The Web can be:
The production and distribution field is leveled because of the Web's characteristic as a distributed, client/server system. The user faces a scattered, dynamic field for attention because of the wide variety of content available on many servers. The content of the Web often is highly enmeshed in porous, dynamic, interactive, and competing Web works.
The idea of developing any product in a user-centered manner-that is, one in which the needs, interests, characteristics, abilities, knowledge, skills, and whims of the user are central in the whole process-might not seem like a radical idea. After all, a web is meant for users to find information and accomplish specific objectives. Not all web developers, however, are sensitive to the needs and experiences of the user. In fact, because user experience often is difficult to plan for and analyze, it often is overlooked. The characterization of user experience in hypertext also is not as simple as it might seem. What is the user doing when experiencing a web? The web developer should have a basis for approaching this question in order to create meaning based on user experience.
Whenever a user navigates the Web with a browser, certain essential needs exist that must be met (such as the capability to view documents and active links in hypertext). A user also will want an additional set of functions in a browser, such as a way to record items in a hotlist or other access to charting features-functions that might not be essential for viewing the Web but that could be very useful for effective navigation. Finally, there is a range of deluxe functions-for example, ways for a user to change fonts and set other preferences in the browser. These functions can help make a user's journeys through the Web more enjoyable.
A Web user's needs consist of a series of activities that can be arranged from lower level survival needs to the luxuries of Web navigation. This progression, because it's arranged with the most basic needs first, reflects the basis for user experience in navigating the Web.
The subsections that follow trace through a Web user's hierarchy of needs, highlighting how these needs are met by Web browser features (using Netscape and/or Mosaic as examples, in some cases). This hierarchy of Web user's needs consists of seven levels:
These needs constitute the set of browser navigation functions that Web users experience. As such, this list of basic needs serves as a basis for development principles covered in the following chapters.
A browser's rendering of HTML or other media gives visual or aural information to the user. This information can be in the form of ASCII characters (such as with the Lynx or the CERN line-mode Web browsers) or graphics used with other browsers (Netscape, Mosaic, Cello, or Prodigy's browser, for example). In all cases, though, the browser resolves the HTML (or other information format) so that the user can experience it. In the case of sound, movies, graphics, or other sensory stimuli, the browser's connection to helper applications should call the proper multimedia player into action for appropriate sensory display. In the case of an embedded executable program (for example, written in a language such as Java), the browser works with the language interpreter to interact with the user.
Despite the potential that these multimedia communications may bring to the Web, the most popular form of communication on the Web remains visual, in the form of text and graphics. The elements of the visual displays of information follow:
The text of the resource (if any) is displayed on the user's terminal or on a graphical area of the browser. In text-based browsers, the control functions accessible by the user often are keyboard commands, so that an elaborate visual reminder of the options is not always in view of the user.
The hyperlinks (if any) that exist within the text of the resource are identified. These hyperlinks are to other information, spaces, or services, and are indicated by hotspots within the text. Browsers identify these hotspots in a variety of ways-reverse video, underlines, numbering schemes, or other markings or symbols to indicate the presence of a hypertext jump that a user can select. Representation of hotspots also may include graphical variations, such as shading, which indicate that the user previously visited a resource.
Special browser-provided symbols to special features such as unloaded images or unloadable images (in the case of nongraphical browsers) are displayed. These include, for example, the symbol that the Mosaic browser uses to show the presence of an inline image in a document that has not been reloaded. Similarly, other graphical browsers such as Netscape include a display of special symbols for unloaded inline images. For nongraphical browsers, inline images, of course, can't be displayed. A nongraphical browser can display a string of characters, however, that the information provider defines within the HTML of the resource (using the ALT field of the IMG tag; see Part III, "Web Implementations and Tools"). The nongraphical browser displays this character string instead of the image.
Graphics in the form of diagrams, pictures, icons, or Java applets. These visual elements-text, hyperlinks, special symbols, and graphics-are the most basic need of a Web user. Information displayed alone is used to convey much meaning on the Web through text, graphics, symbols, and animations.
Although the display of information is a Web user's most basic need, without link activation (the capability to activate a hyperlink so that the browser displays a resource to which the link refers), the Web would not be an associatively linked set of information at all, but just a set of disparate information pages located on servers. Associative linking is the key to the unique way that the Web helps people create meaning, and the capability to traverse these links is the next level in the hierarchy of needs for a Web user.
The fundamental idea of link activation is that a user can select one of the hyperlinks (if any) in the information display. This selection causes the browser to retrieve the resource specified by the selected link. This resource might be another document, an information service, a picture, a sound, or some other sensory stimulus. After the user selects the link, of course, the resource must be retrieved (possibly from the user's computer or from a server on the Internet located across the world). After a resource is retrieved, a user's needs shift back to information display (as discussed previously). A Web user can employ a variety of ways to activate a hyperlink, and these ways vary, of course, according to the Web browser used. Graphical browsers usually employ a mouse-based scheme of point-and-click selection. For nongraphical browsers, keyboard commands (or number selection as in line-mode browsers) frequently are used. The essence of link activation involves a transaction between the user and the Web browser; based on experiencing the information display, the user chooses a hyperlink to follow and conveys that choice to the browser. This process of viewing, choice, and link activation is the essence of Web navigation.
The first two categories of needs for a Web user-information display and link activation-could, theoretically, provide a Web user with everything needed to experience the Web. By following the links on the default home page of the browser, a Web user could follow links until reaching a dead end (until a resource is displayed in the browser that has no hyperlinks in it). Without the capability of movement, however, a Web user, when reaching such a dead end, would have to exit the browser and start all over to follow a different path! It would be possible to navigate the Web with this scheme, but it would be very unpleasant.
Movement, then, is the next level of Web users' needs. Movement is the capability to select a link from a set of previously visited resources or to move directly to a particular resource. Movement is key for a Web user to make good use of the Web, and it enables a Web user to be more flexible in following paths.
The most basic movement function is Back. This function enables users to re-select the resource that was displayed in the browser before the most recent link activation. The Back function helps users retrace their steps after reaching a dead end on the Web. Although this might seem like a very simple procedure, a browser must have a "memory" in order to support a Back function so that it can store the URL of the currently displayed resource when users choose to activate a link. The capability to repeat the Back function multiple times requires the browser to store a stack of previously visited locations. Storing these locations from a nonlinear traversal of the Web into a linear data structure (stack) requires an algorithm that involves recording only certain past paths in the Web.
Another basic movement function is to open an arbitrary URL. Although a very popular way to view the Web is to make selections only from the available set of hyperlinks in the browser's information display, users might want to "go to" a particular place on the Web. Without an Open function to enable this, Web users are doomed to wander only that portion of the Web connected to where they happen to have started. Theoretically, the entire global Web eventually can become connected through spiders and subject trees, but floating islands of hypertext might exist in the Web that are not listed in any spider database or Web tree. These pages also might not be connected via a link to any page of hypertext that is listed in the popular spiders or trees.
Although not as crucial as the Back and Open functions, a Forward function (implemented often in browsers for symmetry and completeness) often is needed by a Web user. A Forward function enables the user to revisit resources that have been backed over from operation of a Back function.
The key for the user to operate the Back and Forward functions with any particular browser is to understand the algorithm used to fill and flush the browser's memory stack that holds these locations. The user's experience of the Web could be (and most often would be) nonlinear, but most browsers use a linear stack method for storing locations in its memory.
The needs discussed so far could give a Web user just about all the functionality to encounter the Web fairly well. There is another, higher layer of needs, however, related to the capability to control information that a Web user often wants. These information-control needs arise from the imperfect nature of the Web. If network connections never failed, retrieval of data across the network were nearly instantaneous, and all Web pages were designed well, these information-control needs would never arise. But the Web isn't perfect, so the Web user must have ways to control information.
First, a Web user needs to be able to stop network information retrieval. Network information retrieval occurs when a user selects a hyperlink referring to a resource on a remote host. If that remote host is not operating, the browser often hangs and keeps trying and trying to retrieve the resource. Or, if the resource is huge, the browser keeps working away, retrieving the resource byte by byte. If left unchecked, these retrieval processes could take a very long time and waste a great deal of network bandwidth. Faced with such a situation, a user needs to be able to request that the browser stop the retrieval. In Mosaic, the famous spinning globe serves this function. In Netscape, the stop sign icon does this (alternatively, clicking Netscape's animated logo stops this, but this action sends a user to the browser manufacturer's home page). Nongraphical browsers sometimes have control sequences to enable this (the keyboard commands Ctrl+C or Ctrl+Q, for example).
Another information-control need is related to the idea of stopping network information retrieval. This is the capability to control image loading. Controlling image loading is an issue, of course, only in graphical browsers, but it is related closely to the need for stopping network information retrieval.
In most graphical browsers, you can turn off image loading. All inline images then are represented by an unloaded image symbol. By being able to control image loading, a user can avoid situations in which massive amounts of inline images are used on Web pages. Large numbers of inline images can be as potentially crippling as a massive resource retrieved from a remote site. Unfortunately, the practice of including many inline images on a page is common on the Web. Therefore, the capability to turn off these inline images for more efficient Web navigation and specialized techniques, such as surfing, is crucial.
Just as the Stop and Turn Off Images functions described previously are key to a Web user's capability to control the information, so is the capability to make use of a (possibly large) resource in a browser's information display. A Find function gives the user a way to search for character strings or patterns within the document text currently displayed in the Web browser. This Find function often works similarly to functions found in word processors to search for the occurrence of a string in a document. Without this Find function for a browser, a Web user must visually search for a string or keyword of interest in a (possibly very) long document. It is possible, but it could be extremely laborious when dealing with long documents.
After users meet the needs for information display, link activation, movement, and information control, their attention turns toward features such as interactivity.
A Web user can move very efficiently if a Web browser readily displays the following information:
A user can progress through the web more efficiently by using the following navigation aids:
Visual aids you can use follow:
In an ideal browser, users would feel that nothing intervenes between themselves and the Web. To accomplish this, users must be trained well in techniques for using a particular browser as well as in general Web-navigation tools and techniques. The browser must have an inherently good design; otherwise, even the most adept users would grow impatient with it. Similar to the emergence of standard applications in word processing software, standard Web browsers probably will emerge with interfaces that elegantly and lucidly meet users' needs. A well-designed Web browser is the essential first step in the process for a user to navigate the content of the Web.
Web users can meet their navigational needs by:
Just as Web users gain a great deal of cues from the kind of information space they're in, they also pay attention to the information's texture. Information texture, as defined here, refers to the medium in which the information is encoded, the structure of the information, and the connections to and from the information. Just as users of a Web look quickly to find cues about the information space they are in, they also look for cues about how the information is presented. By examining cues of media type, information structure, and connections, they quickly get a handle on how to extract information.
Media type is one aspect of information texture. A user entering an FTP site, for example, might encounter a long list of files that display a variety of media types-graphics, a movie, text files, and directories, for example. This variety (or uniformity, in the case of all the same kinds of media presented to the user) is the media type, which is one aspect of the information's texture. A quick look at the possible graphical symbols at an FTP site or a Gopher, for example, quickly creates a set of user expectations about what will be found there and the interface required to sense that information. Users encountering a long list of sound files, for example, knowing that their sound player is not hooked up to their browser, know immediately that the site contains information they can't use.
Interactive media such as Java applets are information textures that allow interactivity. Typically, a Java applet is animated or presents a visual cue suggesting the potential for interactivity (for example, the cells of a spreadsheet).
Another aspect of information texture is information structure. Structure is the overall organization of the information within the display of the browser. The structure could be characteristic of an information space (such as the list of files at an FTP site or a menu from a Gopher), or it might be an ordered or unordered list within an HTML file. Structure is the pattern by which the information is presented. Simple structures, like lists or menus, immediately are recognized by the user.
Other structures, such as the complex interspersing of paragraphs, ordered and unordered lists, figures, and forms using HTML might be more difficult at first for the user to perceive. In either case, the structure of the information sets up expectations in the user about how to deal with the information. If the list shown on the browser display is numbered and it continues down the page, a user quickly forms the expectation that the rest of the list will be available by using the scrollbar. In more complicated structures that are possible in Webs, the structure of the information, although more expressive, might include paragraphs and lists, and the user might not know what to expect on the rest of a page or on other pages.
Another aspect of information texture is the connections to other information that are explicit or implied. An FTP site listing, for example, often includes a folder at the top of the list with the Parent Directory label next to it. This folder icon sets up in users' minds an expectation that the information they are encountering is connected to some other information (hierarchically up, in the case of FTP sites). In the case of a Web, these connections might be to pages that are more general or more specific in information content than the page the users presently are viewing, but not necessarily in a strict, linear hierarchy (not necessarily up). In the case of experiencing any information connections, users wonder, "Where in the hierarchy (in the case of FTP sites or Gopher menus) am I?" or "Where in the mesh (in the case of Webs) am I?" The connections to this other information, revealed by cues, can have a great impact on setting up their expectations about how to deal with the information shown.
Although information texture often is the first thing users notice when entering a web, cues are the next part of users' experiences. Whereas information space and information texture have set up expectations in users about "Where am I?" and "What is this?," cues are the features in a web that say to users, "Here is what this is" (information cues) and "Here is how to get there from here" (navigation cues). Information cues are the features of the text or graphics on a web page that help users know the page's purpose, intended audience, contents, and objective. In other words, information cues help users know what the page is for and what it contains. A careful presentation of information cues can get users oriented quickly, enabling them to more efficiently use information.
An example of an information cue is the title of the document-as it appears in the Document Title window on the browser and the words that appear most prominently at the top of the page (which the user may perceive as the title). A meaningful title that conveys the purpose, audience, and objectives for a web page serves well to orient the user. A title such as "Business Divisions of XYZ Industries, Listed by Region," for example, immediately helps the user know what to expect on that page.
Other examples of information cues are icons, background textures, colors, headings, subheadings, boxed text, or any feature the web developer uses to direct the user on how to get and use information. Considerations for cues play a large role in the design process for a web.
Communicating on the Web is different from communicating through paper-based means (such as brochures, reports, letters, memos, and other documents) because it involves a different kind of encoding process (how a communicator creates hypertext) as well as a different kind of decoding process (how users perceive webs through network-distributed browsers and servers). Because the Web's characteristics and qualities shift user focus and make possible a much more dynamic environment for presenting communication, processes for Web communication differ from many forms of traditional communication in many ways.
Web communication involves different space and time constraints, taking on a different form and employing a different delivery mechanism than traditional media.
When people receive a paper memo, for example, they first might pile it with all the other things to deal with: reports, electronic mail, meetings, voice mail, postal mail, express mail, and so on. All these kinds of communication compete for attention in terms of the space and time they occupy. The memo on a desk is more likely to get attention than the one in the bottom drawer of a filing cabinet or the one that arrived last week. In addition, these forms of communication compete in terms of what form and delivery mechanism they employ. A brightly packaged express mail letter (a special form of communication) usually commands more attention than a plain envelope, particularly when a receiver must sign for the express mail (a special form of delivery).
On the Web, however, the user chooses the time and space for communication. The form of the communication's display (how the hypertext file is shown, in terms of font and appearance) is set by the user's browser, and the delivery mechanism is the same for all information along the hypertext links of the Web itself. Although access to information on the Web is constrained by awareness of it and the skills necessary to retrieve it, all information is potentially equally accessible. Is the 1948 company report in a storage room as accessible as the memo sitting on a desk today, for example? If delivered over the Web, that 1948 company report becomes not only more accessible to a single user but to any number of other users at the same time.
The form of the Web itself-hypertext-is different from the linear flow of print on paper. Whereas memos and other communications offer themselves as separate objects, branches off a hypertext document can link and thus relate one document or piece of a document to another, resulting in contextual relations among documents. Links from Web documents can be to hypermedia resources, interactive documents, or information-delivery systems.
Web communication takes place within a context much larger than a single site or organization, involving social and cultural structures shaped by traditions, shared meanings, language, and practices developed over time. The Web, like many other forums for computer-mediated communication on networks, rapidly has created specialized information and communication spaces. On computer networks, social and information spaces exist that are, by tradition, set aside for particular purposes. Behavior in these spaces is governed by collective agreement and interaction, as opposed to a single organization's rules of operation.
Community norms developed on networks inhibit advertising in noncommercial spaces. Just as going to a public place and shouting "Buy my widgets!", this method of advertising may bring derision, particularly if it disturbs the decorum the people in that public space previously enjoyed. Although there may be no "Net cops" monitoring what is said and done, inappropriate communication risks invoking the wrath of a community. In contrast, the same widget seller in the market bazaar (or the Web-equivalent virtual mall) would be welcomed, because the users going into that marketplace know that they will see ads. The enthusiastic widget seller may be eagerly approached by those looking for very good widgets.
Examples illustrating appropriate and inappropriate advertising demonstrate the developed sense of community responsibility and tradition that has evolved over time in networked communities. Web traffic occurs in the context of these traditions. In contrast, the interoffice memo and the internal report exist within a closed environment-closed not just by proprietary considerations, but by the space and time limits inherent in the paper memo as a communications medium. This is not to say that there are no private, proprietary spaces on the Web. Indeed, an organization or individual would not even have to link the hypertext to the Web, and servers can support restricted access via passwords and machine names. A local community still can evolve on private, internal webs, however, and display all the cultural and psychological effects that have occurred in computer-mediated communications systems for decades-community building, social practices, emotional interactions, and conflicts.
Because communication on the Web exists within a larger community, the information provider must cope with the relationships arising from these connections. Web communities evolve over time, and relationships may cross national, cultural, language, space, and time borders. The challenge for this larger Web community is to negotiate the norms for individual interactions appropriately.
Web communication processes are dynamic. Traditional information-development practices have long recognized the iterative nature of the process of creating and delivering information. Web communication, however, involves not only iterative development, but offers a delivered artifact that is conceptually and physically very different from that of traditional media. Web communication does not need to be fixed in its delivered form, and it exists within an information flux. Someone preparing a report often goes through the process of editing, revising, reviewing, user testing, and revising again. Eventually, the deadline clock ticks, and the information takes a final form. Although changes can be made, and there are very possibly second, third, and more editions of the work created, the sense by all parties involved is that the work is "completed" when it is etched into a medium, such as paper, a CD-ROM, a computer disk, or a video tape.
On the Web, hypertext links, the multiple interactions with and among users, and the changing Web information universe all mark the Web as a medium attuned to flux rather than to stasis. Although a developer can create a web and deliver it to the world through a Web server, the job as an information developer is not done-in fact, it's just starting. The developer not only must manage the technical operation of a server but also must handle feedback from users and information about the web's place in the constant flow of new information introduced on the Web.
Although the implications for how the Web changes communication go beyond even considerations of space, time, form, delivery methods, context, and information dynamism, these issues are enough to raise awareness of how the Web medium differs from traditional media.