|Taxonomic: Using anchors to indicate link types|
|Hypertext developers have
long been postulating link types and classifications. Indeed Vannevar Bush
suggested a "special button" (now termed anchor) that users would
name themselves to provide associations, to realize his hope that "Selection
by association, rather than indexing, may yet be mechanized. [77, Section
A brief history of taxonomic speculation in the hypertext community
In 1983, Randall Trigg postulated a complicated system of link typing : "One of the most important features of the Textnet approach is the extraction of semantic content from text by making the relationships between nodes explicit. This is accomplished by joining the nodes with typed links. "
In 1993, Marc and Jocelyn Nanard took link typing to the anchor level in MacWeb, a system that would map a virtual hypertext organization onto a document base. They postulated that knowledge "is the essence of hypertext itself. . . the base for an intelligent navigation between documents" and therefore "knowledge handled by the hypertext should describe anchors themselves and not only nodes.  "
Mark Bernstein urged a second look at Link Types in Hypertext Now " By increasing our range of hypertextual expression, link types might well give all hypertext writers -- not just scientists -- a denser and richer medium. "
What about a simpler taxonomy?
Others have assumed and given names to links that indicate
the destination, even without a more complex taxonomic system. Anna Gunder
 defines categorized links as those that "provide the
users with more or less infomration on the link destination" and
uncategorized links as those that "give no information whatsoever
to the link destination." This discussion must really be about the
anchor part of the link, unless there is software that would tell the
reader where the link leads as the link is triggered or being traversed.
The only anchors that we found that would be at all categorized were those
in menus or maps
(which indicate the heirarchy or placement of nodes) or denotative
anchors (which indicate the title of the node). Most embedded
or connotative anchors were not categorized.
Many have bemoaned the simplicity of the web link as it precludes systematic taxonomy or even classification. Shorewalker explains and notes a failure: "The links of the World Wide Web create at least three potential problems. For a start, they're dumb; they contain little information. The context of the surrounding words, pictures and sounds may make it obvious that a blue, underlined, hyperlinked word leads to a particular node elsewhere on the Web. But in many cases - as any Web user knows - a link will leave it to the reader to guess what might be on the other end. Somehow or other, you should describe your links. (The Web magazine Feed once attempted to use marginal notes in the text to tell readers what sort of link they were clicking, but the results were not entirely successful.)  "
Shorewalker goes on to relate some solutions that haven't quite worked: "Fifth, you can try implementing a visual code that tells users what sort of link they're clicking. This sounds a neat solution, and it has been patchily implemented here on some pages of the Lighthouse on the Web site, and with more success by MSNBC. One firm, Matterform Media, has gone so far as to invent a whole system of tiny icons called QBullets, intended to signify different sorts of links - to form-based pages, brief notes, lists of other links and so on. In practice, though, these systems look awful. Users don't seem likely to start learning yet another new system to help them understand badly-designed sites. And authors can't easily classify all the documents to which they link - a point discovered early by the pioneer of link classification, Randall Trigg.  "
Anchors have not been widely used to denote types of destinations. Postulating reasons for this phenomena would be pure speculation.