User Research

User Research

When the Tablet PC team formed at Microsoft in the fall of 1999, Microsoft Vice President Alexandra Loeb hired user research advocate Evan Feldman as one of the team s first members. Even before any software developers or testers were hired, Feldman was on the scene preparing for the host of user studies that he and his team would conduct over the next three years. Everyone recognized early on the pivotal role user research would play in developing a successful product.

From fall 1999 to the release of the Tablet PC in fall 2002, Feldman and his team conducted more than 100 user studies with a total of more than 1000 users. The studies varied in duration, but each user received and used a prototype Tablet PC developed by Microsoft. User research discoveries contributed directly to several major design changes and many smaller refinements in what would become Windows XP Tablet PC Edition.

Let s take a look at some of the findings of the Microsoft user studies. By applying the following guidelines based on user research, you will be able to create more user-friendly applications and avoid some of the more common usability pitfalls. We ll divide our discussion into three areas: Tablet Usability, Application Usability, and Pen Usability.

Tablet Usability

The Tablet PC team conducted its most extensive user study in the spring of 2001. The team selected participants from several companies (none directly involved in information technology) to use prototype Tablet PCs for four weeks. During the study, users shared their thoughts on the experience.

The participants perceived two major benefits of tablets over laptops. The first is that tablets overcome a social barrier that laptops pose, namely, that it is not acceptable in many workplaces to bring laptops into meetings. Tablets do not suffer the same stigma even though they are every bit as powerful as laptops in their functionality. People more readily accept the use of Tablet PCs in meetings because of their pen-based interface, which many find less disruptive than traditional laptop keyboards. The second major benefit of tablets over laptops noted by users is also due to the pen: tablets are easier to use in standing positions and confined spaces.

The study also revealed that most participants wanted to have tablets with keyboards. Convertible tablets that switched easily between tablet and traditional laptop modes clearly offered a lot of value in the users eyes. Many users said that although the pen was great for the majority of scenarios while away from their desks, there were occasions where having a convertible tablet with a keyboard would have improved usability. Based on this feedback, Microsoft expects that the majority of tablets sold in the next few years will be convertible or will come with detachable keyboards.

The Tablet PC team began with a charter to bring pen-based tablet computing to the masses. At the time of the team s formation, the decision about which operating system to use was not a given. In fact, for a time the team pursued the idea of a new pen-centric operating system built from the ground up, akin to Palm Inc. s PalmOS or GO Corporation s PenPoint. However, an even earlier user study helped drive the decision to use Windows XP. Users in that study liked how easy it was to use a pen-centric prototype operating system but preferred the convenience of being able to access the same applications that ran on their desktops from their tablets. From that point on, the team knew that running Microsoft Windows instead of a custom pen-based operating system on tablets would be integral to the Tablet PC s success.

Application Usability

User research conducted by the Tablet PC team in a series of smaller studies also yielded some general application design guidelines. The most important of these guidelines are outlined below. You can consult the Planning Your Application section of the Tablet PC Platform SDK documentation for additional information.

It s Unconventional but Boy Is It Optimized

Early on, the Tablet PC team produced a lot of prototype software. The prototypes were designed specifically for pen interaction. Because pen and mice operate differently from one another, these prototypes ended up being rather unconventional Windows applications for example, toolbars were placed on the bottom instead of the top. New UI widgets were created to be pen-accessible, and some UI elements were emphasized over others to make it easier to target them with a pen.

Pen-optimized unconventionality, however, extracted its toll on usability. While users were indeed more productive with a pen in the prototype applications, frequent switching between the prototypes and traditional Windows applications left users confused. After multiple user studies returned the same results, the team resigned themselves to the fact that users are most productive when all applications on a Tablet PC are consistent. Granted, the software prototypes were designed expressly for pen users and were often more effective than their traditional Windows counterparts, but a tablet user s day-to-day experience would benefit more from overall consistency than from a few pen-optimized applications and features. The team therefore shelved radical UI designs in favor of a more traditional Windows interface.

Ink as Ink

In Chapter 1, we recounted Microsoft s deviation from previous tablet computers by treating ink as a native data type. Most of the Tablet PC s predecessors treated ink as an intermediary to text. Ink was merely an intuitive way by which text was entered. But Microsoft s Tablet PC attempts to give ink just as much status as text. Applications on the Tablet PC such as Windows Journal allow the user to leave ink as ink instead of assuming that conversion to text is what the user ultimately wants.

The Tablet PC wasn t always designed this way. However, extended usage studies showed that many users were content with leaving their ink notes and annotations in ink form. In fact, a majority of users preferred reviewing their own notes and annotations as ink instead of as text. Slowly, the team began to design the Tablet PC around the idea that ink is an important data type in its own right, its appearance having intrinsic value that the recognition process strips away upon conversion into text.

The application you plan to write might be one in which ink plays an important role, or it might be one in which ink is a convenient avenue on the way to text. Don t discount the value of keeping ink as ink users often prefer to see their own handwriting!

Visual Distinction of Inkable Regions

One usability issue that has not been fully addressed in Windows XP Tablet PC Edition is how users are to distinguish inkable regions from non-inkable regions in the UI. Inkable regions are the areas of the screen or application that accept ink as input. The problem is that there is no officially recommended UI guideline to make inkable regions visually distinct. Without visual distinctions, the user cannot know which parts of the application to write on. Figure 2-1 shows how serious such a problem can be.

figure 2-1 two types of textboxes. the one on the top is a normal richtextbox and the one on the bottom is an inkable inkedit control.

Figure 2-1. Two types of textboxes. The one on the top is a normal RichTextBox and the one on the bottom is an inkable InkEdit control.

In the extended usage study of spring 2001, it became apparent that without visual distinctions users expect to be able to write everywhere. For instance, users had difficulty transitioning between Windows Journal and Microsoft Word. After becoming familiar with the way that Windows Journal accepts digital ink anywhere on the page, users expected to be able to ink directly onto a Word document as well. Unfortunately, blank pages look like blank pages no matter what the application is.

With portrait orientation tablet and with pen in hand, the illusion of writing on paper was complete. Participants in the extended usage study found themselves trying to ink on documents in all sorts of applications, most of which were not ink-enabled. The lack of visual distinction between inkable and non-inkable regions frustrated the users expectations to be able to ink everywhere.

Windows Journal attempts to mitigate this confusion by ruling its default stationery with lines like note-taking paper has. The intent is to have Windows Journal documents look clearly inkable. If your application presents documents on screen in a paperlike manner, users will expect to be able to write where there appears to be paper.

NOTE
Your application ideally should support inking in the areas where a user would expect it. If that is not possible, provide visual cues that tell the user where digital ink is accepted and where it s not.

High Screen DPI

Experienced application developers understand the importance of writing code that scales well in high-DPI environments. As tablet display technology advances, manufacturers will continue to increase the screen DPI to boost readability. Thus, it s a good idea to be proactive in both writing code that is prepared for high DPI and testing code in high-DPI environments.

High-DPI problems typically manifest themselves in user studies in two ways. The first is that some things are simply too small on high-DPI screens. Sometimes applications adjust images and programmatically sized controls in units of pixels instead of device-independent units of measure. As the DPI increases, the UI elements that are fixed in pixel size become smaller on screen. To avoid this problem, your application should not work in pixels when dealing with UI. Instead, it s better to query the system s display metrics so that your application can size UI elements in a device-independent manner. That way, your UI remains the same physical size regardless of the tablet s DPI.

The second problem is that some controls are not big enough to allow room for the optional larger font size that users may choose when using a high-DPI screen. The larger font makes it easier to read text, especially as the DPI goes above 120 or when the screen is small. However, you will need to test your application in large-font high DPI to make sure that controls are appropriately sized. The good news is that testing for high-DPI compatibility does not require fancy hardware. Simply set your Windows display settings to use a higher DPI and you re good to go. We recommend testing your application at 120 DPI on a 10.5 to 12 LCD.

NOTE
Your application UI should work in device-independent units and be tested in high-DPI, large-font configurations.

Input Modality

Our last concern with application usability is input modality. Modality, in this context, refers to the different states, or modes, in which the application accepts input. An application in one mode might accept a completely different set of commands from those available in another mode.

Not all applications have modal input interfaces. For instance, Microsoft Calculator has no input modality. Although Calculator interprets the numbers you input differently depending on the numeric base you re in (decimal, hexadecimal, or binary), a button in Calculator always retains its functionality. Now consider an application such as Microsoft Paint, an application driven by input modalities. Scribbling around when the pencil tool is active is different from doing so when the eraser tool is active. The various tools in Paint establish input modalities. While you re in pencil mode, your interactions with the canvas are interpreted one way; while in eraser mode, your interactions are interpreted in another way. Some features, such as the line thickness settings, are available only in certain modes. Thus the input mode of the application governs users perception of what they can and can t do. The right input mode in the application must be set in order to perform a specific action.

Input modes make a small set of input possibilities mean a great deal more. By having modes, an application can make the same action (say, dragging the mouse) invoke different functionality. Input modes are therefore useful when working with a limited number of input possibilities. At the same time, input modes can reduce efficiency in some common scenarios because the user has to switch modes before performing actions.

All this talk about input modalities also applies to ink. It would be ideal if the pen, a natural input method, did not require you to explicitly change modes to know what you wanted to do. For instance, it would be useful if you could ink with the tip of the pen and invert the pen to erase without switching modes in the application. Furthermore, imagine that if after writing for a bit, you could just tap on a word and it would be selected automatically. A modeless user interface would make the process of inking on a Tablet PC even more natural, allowing you to capture and edit thoughts without any barriers.

Early in the development of the Tablet PC, application designers placed quite a bit of faith in being able to make Windows Journal a modeless application. After all, the potential efficiency gains are significant if a user does not have to explicitly change modes between inking, selecting, and erasing. However, a key discovery of the extended usage study was that, given the current technology, we won t be able to get rid of explicit mode switches entirely. Today, the back of the pen can be used as an eraser (and implementing this functionality is quite easy with the Tablet PC Platform SDK). But it s much harder to make other mode transitions accurately, such as interpreting that a tap on a word means the user wants to select that word. For that particular mode transition to be implicit, you must first decide that the tap on the word was not meant to dot an i or to make a diacritical mark. These sorts of decisions are hard to make correctly with a high degree of reliability. The imperfect algorithms used to make such implicit mode switches during the extended usage study ended up frustrating users more than helping them. The last thing users want is a mysterious mode switch occurring at an inopportune time.

This is not to say that all implicit mode switches are inherently unreliable. Indeed, some implicit mode switches, such as those triggered by the press-and-hold gesture in Windows Journal, work remarkably well. Good implicit mode switching requires that you use actions which the user can easily trigger intentionally but will rarely trigger accidentally.

Pen Usability

Although pens are natural input devices, they nevertheless bring a set of usability issues to tablet computing. We outline the conclusions of Microsoft user research on pen-computer interaction here, but we will also discuss the practical implications of these findings later in the chapter.

Targeting

Targeting of screen locations is one of the biggest challenges when interacting with a computer through a pen. Throughout a typical day of computer use, users target many areas of the screen with a mouse. Whether it s to place an insertion caret, move an object, or use the Windows desktop, it s of the utmost importance to first be able to move the mouse cursor to the relevant location on the screen. Tablet PC users must fulfill those same targeting requirements with a pen.

The Tablet PC team conducted a series of revealing user studies on targeting early in 2001. In these studies, increasingly smaller areas of the screen lit up in different locations for brief periods of time. By asking users to tap the lit areas as quickly and accurately as possible, the team determined that users can reliably target areas only 5 millimeters wide or larger.

Selection

The final area of user research on which the Tablet PC team spent quite a bit of time is ink selection. From late 2000 to the middle of 2001, the team experimented with different ways to select digital ink. As users responded to these various prototype selection models, the team adjusted the selection models accordingly to better enable users to select the desired ink.

Neither traditional text selection tools nor rectangular area selection tools perform well when selecting the user s ink. The main problem is that few users write (or draw) in neat rows and columns. Instead, a typical user s ink notes often consist of ink written in different sizes and at various angles. This free-form nature of inking makes traditional selection tools a bad fit.

The Tablet PC team finally settled on using a free-form selection tool the lasso tool as the standard way by which users select digital ink. The lasso tool enables the user to enclose all the ink he or she wants to select by simply drawing a closed loop around the ink. Windows Journal and the Tablet PC Platform SDK Ink Picture control both support a lasso tool to select ink. We will return to the topic of ink selection later in this chapter.



Building Tablet PC Applications
Building Tablet PC Applications (Pro-Developer)
ISBN: 0735617236
EAN: 2147483647
Year: 2001
Pages: 73

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net