Just-In-Time Hypermedia 33
nisms to record the location and content of anchors.
Re-identification requires a criterion to revalidate the
virtual elements and some information about the virtual
elements are needed.
2. Static vs Dynamic Links
Hypermedia is a concept that encourages authors to
structure information as an associative network of nodes
and interrelating links [1]. Hypermedia researchers view
the terms hypertext and hypermedia as synonymous and
use them interchangeably. Hypermedia nominally applies
hypertext concepts to multiple media. A hypertext sys-
tem is made of nodes (concepts) and links (relationships).
A node usually represents a single concept or idea. It can
contain text, graphics, animation, audio, video, images or
programs. In most hypermedia systems, a node usually is
a document. It can be semantically typed (such as detail,
proposition, collection, summary, observation, issue)
thereby carrying semantic information [2]. Nodes are
connected to other nodes by links. Links represent rela-
tionships between two nodes. Static links are direct, per-
sistent connections from one node or anchor to another.
Dynamic links on the other hand are computed each time
an attempt is made to traverse them.
The static links suffer from a variety of problems [3].
These include:
• Dangling links: when a link points to a node that is
subsequently moved or deleted, that link becomes stale;
attempts to traverse it will fail.
• Inexpressiveness: some conventional systems lack
link semantics information. Even in systems that do pro-
vide link types, if the number of link types is set in ad-
vance, then the information content that can be attached
to a link is restricted to the defined set.
• Expensive construction: manually constructing
links is time-consuming and expensive.
• Inflexibility: manual links are created once and are
thereafter fixed; they are unable to rearrange themselves
to suit the needs of the moment.
• Duplication: The semantic knowledge implicit in a
particular link or link type cannot easily be reused or
generalized.
• Irrelevance: When a node’s content changes or
other conditions change, the link may no longer valid.
Dynamic links are created at run-time rather than be-
ing generated earlier (pre-computed links). One approach
to dynamism in hypertext focuses on computing links
based on relationships or similarities between texts or
passages of text. In this approach, the link is not defined
as a pointer from one hypertext node to another, but
rather as a query that leads to a different node [4].
Pre-computed links can be constructed at any time,
whereas dynamic links are computed at the moment they
are required [5]. A dynamic link can take into account
the specifics of the current user interaction. The query
might be based on a co mbinatio n of the browsing histor y,
user profile, content of the current document, etc.
Compared to static links, dynamic links are more
flexible than static links. Dynamic links are dynamically
computed according to the underlying relationships be-
tween nodes or elements. Since dynamic links are
computed at the time when users request to do so, there
is no need to duplicate the links manually and it is not
time consuming. On the other hand, dynamic links also
can suffer from the dangling link problems. Dynamic
link production, however, can detect a dangling link
problem when it occurs and can guarantee that no dan-
gling links are displayed. Dynamic links can allow links
to be created from and to material that the link author
does not control by overlaying the link within the docu-
ment sent to a browser. When a node’s content changes
or other conditions change, dynamic link production has
the potential to detect th at the link is no longer valid.
Advantages of dynamic links include: a simplified in-
terface to search functionality, reduced authoring effort,
and greater opportunity for customization based on the
current user’s interaction history or specific task context.
An additional benefit is decreased cost of maintenance.
However, disadvantages include computation of each
link at run-time (instead of using stored, pre-computed
links), over-completeness and problems with false links.
Computation links at run-time can be expensive in a
large system with many users [5]. Over-completeness
occurs when the reader is presen ted with more links than
he/she can comprehend [6]. False links may occur be-
cause of polysemy, i.e., multiple words and sentences
that lead the search algorithm to incorrectly judge the
similarity of text fragments. This may lead to unsound
links [6] in the hypertex t. JIT systems support both static
and dynamic links.
An anchor in the proposed JIT system is a selection
(which could be part of the document or the whole
document) in a dynamically-generated virtual document.
A link in the JIT system is the connection between two
selections or multiple selections in one or more docu-
ments. JIT supports static and dynamic links. Static links
in JIT are manual links created by users, which is similar
to traditional systems. What JIT offers new are
re-location and re-identification processing when previ-
ously-displayed documents are redisplayed. Dynamic
links in JIT system are computed, based on the inherent
relationships between elements, metadata structure,
document structure, relationships among application spe-
cific commands and parameters, user group relationships,
etc. These relationships could be analyzed before the
hypermedia interface is designed or at the time an anchor
is selected, and JIT then dynamically determines the
links. JIT can support node types by analyzing the dy-
Copyright © 2013 SciRes. JSEA