Abstract 1 1 Introduction 2
Example of a Document and Document Space User Interface for Navigation
Download 324.2 Kb.
|
- Navigate this page:
- 4.3.1.1Display for document content
- 4.3.1.2Display of document identifying attributes
- 4.3.2Displaying a Document Collection
- 4.3.3Displaying a Document Set
- 4.3.4Displaying the Document Universe
- 4.3.5Displaying Navigation Information
4.3Example of a Document and Document Space User Interface for Navigation4.3.1Document DisplayA document is presented in two major views, document content and document identity. As content, a document is presented in its own space. A document is frequently presented in terms of its identity and other attributes when it is presented as part of a set. At the same time, there may be a dialog or other kind of display that will allow all of the attributes, identity, and other features to be displayed. 4.3.1.1Display for document contentThe main purpose of a document content display is for reading and writing. Normally, a document content display shows text as produced by the author. “What you see is what you get” is the most common scheme. Given the capability to display information graphically, an encoded document style may be decoded to control a presentation. A viewer or browser program is generally the same as the editor program. Even if they are different, they will use the same method to decode data, producing a presentation consistent with that intended by the author. With the emergence of structured document encoding, tag mode displays are common for editor applications. A control code will be displayed in a way that allows structural attributes to be edited. The structure of text, e.g. headings, quotes, etc., is commonly presented using different text attributes. In small display environments, a document may be presented by moving text from right to left or rapid serial visual presentation (RSVP) at word level and sentence level. The usability of RSVP (Rahman & Muter, 1999) shows that sentence-by-sentence display is as efficient for reading as full page display. However, the experiment was conducted only on short messages. For a full-length text, a difference in reading performance is to be expected. Specific documents may have special views. For example, in program source code, the leading space is not significant to the meaning of the program. A program editor may automatically create indents to reflect the structure of the program. Color may also be used to present statement type. The color encoding is not actually encoded in the program text, but is simply provided at the display level. When a text body is larger than a display area, scrolling is the major mechanism for revealing an unseen part. Page up/page down key control is common. Most GUI applications provide a scroll bar to control which area to display. Some scroll bar systems provide a hint about the size of the part of the document currently on display compared to the whole document size by using the variably sized scroll thumb. There are many variations on the controlling of a scroll bar. Clicking on a scroll bar may take the user to the position that was clicked or may simply change pages in the direction indicated by the location of the click. Cursor position may or may not move when changing pages. In MS IntelliPoint™, mouse and trackball may be used to control the velocity of scrolling instead of position to scrolling to. A hypertext document has many presentations. The structure of a hypertext document can be formed by individual users, who create their own links, or by shared groups of authors. An author may create semantic links as in a guided tour. These guides may be varied by groups of users. In this view, for the same collection of nodes, the structure of the hypertext will appear differently. It applies to a hypertext as a document collection as well. A common presentation of hypertext is a page with anchors. Clicking on an anchor will result in a new page replacing the previous one. Alternatively, it could open the page in a new window or pop one up temporally until the user releases the mouse button. Annotation, highlighting, and markers in text can all be used to indicate parts of documents that are of interest. Many applications support annotation, footnotes, and highlighting in electronic text. The CASCADE interface (Spring, Morse, & Heo, 1996) helps the user to find these marks by using mural views and explicit presentation of the markers. MS Word™ has a speed up scroll to a certain marker such as to a comment, via “go to next comment” and “go to previous comment” buttons. A document overview may be presented as a miniature of the lines instead of as character text to present the overall structure, e.g. in Seesoft (Eick, Steffen, & Sumner, 1992) and Thumbar (Graham, 1999). Color of lines may be used to present attributes of interest. Mural technique is applied to text display in Jerding and Stasko (1997). The fish eye view of program text is found in Furnas (1981) and of a document found in Greenberg, Gutwin, and Cockburn (1996) which addresses multi user focus points. An overview of a hypertext document may be presented by a diagram showing nodes as graphic objects and links as connecting lines. This is similar to displaying a hypertext collection. It will be discussed later in this section. An overview of a document may be presented through its structural units. A table of contents presents a linear list or tree structure. It can be a network structure for hypertext documents. The table of contents list may be static or dynamic. In a printed book, an index is used to find a text section by reference to selected keywords along with a page number. In an electronic document, finding can be done via a search function, commonly provided by the application. At the document level, common implementations allows only single-phrase matching, no logical operation on sets of phrases are allowed. The matched phrase will be highlighted, the cursor will be moved to its position, and the screen may be scrolled. The “find next” is also provided. 4.3.1.2Display of document identifying attributesA document may also be displayed in terms of its attributes. These displays provide a way to target an “accessing” action. In command interfaces, a file is shown by its name as a text label. In most GUI file managers, a file is presented by an icon with the file name as label. The icon image normally identifies the file type. Miniatures of the first page of a document may be used on the icon as are used in the Freestyle system (Levine & Ehrlich, 1900) and the data mountain system (Robertson et al., 1998). A complex glyph such as in InfoBUG (Chuah & Eick, 1997) encodes several data attributes into one icon. In a hypertext browser (not to be confused with WWW browsers) view, documents are displayed in terms of identifying attributes in a graph structure. In a WWW browser, an anchor in a source node is a part of the content of the source node. The anchor description may not reflect anything about the destination node. The semantics of surrounding text in content may indicate where the link will lead. The identity of a document in this view is hidden. The link may point to another document, or to some other part of the document within which the link is found. Many hypertext systems present anchors differently for different types of link. In the real world, a printed book gives a lot of visual cues about its physical condition. These physical conditions may be applied in electronic documents as well. For example, the thickness of a book provides a clue about its size. Electronic documents might be measured by file size. However, for those documents that maintain revision history and allow combination of text and pictures, file size may not be a good approximation of size. It is possible to compute a “page size” and store it as data for a document. Number of pages may imply, among other things, the time required for reading. Other physical characteristics such as age, usage, annotations, etc. may also be approximated in electronic form. These attributes may be used to create complex icons for a document, which may help the user to decide which document to view. 4.3.2Displaying a Document CollectionTypically, a document collection display presents the documents in a collection and the relations among them. For a file system, a tree diagram is used to present hierarchical relations. For a database, metadata may be represented as a list of records. The relations in a hypertext system are more complex than those in a file system. Hypertext uses a network model for relations among nodes. A hypertext collection is presented as a network diagram. However, the network relations are sometimes simplified as a hierarchical structure. There are various forms for presenting a hierarchical structure. The overview of a hierarchical structure is normally presented as a tree diagram. Samples of hierarchical overviews are shown in Figure 1. Expanding and collapsing sub-trees is a general strategy to avoid too much node information in a view. A network structure, e.g. hypertext and Web, is commonly presented by a diagram, graph-structure. In a large size network, many links on a display may appear overwhelming. Many tactics are used for reduced display data including grouping nodes and showing only one link between two groups. A group of nodes can be zoomed in on to show hidden nodes. Figure 1 Samples of presenting a hierarchical structure. 
An overview presentation offers its own unique problems in dealing with large amounts of data. The design is always a tradeoff between data that can be displayed and data that will not be visible at the moment. To display all data at once on a limited display space, a data point has to reduce to a very small point. On the other hand, if data are visible at a size that can be readable or selectable by mouse, some data points will be not visible due to occlusions from other data points, or to being out of the boundary of display. A large data set also slows down an interaction process. Many strategies are used to solve these problems, including the following:
A document collection may be viewed in terms of its semantic structure. A site map or table of contents for a Web site might constitute such a semantic structure. A semantic map may be created by analyzing document content and creating a 2D mapping. A common presumption is that if two documents are presented near each other in a spatial display, they are semantically related. All computed distance is relative and commonly, direction has no meaning. Kohonen (1998) uses Self-Organizing Maps to create a document collection map. While conceptually attractive, the method requires an extensive amount of computation, and is generally only suitable for stable collections. 4.3.3Displaying a Document SetA document set is some subset of a document collection -- generally one that is constrained in some way. It may be a result of a query or a document selected by interaction. One important document set is a partial set of a document collection displayed while navigating. In a tree structure, children of any given node might be considered a document set. The result of a Boolean query is an unordered set commonly displayed as an ordered linear list. The result of a query in the vector model is an ordered list, ordered by distance value from a given target, e.g., query. This result is commonly displayed in a linear list form. In both cases, if the result set is large, a limited number of items in a display or a limited number of results may be provided. Visual information retrieval interfaces (VIRIs) display a query and result by using visualization techniques. A summary of systems can be found in Korfhage (1997). VIBE will be discussed as an example of this type of system. Visual Information Browsing Environment (VIBE) (Olsen, Korfhage, Sochats, Spring, & Williams, 1993) uses multiple reference points to create a projection of documents into a 2D display. A reference point is a vector of interesting keywords. Through positioned reference points on a plane, documents can be mapped to a display space according to their distance from those reference points. Documents are presented, bounded by the reference points. Changing the position of a reference point, by interaction, causes document positions to change relative to that reference point, i.e. documents with relatively high values move a distance equal to that which the reference point moves, while documents with relatively lower values move less distance. The navigation in a tree will have two basic directions: from a set of children to the parent and from a parent to its set of children. While there is only one way to go up, there are many ways to select to go down into a sub tree. Commonly in a file manager system, a click on a directory will open a new window or replace the current view showing the new current directories and files in that directory. The parent node may be shown as “..”. In order to go to the parent node, a user may click on the parent node icon or click on the menu. A document set may be presented in a separate view. It may be shown as an ordered or unordered list. Its current position or the constraints that define the set are commonly presented as a position indicator or label of condition. A flexible sort order of a list is used in many systems. Sorting a document set may help in navigation or searching. A document set may be presented in a display of a document collection by changing a graphic attribute of the selected objects. This is commonly used in selection processes. For example, when selecting files in the Windows file manager, the icons of selected documents change color. Information about the document set may be provided in the presentation -- e.g., the total size of the set, the number of items in the set, etc. This information may be aggregated, enumerated, or statistical. Document sets may be dynamically created through interaction. The selection process becomes complex when one considers handling multiple sets simultaneously. Many alternatives are possible: for example, the selected objects of the second selection may be added to the first set, or only the second selected objects are in the selected set. A taxonomy of selection processes and an analysis of possible interactions may be found in Wills (1996). Complex combinations of selecting in a selection process may be described as a Boolean operation of sets, i.e. AND, OR and NOT. The Boolean operations are used in query formation as well. 4.3.4Displaying the Document UniverseIn our definition, the document universe will contain all of the document collections. There is no general presentation of a document universe. A document universe is something at the level of the WWW, or the Library of Congress collection. At this level, we consider how to present and navigate a very large space. The document universe is not homogeneous. It is expanding. It is changing. It may be easier to view on a document collection level, rather than at a level of detail within each document in a universe. The metadata will be a primary source of data rather than a document itself. One attempt at creating a map of the WWW is found in Bray (1996); however, the WWW map is presented at a Web site level. A collection of Internet related maps can be found at the Atlas of Cyberspaces Web site (Dodge, 1998). 4.3.5Displaying Navigation InformationNavigation information is the information that captures the user’s interaction with a document space as it pertains to the navigation process. This information includes a history of interaction, both in terms of timing and location, as well as information about bookmarks, backtracking, etc. The navigation information may form a separate presentation from a document space display, or it may somehow be combined with the presentation of the space. Many hypertext systems provide an active trail list. A history list in Netscape™ is shown as a list of visited sites including title of page, URL, first visit date, last visit date, visit count, etc. It can be ordered by those attributes. In Internet Explorer™ version 5, a history list is shown by date, by site, by most visited, and by order visited today. In “by date” and “by site” views, history is shown as a two level expand/collapse tree. Items are grouped by either site name or date. It is interesting that the date grouping is non-linear -- today, days in a week, last week and last 2 weeks. Instead of showing a trail by ordered list, a trail may be viewed based on visited nodes. In a Web Browser, an anchor may change color when it has been visited. In an overview map of a document collection, nodes and links visited while navigating may be highlighted. One may present only visited nodes and links, similar to an overview map. It will dynamically form network relations while navigating. This is considered as presenting visited sub-space. Samples of this system are WebNet (Cockburn & Jones, 1996), Footprint Site Map, and Footprint Paths (Wexelblat & Maes, 1999). Controlled experiments were conducted and gave positive results in the utilization of tools. A trail list of a current session shows a list of visited pages. The trail list can be shown by expanding the back button (in many Web Browsers). The top item is the most recently visited page – the destination of a click on the back button. There are many schemes that might be used to create this list. The stack-based scheme is commonly used. Dix and Mancini (1998) investigated six history and backtracking mechanisms. The formal definitions are provided. They indicate that the back button is used in different ways in many applications. In general, the linear traversal of links will give the same result. However, when the list includes a node that has been visited several times, each mechanism treats the visits differently. As a result, “go back” will go to different positions. A bookmark is a list of marked locations. In the WWW environment, a bookmark list is shown as a title list of marked pages. Bookmarks tend to be large. Most Web browsers provide hierarchical organization of bookmark items. MS Windows™ implements a bookmark item as a linked file. While bookmarks are still accessible from the menu, MS Files manager™ can also access them. In some implementations, bookmarks are built into an HTML file. Document usage data is also useful, especially for managing a document space. Animation of the number of accesses per day on a given Web site could be very effective for identifying new hot pages on a site. It could also show pages that, over time, are cooling down or becoming of less interest. Similarly, document usage data makes possible to see general growth patterns and clusters of activity. Animation of visitors of a group of Web page can be found in Minar and Donath (1999) Download 324.2 Kb. Share with your friends:
|