TextRoller 2.1 (Windows)
TextRoller will query Internet search engines, retrieve and digest the documents. Alternatively, it can be addressed to HTML / DOC files collections on your local drive. TextRoller has 3 functional modes. In the Build Digest mode, the program produces a summary of each document according to the requested degree of conciseness, and also displays the main concepts (single terms & phrases) found in the processed material. TextRoller not only shows what the document is about (say, in order to find out whether it is worth reading), but retrieves the most important information. It will not present you with sentences that refer semantically to the information (passages) missing in the output text. So, TextRoller Digest can in many cases be used as a reliable information source by itself, instead of the original document. Query-Biased Digest. If this option is checked, you get only those passages from the digest of document, that contain the terms/phrases that you've specified. Presence of these terms in the produced digest (or its close surrounding) makes it more focused on your information need. In the Build Report mode, TextRoller assembles a new query-relevant text from multiple documents passages. This is performed by ways of extraction and ordering the query-relevant information. Only pasages not refering to their surrounding in the source text are picked up. Extracted passages are arranged in such a way that they can be read as a new independent information document relevant to your query (with links to originals).