170 likes | 367 Views
Hypertext. Hypertext History (1). Many early attempts to organize human knowledge Thesaurus (Roget) Encyclopedia Memex Conceived by Vannevar Bush in 1945 Electro-mechanical device cataloging and collecting human knowledge. Hypertext History (2). Earliest systems Augment (Engelbart)
E N D
Hypertext History (1) • Many early attempts to organize human knowledge • Thesaurus (Roget) • Encyclopedia • Memex • Conceived by Vannevar Bush in 1945 • Electro-mechanical device cataloging and collecting human knowledge
Hypertext History (2) • Earliest systems • Augment (Engelbart) • Xanadu (Ted Nelson) • Complete, distributed, interlinked system containing all documents in the world • Not technically feasible • Computers/networks weren’t powerful enough • Ted Nelson is too disorganized • AutoDesk tried to resuscitate project in 1990s
Hypertext History (3) • Initial production systems • Intermedia (Brown), NoteCards (Xerox PARC), KMS (Carnegie-Mellon) • Local network, often built on a database • HyperCard popularized NoteCards model • Imitators: SuperCard • Competitors: Guide • The Web • Less functionality, but world-wide
Hypertext Concepts and Goals • Text documents are connected by links • A browser is used to see documents and traverse/follow links • Users are encouraged to read non-linearly • Non-linearity is overrated, convenience is key • Many arty types are trying to construct radically different fictional works exploiting non-linearity
Nodes and Links • A hypertext is a graph with nodes and links • Links generally connect parts of documents • The parts are marked by anchors • sometimes called persistent selections • Links come in many types • Unidirectional (source and destination) • Bidirectional (two anchors, no distinction) • Multi-links (>2 anchors)
Link Example (PDF) • Anchors are rectangular regions of a page • Links are unidirectional • stored in the source document • Acrobat Reader interface • source is outlined • destination is highlighted • Links defined by WYSIWIG interface • Click and drag regions • Normal file/page navigation to connect source/dest
HTML Links • Unidirectional links • “A” elements are anchors • Source must be an anchor • Destination may be an anchor • Uniform Resource Locator (URL) • Simple language for specifying link destinations • One of the three critical features of the Web
URL has 3 parts, plus … • Protocol: http, ftp, smtp, telnet, … • Domain name: host providing protocol’s server • www.cs.uwm.edu • Path: normally a path for file system • e.g. http://www.cs.uwm.edu/~munson
URL Details (1) • CGI queries appear after question mark • Partial URLs provide a path relative to source document • Fragment identifier specifies part of a document • “A” element with matching “name” attribute • Any element with matching “id” attribute • HTML 4.0
URL Details (2) • Link behavior is more complex with frames • must specify which part of interface is changed • specify name of frame or use keywords • _self, _blank, _top, _parent
Other Link Concepts • Earlier hypertext systems stored links separately from the documents • Bidirectionality is easier to support • Link validity can be tested • Web can be visualized and analyzed • Queries can be made efficiently • Web is a step backwards conceptually
XPointer and XLink • XML advocates hope to support more powerful linking services • XPointer specifies link targets within documents • XLink describes complex links
XPointer • Tree selection language • start at root(), or id(name) • navigate with child, descendant, ancestor, preceding, following, psibling, fsibling • four arguments: • number (ith instance found) • element name sought • attribute name • value of attribute • id(intro).following(2, example, quality, bad)
XLink • Supports simple and extended links • Extended links can be multilinks • Discussion of how to store links separately • Can this work on the Web?
Navigation in Hypertext • Easy to get disoriented • Many navigation aids have been proposed • Bookmarks • History lists • Site maps • Web directories • Search engines