This paper introduces an efficient incremental LL(1) parsing algorithm for use in language-based editors that use the structure recognition approach. It features very fine grained analysis and a unique approach to parse control and error recovery. It also presents incomplete LL(1) grammars as a way of dealing with the complexity of full language grammars and as a mechanism for providing structured editor support for task languages that are only partially structured. The semantics of incomplete grammars are presented and it is shown how incomplete LL(1) grammars can be transformed into complete LL(1) grammars. The algorithms presented have been implemented in the fred language-based editor