Technical documents use cross-references to numbered equations and theorems to make the presentation succinct. Cross-references in the running text typically use a system-generated number, e.g., equation [tex2html_wrap5896]. Even when reading printed material, this convention can present problems. Consider, for instance, a proof that reads:
By equation 3.1 and theorem 4.2 and equation 8, we get equation 9 and hence the result.If the above is difficult to understand when reading it in print, it is unusable when encountered in a spoken document, where the listener does not have the luxury of quickly scanning back to the cross-references.
The ability to follow a cross-reference tag during the rendering does mitigate this problem to some extent. However, following such cross-references can prove distracting. AsTeR overcomes this by allowing the listener to label cross-referenceable objects with meaningful labels when they are rendered. These user-supplied labels are used when rendering cross-reference tags, almost eliminating the need to follow cross-references.
To give a concrete example, consider listening to a book on Fermat's last theorem. The first chapter might introduce the subject by stating the theorem. Assume that this theorem is numbered 1.1. The rest of the book might refer to this theorem by number, as in:
As a corollary of this result, we can prove theorem 1.1.When the corresponding document is rendered by AsTeR , the listener hears the theorem, the system generated number (theorem 1.1), and a fleeting sound cue. At this point, the user can give the theorem a more meaningful name by pressing y and entering an appropriate name. Assume the listener enters ``Fermat's last theorem''. Later, when rendering cross-references to this theorem, the newly entered label will be used instead of the system generated theorem number. Thus, the example shown above would be rendered as: As a corollary of this result, we can prove Fermat's last theorem.