The technique of error-trapping decoding for algebraic codes is studied in combinatorial terms of covering systems. Let 

 , and 

 be positive integers such that 

 . An 

 -covering system is a pair 

 , where 

 is a set of size 

 and 

 is a collection of subsets of 

 , each of size 

 , such that for all 

 of size 

 , there exists at least one 

 with 

 . Let 

 denote the smallest size of 

 , such that 

 is an 

 -covering system. It is shown that the complexity of an error-trapping decoding technique is bounded by 

 from below. Two new methods for constructing small 

 -covering systems, the algorithmic method and the difference family method, are given.