You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
182 lines
7.3 KiB
182 lines
7.3 KiB
<html xmlns="http://www.w3.org/1999/xhtml">
|
|
<head>
|
|
<meta name="generator" content=
|
|
"HTML Tidy for Mac OS X (vers 31 October 2006 - Apple Inc. build 13), see www.w3.org" />
|
|
<title>
|
|
HyperText Design Issues: Navigational techniques
|
|
</title><!-- <NEXTID N="13"/> -->
|
|
</head>
|
|
<body bgcolor="#FFC060" text="#302005">
|
|
<a href="OldDocs.html"><img src=
|
|
"../Icons/WWW/arch1990" /></a>TimBL
|
|
<hr />
|
|
<h1>
|
|
Navigational Techniques and Tools
|
|
</h1>
|
|
<address>
|
|
<a name="anc5" href="../People.html#BernersLee" id=
|
|
"anc5">TBL</a>
|
|
</address>There are a number of ways of accessing the data one
|
|
is looking for. Navigational access (i.e., following links) is
|
|
the essence of hypertext, but this can be enhanced with a
|
|
number of facilities to make life more efficient and less
|
|
confusing.
|
|
<h2>
|
|
Defined structure
|
|
</h2>It is sometimes nice for a reader to be able to reference
|
|
a document structure built specifically to enhance his
|
|
understanding, by the document author. This is especially
|
|
important when the structure is part of the information the
|
|
author wishes to convery.
|
|
<p>
|
|
See a <a name="anc2" href=
|
|
"../../Conferences/ECHT90/Structured.html" id="anc2">separate
|
|
discussion of this point</a> .
|
|
</p>
|
|
<h2>
|
|
<a name="anc6" id="anc6">Graphic Overview</a>
|
|
</h2>A Graphic overview is useful and could be built
|
|
automatically. Should it be made by the author, server, browser
|
|
or an independent daemon?
|
|
<p>
|
|
Can one provide an overview with less granularity than the
|
|
basic web by grouping nodes in some way? The user could
|
|
select from <a name="anc3" href="Topology.html#4" id=
|
|
"anc3">link types</a> used to imply the tree
|
|
structure.<a name="anc4" href="../People.html#groff" id=
|
|
"anc4">(JFG)</a>
|
|
</p>
|
|
<p>
|
|
I think this depends on how long it will take. It might be
|
|
interesting to experiment with daemons which will
|
|
independently make and update maps of the web. This is not
|
|
essential for a first pilot model.
|
|
</p>
|
|
<h2>
|
|
History mechanism
|
|
</h2>This allows users to retrace their steps. Typical
|
|
functions provided can be interpreted in a hypertext web as
|
|
follows:
|
|
<dl>
|
|
<dt>
|
|
Home
|
|
</dt>
|
|
<dd>
|
|
Go to initial node
|
|
</dd>
|
|
<dt>
|
|
Back
|
|
</dt>
|
|
<dd>
|
|
Go to the node visited before this one in chronological
|
|
order. Modify the history to remove the current node.
|
|
</dd>
|
|
<dt>
|
|
<a name="z12" id="z12">Next</a>
|
|
</dt>
|
|
<dd>
|
|
When the current node is one of several nodes linked to the
|
|
ªbackº node, go to the next of those nodes. Leave
|
|
the ªBackº node unchanged. Modify the history to
|
|
remove the current node and replace it with the "next" (new
|
|
current) node.
|
|
</dd>
|
|
<dt>
|
|
Previous
|
|
</dt>
|
|
<dd>
|
|
When the current node is one of several nodes linked to the
|
|
ªbackº node, go to the preceding one of those
|
|
nodes.
|
|
</dd>
|
|
</dl>In many hypertext systems, a tree structure is forcibly
|
|
imposed on the data, and these functions are interpreted only
|
|
with respect to the links in the tree. However, the reader as
|
|
he browses defines a tree, and it may be more relevant to him
|
|
to use that tree as a basis for these functions. I would
|
|
therefore suggest that an explicit tree structure not be
|
|
enforced.
|
|
<p>
|
|
(If a default tree is needed by the system for some reason,
|
|
then we can always use the creation order: when a node is
|
|
created it is always created with a link to an existing node.
|
|
Such links, whatever their type, may be used to define a
|
|
tree. If they are deleted, an alternative link must be chosen
|
|
to become a tree link.)
|
|
</p>
|
|
<p>
|
|
If authors want to write a tree structure into their
|
|
documents, then the words "after", "before" and "above" could
|
|
be used to mean a static structure.
|
|
</p>
|
|
<h2>
|
|
Intelligent navigation
|
|
</h2>See A. Secret's <a name="z11" href=
|
|
"Intelligent_Navigation.html" id="z11">discussion of
|
|
intelligently navigation techniques</a> .
|
|
<h2>
|
|
<a name="anc8" id="anc8">Index</a>
|
|
</h2>An Index helps new readers of a large database quickly
|
|
find an obscure node. Keyword schemes I include in the general
|
|
topic of indexes. <a name="anc9" id="anc9">The index must, like
|
|
a graphic overview, be built either by the author, or
|
|
automatically by one of the server, browser, or a daemon</a> .
|
|
The index entries may be taken from the titles, a keyword list,
|
|
or the node content or a combination of these. Note that
|
|
keywords, if they are specifically created rather than random
|
|
words, map onto hypertext ªconceptº nodes, or nodes
|
|
of special type ªkeywordº. It is interesting to
|
|
establish an identity relationship between keywords in two
|
|
different databases -- this may lead a searcher from one
|
|
database into another.
|
|
<p>
|
|
Index schemes are important but indexes or keywords should
|
|
look like normal hypertext nodes. The particular special
|
|
operation one can do with a good keyword index system which
|
|
one can't do with a normal hypertext system is to do a fast
|
|
search on multiple keywords. This must to be provided as an
|
|
extension to the hypertext navigation scheme. However, it is
|
|
in fact analogous to a trace starting with more than one
|
|
node, which is a valid hypertext tracing operation. The
|
|
difference is that the tracing would normally be done by a
|
|
browser, but the indexed search done by the server.
|
|
</p>
|
|
<p>
|
|
When many nodes in a web represent different indexes, then a
|
|
query search can chain between them (See " <a name="anc7"
|
|
href="ManyIndexes.html" id="anc7">Web of indexes</a> ").
|
|
<a name="z10" href=
|
|
"http://www.vuw.ac.nz:80/who/Nathan.Torkington/ideas/reggie.html"
|
|
id="z10">Nat Torington's musings</a> .
|
|
</p>
|
|
<p>
|
|
See also: <a name="anc1" href=
|
|
"../../Conferences/ECHT90/HTandIR.html" id="anc1">HyperText
|
|
and Information Retrieval</a>
|
|
</p>
|
|
<h2>
|
|
Node Names
|
|
</h2>These allow faster access if one knows the name. They
|
|
allow people to give references to hypertext nodes in other
|
|
documents, over the telephone, etc. This is very useful.
|
|
However, in Notecards, where the naming of nodes was enforced,
|
|
it was found that thinking up names for nodes was a bore for
|
|
users. KMS thought that being able to jump to a named node was
|
|
important. The node name allows a command line interface to be
|
|
used to add new nodes.
|
|
<p>
|
|
I think that naming a node should be optional: perhaps by
|
|
default the system could provide a number which can be used
|
|
instead of a name.The system should certainly support the
|
|
naming of nodes, and access by name.
|
|
</p>
|
|
<h2>
|
|
Menu of links
|
|
</h2>Regular linkwise navigation may be done with
|
|
ªhotspotsº (highlighted anchors) or may be done with
|
|
a menu. It may be useful to have a menu of all the links from a
|
|
given node as an alternative way of navigating. Enquire, for
|
|
example, offers a menu of references as the only way of
|
|
navigating.
|
|
</body>
|
|
</html>
|