Text::Document is a collection of modules which allow to operate on text documents from the perspective of Information Retrieval. Text::Document scans documents, extracts terms, compares pairs of documents using the Jaccard and Cosine similarity measures. Text::Bloom allows to compute Bloom filters which compactly store information about term presence in documents, thereby allowing for efficient storage of document 'signatures'. Text::DocumentCollection is a collection of documents, allowing for persistency and for such calculations as the Inverse Document Frequency (IDF). Version 1.04 of the package Text::Document is Copyright (C) 2001 Andrea Spinelli and Walter Vannini All documents in this package can be used with the same limitations as Perl itself. Anyway, we are eager to know about your experiences with this thing, at spinellia@acm.org and/or walter@humans.net. Version 1.05 eliminates the binary checksum of Text::Document and introduces a textual one which is sexier, e.g. in XML. Version 1.06 corrects a quirk in the tests and changes the size of a prime, p, used in Text::Bloom, which caused errors on my linux box. Version 1.08, nineteen years later (sigh!), solves a missing requirement quirk.