A link is the traversable connection between two nodes. An anchor is the visible region which must be selected to activate the link. These may vary in size from one word to the entire contents of the node. For example, in this current node, the word "nodes" in the first sentence is an anchor. It is the visible part of a link between the nodes "Links and Anchors" and "Node." In theory, anchors may overlap; in practice, most systems insist on their being distinct, in order to make the underlying mechanics simpler.
The granularity of link destinations is an important qualifier. Some systems support only basic links, eg. those which connect two nodes without specific anchors. Source anchors are available in almost all hypertext systems. Destination anchors allow a particular region of a node to be addressed; this becomes the focus when the link is traversed. For multi-screen nodes, this is a helpful feature. However, destination anchors are supported in only some hypertext systems. More often, the entire destination node is referenced, and the reader is placed at the beginning of the text. This may result in disorientation, particularily if more than a screenful of text must be traversed before the passage relevant to the link is discovered.
Many software packages refer to anchors as "buttons," though the latter is more properly considered a feature of the user interface and not the underlying functionality. Anchor appearance varies in different systems.
Once an anchor is selected, the link is activated. If more than one link is available for the anchor, either a choice is presented to the reader, or the system selects from the possibilities according to pre-defined criteria. The actual navigation of the link proceeds according to the link class.
Links may have other characteristics such as name, author, creation date, and access rights.
A discussion of the rhetoric of link traversal is available in a nearby node.