http:^^www.cs.washington.edu^homes^joebob^590hf^project.html
来自「This data set contains WWW-pages collect」· HTML 代码 · 共 284 行
HTML
284 行
Date: Tue, 10 Dec 1996 21:39:28 GMTServer: NCSA/1.4.2Content-type: text/htmlLast-modified: Thu, 08 Jun 1995 17:02:11 GMTContent-length: 11288<html><head><title>CSE 590H Course Project: </title></head><body><h1>Organizing Web Documents In A Desktop</h1><h2>Andrew Calkins and <a href="index.html">Joe Sherman</a></h2><hr><h2>Motivation</h2><p>The World Wide Web provides addressing and resource namingconventions that permit several kinds of computer resources to beshared globally. These resources include:<ul><li>Hypertext and hypermedia documents<li>Still images and movies<li>Interactive menus, buttons and dialogs<li>File transfer</ul><p>Access to these resources is available through application softwareknown as Web <em>viewers</em> or <em>browsers</em>. There is no inherent,centralized organization to the Web. As a result, Web indexes and searchmechanisms have been developed. Users are free to build their ownhypertext index or save pointers to Web pages in lists (e.g. Mosaic's<em>Hot List</em> or Netscape's <em>Bookmark</em> feature). These lists arecontrolled by menus included in the browser's interface.<p>The NSF has designated the integration of Web browsers into the clientworkstation desktop as a<a href="http://www.cc.gatech.edu/gvu/nsf-ws/report/Visualization.html#HDR3">research priority for the World Wide Web</a>.This integration would eliminate the lack of symmetry between a user'slocal files, documents and services, and those available on the World WideWeb. If done right, we believe, it would assist the user in theorganization and access of this content so that items found once couldbe easily found again.<hr><h2>What Does "Integrate Into The Desktop" Mean?</h2><p>The NSF's research priorities don't explain in detail the phrase,"integrate into the desktop". For this project we will assume that"desktop" refers to user interface software such as:<ul><li>Microsoft Windows Program Manager and File Manager<li>Apple Macintosh Finder<li>X Windows window managers (mwm, twm, ...)</ul><p>These user interfaces provide the following services, to varyingdegress:<ul><li>Organize, view, present available files, resources, documents<li>Launch applications, tasks, projects<li>Permits user to personalize <em>their</em> workstation</ul><hr><h2>How Do We Organize Resources Today?</h2><p>In today's "desktop" interfaces there are different ways of organizingand displaying local and Web resources:</p><h3>Local Resources</h3><ul><li>Hierarchical file systems<li>Hierarchical viewers (2D and 3D)<li>Nested "folders"<li>Program groups<li>Varying degrees of support for direct manipulation ("drag-and-drop")</ul><h3>Web Resources</h3><ul><li>Hot list (with or without limited nesting)<li>Bookmarks (with or without limited nesting)<li>Menu driven or dialog driven control<li>...or we write our own HTML pages</ul><hr><h2>Problems</h2><p>Before discussing what exactly is meant by "integration of a browserinto a desktop," let's list some of the problems casual users seem to haveusing the Web. In each case, let's also ask if this problem might appearwith any distributed, decentralized collection of resources.<p>We first noticed several of these problems watching a user navigatethe Web while trying to build a set of Web pages to serve as a topictutorial.</p><ul><li>There is no support for direct-manipulation arrangement of Web views.Today's browsers have at most only primitive mechanisms for organizingWeb resources according to personal preference. Hot lists are hierarchicalbut menu-driven and non-visual. Bookmarks have similar limitations.Marc Andreeson, "inventor" of Mosaic and founder of Netscape, admittedin a recent talk that "indexing, navigational aides and contentorganization will quickly supersede current layout and design issues"and that "ultimately, user will have complete control over how content isviewed."<li>Cumbersome menus prohibit "on-the-fly" organization.<li>Users often spend time finding the same page more than once.Accessing a Web page you've seen before is not quick and painless.Some people can remember long URL strings, many people can't. Hotlistsand bookmarks help, but as these grow in size, they become difficult tomanage. In a <a href="590h3.html">brief experiment</a> where a person wasinstructed to locate and arrange Web pages, the subject often had to locatethe page several times because he had forgotten the URL.<li>Users must remember and type URLs.<li>Home page uses may conflict. Is the purpose of a home page toorganize the owner's view of the Web? Or is the purpose to let othersknow about the owner and the owner's interests? These are not the samegoals.</ul><hr><h2>Preliminary User Studies on Organization and Navigation</h2><a name="study"></a><p>To focus on the ways in which a user retains and organizes access toWeb documents, we studied a novice Web user in a<a href="./study-1.html>quick usability test</a>.<hr><h2>Initial Investigations</h2><p>We thought about repeated access to Web pages, and how a personorganizes pointers to Web resources for future retrieval. It seemed to usthat many of the direct manipulation techniques included in currentcomputer"desktops" could be applied to Web resources. In this manneraccess to the Web would be "integrated" into the desktop.</p><p>In the next section we'll try to explain how this would work.Comments and suggestions are welcome.</p><p>Later on we'll have more to say regarding the desktop metaphor.<p>The following ideas guided our design:<ul><li>Leverage interactions with current desktop software.</li><li>Visualize Web similar to storage devices (like disks).</li><li>Use icons to represent Web pages</li><li>Use hierachical or nested views to organized Web pages.</li><li>Provide direct-manipulation operations on Web page icons.</li><li>Activate icon to launches pages.</li><li>Icons can be generated and saved to desktop view</li></ul><hr><a name="proto"></a><h2>Paper Prototype Design</h2><p>We've included a description and drawings to<a href="./design-1.html">tour our design</a>.<hr><h2>User Study With Prototype</h2><a name="study2"></a><p>We invited out test subject to play with our paper prototype.<a href="study-2.html">Our observations and her suggestions</a>are described in this section.<hr><h2>Discussion</h2><p>In one sense, integrating Web access into the computer desktopreally means integrating Web "objects" into other applications on thedesktop. The desktop serves to organize and launch the computer'sresources in support of some activity. Many common task can beinitiated by direct manipulation techniques that move a document ontoicons representing available services (like "Printer" or "Mailbox").<p>Newer platforms are capable of supporting 3D interfaces(see VRML and Hot Java). This work could be implemented in a 3D interface,but would not take advantage of perspective, eye position and viewingdirection. Several 3D hierachy views could be used to visualize theuser's organization of Web documents. 3D visualization may be moreuseful for navigation and exploration of the Web than organization.<hr><h2>Metaphors</h3><p>We haven't tried to apply a novel, reality-based metaphor.That may be either a curse or a blessing.Instead we've decided to extend the popular "desktop" metaphor.We feel that this metaphor, in several of its implementations, hasassumed a "form of life" of its own almost separate from physicaldesktops. Computer users today for the most part are familiar withsome desktop GUI. The use of folders to organize directory trees hasproven useful in desktops, even though physical file folders are notoften hierarchies.<p>We have envisioned an extension to thesemetaphors that attempts to make network resources appear similar tolocal resources. This has already been done to some extent withprporietary networks and research prototypes, but not specifically withthe Web.<p>What we've envisoned is consistent with standalone computer desktops.We need however to address network delay and reliability in accessing networkresources. In what way would a user accept asynchronous activities(for example, would it make sense to delegate something to an "agent"or dispatch it, and ask for an interruption upon completion?).<p>This approach retains the key problem interfacing with a large or deephierarchy. Items deep in the hierarchy are hard to see. A "Find File"or "Search" tool is usually provided to help users find what they haveburied.<p>In class we mentioned several metaphors for Web access that differgreatly from the desktop metaphor: The Web as an ocean, the Web as ahighway, for example. We discussed these metaphors and decided thatthey have more use as metaphors for navigation and browsing (e.g. "fishing","trawling", "wandering", "road trip"). When applied to organizationproblem, they don't appear to offer assistance. In real life we cannotarrange the ocean or the interstate highway system to suit our ownneeds.</p><hr><h2>Summary</h2><p>We have discussed several extensions to this approach:<ol><li>The drag-and-drop mechanism could be used to connect documents andobjects to services. The objects and the services could be local orremote. Example objects include movies, audio, pictures, spreadsheetsand commodities for sale. Services could include viewers, word processors,spreadsheet applications, mailers, printers, fax software, shopping andpurchasing software, and personal organizers.<li>The Web information structure could be extended to support theseadditional object types. This would associate services with objectssimilar to the MIME concept. Default operations for direct manipulationcould be associated with the service/object combination.<li>The desktop appears to be a suitable way for users to organize theirpersonal views of the Web. But it does not seem to provide an idealmetaphor for exploration or navigation. Metaphors appropriate forexplorations are still needed.<li>The interface could support personal annotations by overlaying thedocument page with "Post-It" type annotations that could be stuck anywhereon the document. These would be associated with, and accessible from,the document icon. They would be stored locally and attached wheneverthe document was referenced.</ol><p>Our approach could suggest modifications to the Web information structureand HTML. Authors and users often collect links into a page thatacts as a directory or folder. Often these pages have no content of theirown, or may contain personal annotations. What if HTML were extendedto incorporate directories or folders (instead of logical types likeordered and unordered lists)? Browsing software and viewers could thensupport presentation metaphors suited more for organiztion than forreading. As an extension it would neither force authors to use it norviewing software to present it differently from a document.<p>We feel that Web viewing software can be easily enhanced to supportdynamic, personalized organization. Leveraging existing desktop softwareis a natural way to do this.</body><hr><address><a href="index.html">Joseph M. Sherman</a><br>Last modified: Friday, May 5th, 1995.</address></html>⌨️ 快捷键说明
复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?