Determining the extent of lookahead in syntactic error repair

Abstract

Many syntactic error repair strategies examine several additional symbols of input to guide the choice of a repair; a problem is determining how many symbols to examine. The goal of gathering all relevant information is discussed and shown to be impractical; instead we can gather all information relevant to choosing among a set of "minimal repairs." We show that finding symbols with the property "Moderate Phrase-Level Uniqueness" is sufficient to establish that all information relevant to these minimal repairs has been seen. Empirical results on the occurrence of such symbols in Pascal are presented.

Natural and Flexible Error Recovery for Generated Parsers

“…An alternative approach to local or global recovery is to consider only the direct context of the error, by identifying the region of code in which the errors reside [16,18,21]. Using regions for error recovery has three main advantages.…”

“…Regional error recovery methods [16,18,21] select a region that encloses the point of detection of an error. Typically, these regions are selected based on nearby marker tokens (also called fiducial tokens [21]), which are language-dependent.…”

A formal frame for robust parsing

“…Following Mauney and Fischer in [4], we talk about the error in the input to mean the difference between what was intended and what actually appears. We talk about the point of error as the point at which the difference occurs.…”

“…Since they expend equal effort on all parts of the program, including areas that contain no errors, much of that effort is wasted. Finally, the main problem to be dealt with in regional approaches [4] is how to determine the extent of the repair in order to avoid cascaded errors.…”

“…The repair quality is equivalent to global approaches. Compared to other regional algorithms [4], we provide a least-cost dynamic estimation of this region, which is an advantage in the design of interactive tools, where efficiency is a priority challenge.…”