tabs/Documents
- Switching Between Tabs/Documents
Figures 3.16 and 3.17 show that the number of groups of single monitor users and dual monitor users when the overlapping threshold is set to 25%. The number of tabs/documents of a TDI/MDI application is a basic and very important factor.
Active Window Sequences
Third, there was little difference between single-screen and dual-screen users in interaction pattern (Figure 3.21). Single screen Dual screens Average of the percentage of three types of interaction pattern with NTChild mode.
Conclusion
Push-and-Pull Switching
Group Switching
Overlap for a given window is calculated as the percentage of pixels covered by the rest of the group. In the following sections, we will refer to this frontmost window with keyboard focus as the active window.
Restacking the Focused Window
When you release the Ctrl key, the window with the highest Z order gets keyboard focus. We chose to focus the keyboard on this window because it is the last accessible window in the group and we think the user is more likely to interact with it.
Experiments
Experiment 1: Group Switching
- Apparatus
- Participants
- Experimental Design
- Procedure
- Results
In the second scenario, participants were asked to switch back and forth between the array represented by windows (W7, W8) and window W6 in Figure 4.4b. In the fourth scenario, participants were asked to switch back and forth between the arrays represented by windows (W7, W8) and windows (W5, W6) in Figure 4.4d.
Experiment 2: Restacking the Focused Window
- Apparatus and Participants
- Experimental Design
- Procedure
- Results
Longitudinal User Study
They mainly used direct pointing (47%) and the taskbar (36%), while Alt+Tab was used 2% of the time and tap-drag switching 15%. Participants reported that they mostly use the Push and Drag toggle when they want to have two or more windows merged.
Application
Conclusion
We conducted an experiment to compare the performance and error rate of the stack scanning technique with other four commonly used window switching techniques under different visual factors (e.g., the number of windows and their visual similarity). We describe an experiment to compare the performance and error rate of the stack scan technique with other four commonly used window switching techniques (Direct pointing, Taskbar, AltTab, and Expos´e) under a variety of combination conditions (the number of windows, visual similarity, and window layout).
Stack Scanning
Moving the mouse over a button brings all windows within the layer to the foreground. Moving the mouse to pass the button brings all windows within the layer closer to the foreground (b), the mouse moves away from the left side of the widget, bringing only the clicked window within the layer to the foreground (c), the mouse moves away from the right side of the widget, bringing all windows within the layer closer to the foreground (d).
Window Switching Time Model
When the mouse leaves the widget, the mouse can be moved from the right or left side of the widget. For stack scanning, first press the mouse wheel to call it and move the mouse to bring windows closer to the foreground (the distance to be traveled with the mouse depends on the layer order), then a visual search and a target selection.
Experiment
- Visual Factors For Window Switching
- Hypothesis
- Apparatus
- Participants
- Experimental Design
- Procedure
Meanwhile, the degree of overlap determines the number of layers in the stack scanning technique; TheMLs defines the number of windows in total Num (distractors) is similar to the target window.
Results
Switching Time
- Main Experiment
- Second Experiment
For the first level of matching, we observed a significant difference between the taskbar and other techniques (p<0.001). On average, the taskbar is 17% faster than stack scanning, Expos´e and Alt+Tab, and 72% faster than direct pointing (Figure 5.10).
Error Rate
- Main Experiment
- Second Experiment
For the HVS condition, Expos´e is slower than the other techniques for each Num condition. For LVS and NVS conditions, the taskbar is faster than other techniques for each Numcondition, so it is the best choice for users.
Qualitative Results
Eight participants said they used Direct Pointing less often with the window of interest obscured, but seven participants (six out of the eight participants) mentioned Direct Pointing as their preferred way with the window visible (the experiment has . proved) that it was faster than other techniques when the target window is visible). Participants said that it was inconvenient and uncomfortable to press the Shift key in this case (press the Alt and Tab keys at the same time).
Discussion
Participants who used Alt+Tab never used the Shift key to move back through the list of windows when they missed the target window, preferring to iterate through the entire list. Three participants who used the Alt+Tab technique said that changing the order of the windows in the list could confuse them, as stability is very important to them.
Conclusion
A window tagging mechanism was used to allow users to explicitly create or modify tag groups. Drag and drop and group sharing functions also allow users to explicitly define or modify groups.
Design Principles to Support Task Switching
Group Creation
Group Edition
Group Access
Group Deletion
WindowsTagging
Group Creation
- Implicit Group Definition
- Explicit Group Definition
The border color of the dragged thumbnail is then updated with the color of the group for which the cut area is most important (Figure 6.5 (c1) and (d1)). The border color of the thumbnail window being moved or copied is constantly updated with the color of the group being hovered (c1, d1; c2, d2).
Window & Group Access
- Window&Group Switching
- Finding Windows
Each time you enter or remove a new character, the list of windows that match the search is updated, and windows that do not match are dimmed, leaving their contents visible due to transparency. Each time a new tag is added or removed, the list of windows that match the search is updated, and windows that do not include tags are dimmed, leaving their contents visible due to transparency.
Group Edition
- Adding/Removing windows in groups
- Fixing Group
- Splitting Group
Window & Group Deletion
Implementation
- Group Layout
- Clipping Rectangle
- Splitting Group
Initialization: Create a rectangle container vecRectangles for each group Gi that intersects the drawn window do. Discover all the largest void rectangles when ri is added to the drawn window Bell & Feiner (2000) and then add all the largest.
Experiment
- Hypothesis
- Apparatus
- Participants
- Experimental Design
- Procedure
- Results
- Switching Time
- Error Rate
- Qualitative Results
On average, for the 12 windows, SFWindowsTagging was 25.9% faster than NSWindowsTagging and 42.1% faster than Expos, NSWindowsTagging was 12.8% faster than Expos, and for 16 windows, SFWindowsTagging was 27.5% faster than NS0.7% faster than NS0OS Expos and 6% faster. e, NSWindowsTagging was 25.5%. For HVS and LVS conditions, we found no significant difference between NSWindowsTagging and Expos (On average NSWindowsTagging was 13.2% faster than Expos) (Figure 6.10).
User Evaluation
Using Step-by-Step
After five days, participants were interviewed to gather details and comments about how they used WindowsTagging and how useful it was. Finally, participants were interviewed to gather details and comments about how they used WindowsTagging and whether they found it useful.
Results
Meanwhile, participants were instructed that they could use some new functions (manual control), including a drag-and-drop operation, a window snapshot operation, an edit group, a split group, and a delete group. After another five days, we interviewed them and collected details and comments about how they used WindowsTagging and how useful they found it.
Discussion
Overlapping Mechanism
Grouping Mechanism
Search Mechanism
Stable Spatial Layout strategy
Conclusion
The window tagging mechanism provides a more powerful and flexible grouping technique to help users explicitly define groups. A search function is also provided to help users find the desired windows by their titles or tags.
Future Work
The main purpose of this dissertation was to understand and improve window replacement techniques. The window tagging mechanism was first used in the field of task switching techniques to allow users to explicitly create or modify groups by tags.
Conclusion
Provide eleven design principles and introduce a window tagging mechanism to help designers design new window switching techniques. Researchers can use these findings as a platform for future observations of switching techniques, and designers can use these observations and design principles to design new switching techniques.
Future Work
- Window Tagging
- Theorize Window Switching
- Implicit Grouping Techniques
- A Standardized Evaluation Framework
InCHI ’04: Proceedings of the SIGCHI conference on Human Factors in Computing Systems, 175–182, ACM, New York, NY, USA. InCHI ’91: Proceedings of the SIGCHI conference on Human Factors in Computing Systems, 161–166, ACM, New York, NY, USA.
