In the indexing method presented in this section the index is computed off-line, without parsing a training corpus. The index key for each daughter is represented by its position (rule number and daughter position in the rule, hence the name ``positional'' indexing), therefore no time is spent during parsing for computing the index keys.
A related method (rule filtering) was proposed in [Kiefer et al.1999] (described in Section 4.4.3). However, the authors use a filter to avoid unsuccessful unifications. Using an index offers the advantages of a flexible, yet organized approach: as will be shown over the next sections, the indexing scheme developed here can be enhanced with information extracted from further static analysis of the grammar rules, while maintaining the same general indexing strategy.