The present invention relates to a method of generating a distributed text
index for parallel query processing by a number of nodes. A set of node
indices is generated for text indexing a set of documents, each node text
index covering a subset of the documents. For each node text index, a
local frequency measure for each term of the node text index is
calculated on the basis of a frequency of documents containing the term
in the subset of the documents of the node. A global frequency measure
for each term is calculated on the basis of a frequency of documents
containing the term in the set of documents. A quality measure for each
node text index is calculated on the basis of the local frequency
measures of the terms of the node and the global frequency measure of the
terms of the node.