Notes on Cooper, chapters 21 & 22

Chapter 21 covered the different controls on a user interface. While sometimes it is helpful to bring up a dialog box to direct the user’s attention at some action, most controls should be somewhere on the interface itself to save time. Cooper outlines four types of controls for UI design.

The first main control type is the imperative control. When the user clicks this control, some immediate action is expected to happen. The most common example is the button. This is usually part of a dialog box and contains text to explain its purpose. It usually appears raised when it can be clicked, and depressed when the action is executed. Similar to the button is the butcon, which appears on a toolbar. It can have an image instead of text inside, and can act as a shortcut for some action in a menu. One last imperative control is the hyperlink, which brings users from one page to another.

The second type of control is the selection control. This allows the user to pick from a set of options on the application. The most basic forms are the check box, radio button, and list control. There are some ‘hybrid’ controls as well, such as the combutcon (a butcon with a dropdown selection list) and the combo box (a drop down list, with an edit box at the top for the user to simply type in the option they want). One additional control is the tree, which should only be used when there is a clear hierarchy.

Third is the entry control. This allows a user to enter custom data. Entry controls can be characterized depending on whether or not users can enter any values: bounded or unbounded. Bounded controls include spinners, dials, sliders, and thumbwheels. Unbounded entries are usually text fields. While these allow users more freedom, they need some kind of validation to prevent errors. Active validation is the rejection of text while the user is entering. Passive validation is changing the values of a text box after a user finishes inputting data. When using text fields, it’s important to be able to recognize user-entered units.

The last type is the display control. These give the user options for displaying what’s on the page. They include text controls, scrollbars, splitters, drawers, and levers.

Chapter 22 covers menus. It begins with a brief history of interfaces, beginning with command lines. Today, menus should organize commands in a way that helps advanced users, while giving clues to new users about what the application can do.

It is not necessarily wise to follow the standard menus used by other applications. A menu should be goal-directed, or designed to reflect the way the user will think. Cooper had comments about many of the most common menu items.

• File: this should contain commands related to the file itself, and should be tailored to the application (e.g. a music app should have “Song” instead of “File”)
• Edit: A helpful category, but to avoid making it a catch-all for every command that makes some kind of change
• Windows: only include if the application allows multiple documents at the same time
• Help: A prime example of how the standard format is not the best. Most help menus are not at all helpful

Some optional menus include View, insert, Settings, Format, and Tools. To further aid users, disable menu items that cannot be used, and don’t put imperative controls directly in the menu bar. Menu shortcuts should be encouraged as long as they follow standards.

AA4: Influence of Computer Anxiety on Ease of Use Perception, D. Fakun

Summary

The purpose of this study was to examine the relationship between computer anxiety and perceived ease-of-use for hypermedia enterprise software.

Studies have shown that a user’ preconceptions about the difficulty of software affects how they adapt to using it. Two factors in particular are computer anxiety (fear of having to use a computer) and computer self-efficacy (how good a user thinks they are at using computers).

The researchers defined the hypothesis of the study (H1) as follows: “Computer anxiety has a significant negative influence on ease of use perceptions.” (Fakun, 2009, p. 223)

To avoid confounding variables, the researchers also sought answers to two questions. The first, Q1, asks: would the hypothesis be valid if the software’s user interface was well-designed and exceeded user expectations (p. 224)? The second, Q2, asks: would the hypothesis be valid if the user group was allowed to try out the software and affect its development before they are required to use it (p. 225)?

The study looked at hypermedia applications designed for two separate assembly plants: one for Ford cars and another for Ericsson cell phones. In the case of the Ford plant’s software, users could test it and influence its development as much as they wanted before the release. Users had very limited opportunity to do this in the Ericsson plant.

During the usability studies, users were asked several questions about their perceived level of computer anxiety on a Likert scale. Next, they were asked to complete several tasks of varying difficulty. After completing the tasks, they were asked additional questions in a post-task questionnaire. The results of this study supported H1 for Ericsson, but rejected H1 for Ford.

The researchers found three important points as a result of the study. First, user participation is a necessary part of building software that is suitable to their needs. Second, if a user cannot try out an application free from the pressures of their job before they have to formally evaluate it, their anxiety is negatively impacted. Thirdly, if a software application does not fit well with a user group’s task requirements, general computer abilities, and available technology, H1 will almost always be supported.

The researchers ended with four recommendations to developers of hypermedia applications. First, testing if a user interface surpasses the user group’s expectations must be done after the group has had a chance to use it. Second, usability experts should be consulted in developing the interface. Thirdly, the user group’s abilities and requirements should always be considered in the development process. Finally, Q1 and Q2 as asked above should be considered together when analyzing usability data on ease of use.

Analysis

I was surprised at how far the researchers went to avoid confounding variables. They took part in the development of two separate software systems, and placed strict conditions on their hypothesis. I can imagine computer anxiety is a nebulous concept, so a study on its effects would need to be exhaustive.

I think the application of this study is particularly important. In places such as manufacturing plants, computer skill is not necessarily a primary concern. I could understand why many of these workers would be anxious about having to perform critical, high-risk tasks with a machine they don’t understand. The results of the study make sense: if the workers of a plant are to be comfortable with an application, they should have the chance to play with it before it becomes a necessary part of doing their jobs.

Do you think computer anxiety should be considered in more usability studies, or only when the interface is required in a job that doesn’t emphasize computer ability?

Reference

Fakun, D. (2009). How to mitigate the significant negative influence of computer anxiety on ease of use perceptions. Behaviour & Information Technology, 28(3), 223-238.

Notes on Cooper, chapter 14

Chapter fourteen gave an overview of the principles for visual interface design. Cooper made clear that interface design is an entire field of its own, and cannot be adequately explained in one chapter. It has some overlap with fine art and print design, but several characteristics set it apart. First of all, fine art has a primary goal of conveying emotion, and is created to satisfy the artist. Interface design, on the other hand, has a primary goal of conveying information, and is created to satisfy the end user. Aesthetics are important, but must come after function. As for graphic design of print media, artists from this background do not have to deal with interactivity or an understanding of software.

Cooper describes the basic visual elements that contribute to a user’s perception of an interface object. There is shape, which lets the user know what an object is supposed to be, but is not necessarily the most noticeable quality. Next is size, which helps a user determine how important an object is relative to others on the screen. Third is color; higher contrast makes objects easier to see, but high saturation should be avoided. Orientation on the screen determines when a user will see an object, and how they judge its role. Texture can be useful for showing an object’s function, but can be distracting.

The visual layout of an interface should be designed around the way a human mind thinks. Related elements should be grouped together and given similar visual properties. Hierarchies should be used to show the relative importance of different elements. As a whole, the interface should encourage structure and flow, so the user doesn’t get confused. A grid can be a useful approach to support this. It establishes different areas of the screen for different objects, helping maintain consistency. At the same time, a grid should be flexible enough so that it can be used on pages with different purposes. If used well, it can make information easy to find, it gives the site a clean and ordered look, and it makes the site easier to maintain.

Images can be helpful if they are used appropriately, otherwise they can distract or confuse the user. Especially important on the web is the fact that certain colors and images have different meanings across different cultures. When used well, images can help users understand functionality without forcing them to read very much.

Several rules apply to interface design on handhelds in particular. Onscreen controls should be large and noticeable. Visual landmarks should be employed, so users don’t get lost moving across different screens. If the handheld has a touchscreen, controls should be large. Text should be large with sans-serif fonts. Lastly, if there is any data off-screen, make this very clear.

Notes on Cooper, chapters 8 & 11

Chapter 8 was focused on outlining the key elements of good design. Overall, a website should try to minimize the amount of work given to the user. The structure of the website should be understandable from a user’s perspective. It should require as little memory recall from the user as possible. It should have a clear visual layout, so a visitor can tell immediately where they should look to find what they’re looking for. The amount of physical work (keystrokes, mouse movements, etc) should also be kept to a minimum.

There are several design values that all interface designers should keep in mind. They should always try to improve a person’s experience and never make things hard for them. At the same time, however, designers should try to avoid insulting the user’s intelligence. The design should always focus on the purpose of helping the users reach their goals. Designers must be pragmatic, building interface designs around the limitations of the real world, such as budget and team size. Lastly, design should be elegant. It should keep everything as simple as possible while maintaining maximum usability.

Overall, Cooper stresses that there is no single design pattern, and that an approach should be unique to the problem at hand.

In chapter 11, the authors describe several ways to identify and eliminate excise. This is defined as a task that doesn’t contribute directly to the goal. It isn’t immediately necessary, and should be reduced whenever possible. This can be counterintuitive to designers sometimes. For example, there is the need for software to include extra features that cater to new users. Novices will quickly “outgrow” such features, making them excise for the majority of users.

An easily identifiable form of excise is where the user is forced to stop and respond to dialog boxes. But there are different, more subtle versions of excise to identify. Pure excise is any work that is needlessly forced on the user. Visual excise is excessive visual information that can confuse the eye, and takes away from the ease of navigation. Navigation excise is where the user must frequently switch between different windows or panes.

To reduce this extra work, interfaces should try to keep the number of places to visit at a minimum. Navigation should include signposts and overviews, to give the user a better sense of where they are on a webpage. Lastly, though hierarchy is a necessary concept in computer science, it should be reduced as much as possible on websites.

AA3: Phone Design for Older Users, Renaud et al.

Summary

The purpose of this study was to design a cell phone that emphasizes functionality and ease-of-use for the elderly.

Nearly everyone sees cell phones as essential tools in their everyday lives, and senior citizens are no exception. Studies have shown some very good reasons for them to have cell phones. It helps maintain their autonomy; they can live and travel where they please, without worrying about isolation from aid. If their memory is failing, it can help keep their daily lives organized with a reminder system. Most of all, it encourages them to talk more often with loved ones, which can greatly improve quality of life for retirees.

Unfortunately, many older people struggle using their phones. This age group has a decreased ability to learn, as well as problems with vision, hearing, dexterity, and memory. Clearly, there is a demand for cell phones that address these accessibility concerns. However, the response by cell phone manufacturers has mostly been in the form of extremely oversimplified designs. These phones typically lack internal memory or extra features, and a few do not even have the standard 12-button keypad. In the latter case, users must memorize a mapping of 3-5 buttons in order to dial a number. Since many older people have memory trouble, this is a poor design choice.

Instead of reducing a cell phone to the most basic possible design, the researchers chose to approach from a “worth-centered” perspective. This is the ability of a device to fulfill a user’s needs. The researchers outlined two sets of mobile needs, loosely based off Maslow’s Hierarchy of Needs. The core set consisted of Organization, Safety/Security, Personal Information, and Relationships/Communication. The set of secondary needs included Entertainment, Photos, History, Non-personal information, and Mobile Commerce.

There were three parts to the research design. First, the researchers conducted one-on-one interviews with elderly subjects regarding cell phone use. Second, they asked these subjects to design their ideal phone. Lastly, the researchers themselves created a prototype based on their findings.

The study looked at 34 users from South Africa. The demographics were as broad as possible: male and female, ages 60-92, with a variety of education levels and ailments.

During the interview, the subject was asked their opinion on a number of scenarios. For example, they were asked to suppose that an elderly woman has previously survived a stroke, but worries it might happen again. Should this woman keep a cell phone for emergencies? While most subjects believed a cell phone would be useful for emergencies, they also felt most older people would not be able to operate the phone under duress. Overall, answers to the interviews suggest that the elderly mostly use the phone for calls and texts.

Next, the subjects were asked to design their ideal phone, using a cardboard phone cutout and buttons. Buttons that were placed most frequently included Nearest & Dearest, SMS Write, SMS Read, Police/Ambulance, Own Number, and Phone Book. Seldom used buttons included Alarm, Directions Service, and Directory.

The researchers came up with a list of requirements for an elderly person’s phone.  The buttons should not be too small, and their labels should be as large as possible. There should be shortcut buttons that activate frequently used functions, but the face of the phone should not be crowded. All buttons should be on the front, with enough space between them. The prototype phone was designed off of these requirements.

The researchers noted that more user studies need to be conducted in other countries. The desire for SMS shortcuts, in particular, may be biased in South Africa. Telephone calls are very expensive there, so texting is the primary method of long-distance communication.

Analysis

It was surprising to me the kinds of devices sold to older people. I am familiar with phones like the Jitterbug which only make and receive calls, but phones like the Easy5 and the Mobi-Click seem a little ridiculous. A user study involving elderly people only makes sense.

I don’t think they used the best study design. The interview in particular didn’t seem all that useful, asking very broad questions about cell phone usage among seniors. The answers were mostly along the lines of “That would be a great feature, but I don’t think the elderly know how to use it,” confirming what the researchers already knew.  The second part of the study was better, since it was more engaging for seniors and helped give guidelines for the design of future senior-oriented phones.

What do you think? What other kinds of usability studies should be performed with regards to seniors and cell phones? Should senior phones be as simple as possible, or should manufacturers try to include some more features that are useful to the elderly?

Reference

Renaud, K., van Biljon, J. (2010). Worth-centred mobile phone design for older users. Universal Access in the Information Society, 2010.

AA2: Emotion on the Road, Blaschke et al.

Summary

This research was focused on improving car safety with the introduction of a virtual car assistant.

A driver’s mental state is one of the most crucial factors in road safety. In traffic, people often experience negative sensations, like anger, confusion, and fatigue, that impact their ability to drive safely. In addition, distraction is becoming more serious with the introduction of more sophisticated gadgets. It’s not unusual for a driver to have a GPS navigator and iPod to pay attention to along with the critical items of the dashboard.

A virtual assistant could speak to the driver to positively impact his/her emotion, as well as  providing an interface for the aforementioned gadgets. To be effective, it must take facial expressions, vocal cues, and the outside environment into account. This will increase the driver’s comfort in the absence of a passenger, since he/she will not be alone in a stressful situation. If the driver is tired, the assistant will loudly suggest taking a break. If the driver is angry, it quietly tries calming him/her down. In longer trips, it will try making small talk. To make for a more comfortable driving experience, it might change the lighting, temperature, or music.

Accurately measuring a person’s emotions and vocal commands are considerable challenges in their own right. For the purpose of user studies, the researchers used a “Wizard of Oz” model. That is, the virtual assistant was actually one of the researchers speaking remotely. The user participated in a virtual driving simulator while the supposed virtual assistant spoke. The scenario was driving to work on a Monday morning. In the meantime, the user’s tasks included entering a destination on the GPS, scheduling an appointment on the calendar, and calling the office on their cell phone. A traffic jam and bad cell phone signal were simulated to add stressful factors. Whenever the user appeared annoyed or confused, the computer offered assistance.

After their virtual driving sessions, users scored their experience based on its usability and their experienced effort. Generally, users gave high scores, particularly to the virtual assistant’s ability to assist a confused driver. On the downside, users had varying opinions on whether a car should initiate conversation or try to guess stress levels. Most users wanted the option to turn off the assistant.

The researchers conducted a second study to test how accurately a computer can track driver distraction. Many more subjects were used this time, and all were asked to drive the same stretch of road while using head-tracking devices. Some of the subjects were asked to perform distracting tasks, such as changing the radio settings, dialing a phone number, or setting the destination on the GPS navigator. Various performance metrics were recorded and analyzed with a neural network called long short-term memory (LSTM). Overall, this system was around 90 percent effective at detecting whether or not a driver was distracted.

Analysis

I’m skeptical about the push to include smart technology in cars, such as corrective steering and automatic braking. It seems to me that even if the possibility for error is small, the result could be disastrous. Cruise control is an old and comparatively simple technology, yet it’s not without its problems. I have a similar concern for voice recognition; I’m familiar with the automated voice commands of tech support, and I wouldn’t want to deal with that sort of thing while on the road.

However, the first experiment of the article takes an interesting perspective. Instead of, “Can we get intelligent cars on the road?” the question seems to be “If we had intelligent cars on the road, how would drivers react?” I think it’s definitely worth looking at the human side of this problem.

If anything, the second experiment showed the kind of gap in technology that is still present. People typically have very little patience for machines. If a virtual assistant incorrectly assesses a driver’s emotions one out of ten times, the illusion of intelligence is shattered and the driver is once again dealing with tech support. So much for stress reduction.

What do you think? Is this kind of thing worth pursuing? Is it even possible to get right?

Reference

Blaschke, C., Färber, B., Eyben, F., Nguyen-Thien, N., Schuller, B., Poitschke, T., Wöllmer, M. (2010). Emotion on the road- necessity, acceptance, and feasibility of affective computing in the car. Advances in Human-Computer Interaction, 2010.

Article Analysis on Teichert et al.

Summary

The purpose of this research project was to develop a large, multi-touch, multi-user surface that is suitable for real world applications.

Despite commercial success in mobile devices, multi-touch in large “smart” surfaces is not fully developed. Applications in areas such as computer-aided design and animation cannot be made without decent hardware.

It is not uncommon, however, for innovative hardware to take many years to mature. The computer mouse, for example, was invented in the 1960s but was not shipping with personal computers until the 1980s. In the case of multi-touch surfaces, many years of scientific research have gone into perfecting the technology. Several methods of tracking exist, including frustrated total internal reflection (FTIR) and diffuse illumination (DI). The latter technique allowed more flexibility in terms of surface material and touch sensitivity, as well as some degree of hovering input. The technique is to shine infrared light on a diffuse surface and track it with an infrared camera. The experimental table, named mrT, was built with the intention of improving upon DI technology.

For effective multiple user collaboration, a multi-touch table would need a large surface area as well as high resolution. Typically, DI devices have poor image clarity because they use short-range projectors. mrT had a surface of 51” in diagonal and a resolution of 1920 x 1080 pixels. This was achieved by folding an HD projector into the case and reflecting it along three separate mirrors, creating the necessary virtual distance.

Another problem with the DI technique is its sensitivity to outside light sources. Direct sunlight, for example, usually takes its toll on tracking ability. To compensate, the mrT table used hundreds of high-power infrared LEDs, creating a highly sensitive and homogenous surface area that would work well in any lighting environment. Because of the number and intensity of infrared lights used, some internal reflection constraints had to be considered to prevent infrared light from shining directly into the camera.

The mrT table was designed with longevity in mind. For example, DI tends to cause overheating. To keep the machine operational even after prolonged use, the ventilation system was designed to keep it at 36 degrees Celsius.

The findings of this project were that, with some creative tweaks, existing approaches to multi-touch can make for more durable and usable devices. Building mrT was a time-consuming process, but much of that time was the overhead in designing the prototype. If mrT’s improvements were accepted into common practice, real-world applications for multi-touch surfaces will be easier to develop.

Analysis

I have not heard much about multi-user, multi-touch surface interfaces before, which is why this got my interest. While I was hoping for more elaboration on mrT’s applications, this paper was an interesting look at the hardware side.

The section on literature review described many examples of smart surfaces, but only touched lightly on the principle technical terms, FTIR and DI. I understand that these might be common knowledge to researchers in the field, but since the majority of the paper describes how mrT improves on DI, a brief summary of its function would help a great deal. The descriptions of the device’s inner workings were very detailed, but I was sometimes unclear on what was being  improved upon, and how (the ventilation system, for example).

Overall, I’m most fascinated by what the authors hint at in their Applications section. What kind of user interfaces could you have on a machine with multi-touch and hover detection for up to six users? How might practices like modeling and animation change?

Reference

Teichert, J., Herrlich, M., Walther-Franks, B., Schwarten, L., Feige, S., Krause, M., Malaka, R. “Advancing Large Interactive Surfaces for Use in the Real World.” Advances in Human-Computer Interaction, May 2010.

Article Dates

I will post article analyses on the following dates:

Friday, September 10th, 2010

Friday, October 1st, 2010

Friday, October 22nd, 2010

Friday, November 12th, 2010

That’s all for now.

-Billy