Artha is a handy thesaurus based on WordNet with distinct features like global hot key look up, passive desktop notification, regular expression based search, etc. Artha may be used as a free open-source replacement/clone to the proprietary WordWeb Pro thesaurus (which is also based on WordNet) on Unix-like and Windows operating systems.
HogTrans provides an automatic word translation engine built on statistics of text translations used for free software. It basically provides an automatically created dictionary with multiple translations and example usages for each. HogTrans can import translations from standard GNU .mo-files.
JWKTL (Java-based Wiktionary Library) is an application programming interface for the free multilingual online dictionary Wiktionary. Wiktionary is collaboratively constructed by volunteers and continually growing. JWKTL enables efficient and structured access to the information encoded in the English, German, and Russian Wiktionary language editions, including sense definitions, part of speech tags, etymology, example sentences, translations, semantic relations, and many other lexical information types.
MAPDAV (More Accurate Password Dictionary Attack Vector) is designed to use what is known about users via the /etc/passwd file on Unix/Linux systems to generate a dynamic dictionary of more accurate guesses as to what their possible password may be. It does this by mangling the user's username and user information in various user-specified ways to look for bad password protection practices.
QuickDic parses data from en.wiktionary.org and Beolingus to generate dictionary files that can be used offline. It currently has built-in dictionaries from English to: German (Deutsch - Englisch), Italian (inglese - italiano), Afrikaans, Albanian, Arabic, Armenian, Belarusian, Bengali, Bosnian, Bulgarian, Catalan, Croatian, Czech, Chinese|Mandarin|Cantonese, Danish, Dutch, Esperanto, Estonian, Finnish, French, Greek, Hawaiian, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Irish, Latin, Latvian, Lithuanian, Japanese, Korean, Kurdish, Malay, Maori, Mongolian, Nepali, Norwegian, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Sanskrit, Serbian, Slovak, Somali, Spanish, Swahili, Swedish, Tajik, Thai, Tibetan, Turkish, Ukrainian, Vietnamese, Welsh, Yiddish, and Zulu.
TAO Translator (formerly, The Advanced Online Translator) uses online translation services to provide translations. Currently supported services: Google Translate (supports translation, language detection, dictionary, and reverse translations for single words); Microsoft Translator, a.k.a. Bing Translator (supports translation only); Yandex.Translate (supports translation and language detection); and Yandex.Dictionaries (supports dictionary with synonyms and reverse translations). More services are possible in the future.
Unibas is a program (a "fat client" of PostgreSQL) to manage frequently used entity types. Its mission is to create a relational database that is both human- and machine-readable (semantic) to collect data about people, documents (books, songs, movies, etc.), products (CDs, DVDs, etc.), fictional characters, events, places and other entities for personal and collective use. It features complete archive management. You get a document (text, image, music, video) from somewhere (e.g. from the Internet) and tell Unibas to take care of it. Unibas does the rest. Album management: CDs can be lost or destroyed by scratches or heat. Unibas makes it easy to back them up, including most of the metadata (composers, artists, titles, etc.). Organize your knowledge in a tree structure like most modern scientific books, yet extended over the complete human knowledge. Link your knowledge with existing knowledge in the tree and with external documents. Tap the many human-readable sources on the Internet and put their content in an ordered, machine-readable, semantic form. Explain words in a dictionary through well understood language-agnostic notions.