Generation of Compatible Queries

In order to study query processing efficiency and effectiveness using a simulated text corpus, it is necessary to obtain a set of compatible queries and judgments. The text generation methods implemented in Syntha Corpus make no pretence of being able to generate meaningful natural language. Consequently, it is presently out of the question that ad hoc queries along the lines of those in the TREC1 ad hoc task, could be devised or judged.