Abstract

When input sentences are processed in some manner, say evaluated with some set of rules, the resulting sentence must be pretty-printed in order to be presented to the user. We introduce a technique called "Share-Where Maintenance'' which is used to preserve layout information by annotating abstract representations of visual sentences. The annotations in the abstract representation point to the presentation where sub-terms originated which were created either by the user (initial term) or by the language specifier (equations, as in introduced terms). Both kinds are visual presentations which are used for presenting the new term.