Hunspell

Hunspell
Name	Hunspell
Developer	* László Németh * OpenOffice.org * Mozilla Foundation
Released	2000s
Programming language	C++
Operating system	Linux, Microsoft Windows, macOS
Genre	Spell checker, Morphological analyzer
License	GPL, LGPL, MPL

Hunspell is an open-source spell checker and morphological analyzer widely used for complex languages and cross-platform applications. It builds on precedents in computational linguistics and natural language processing pioneered in projects such as MySpell, Aspell, Ispell and draws on lexicographic traditions exemplified by the Oxford English Dictionary, Merriam-Webster, and national language institutes like the Instituto Cervantes. Hunspell's design emphasizes extensible affix handling, rich dictionary formats, and integration into major software ecosystems including LibreOffice, Mozilla Firefox, and Google Chrome.

History

Hunspell originated in the early 2000s as an evolution of MySpell to meet the needs of morphologically rich languages such as Hungarian and Turkish. The project was led by László Németh and emerged alongside language engineering efforts from organizations like OpenOffice.org and the Mozilla Foundation. Its adoption accelerated when desktop environments such as KDE and GNOME sought robust multilingual support, and when large projects including LibreOffice, Google Chrome, and Microsoft Office-compatible toolchains required adaptable spell-check backends. Hunspell's trajectory intersects with initiatives from academic centers such as University of Szeged and standards bodies like the Unicode Consortium and internationalization groups within W3C.

Architecture and features

Hunspell's architecture is implemented in C++ and designed as a modular library exposing APIs usable by applications including LibreOffice, Mozilla Firefox, and Thunderbird. Key features include support for complex affix rules inspired by finite-state morphology research conducted at institutions such as University of Helsinki and University of Cambridge, character encoding handling aligned with Unicode standards, and compatibility with legacy formats from MySpell and Aspell. The engine supports custom suggestion algorithms, compound word analysis influenced by techniques used in Lucene and Elasticsearch, and extensibility for contemporary projects like Visual Studio Code and Sublime Text.

Affix and dictionary format

Hunspell uses paired affix (.aff) and dictionary (.dic) files, an approach inherited from earlier tools such as MySpell and reformulated to accommodate agglutinative languages like Finnish, Hungarian, and Turkish. The .aff file encodes morphological operations—prefixes, suffixes, cross-product flags—aligned with lexicographic practices used by national academies such as the Real Academia Española and the Académie française. Dictionary entries can include part-of-speech flags and morphological codes used in corpora produced by projects like the British National Corpus and the Corpus of Contemporary American English. The format also allows encoding of locale-specific rules, character mappings, and UTF-8 normalization strategies promoted by the Unicode Consortium.

Spell checking and morphology handling

Hunspell integrates spell-checking with morphological analysis to correctly handle inflectional paradigms found in languages such as German, Russian, and Polish. Its algorithmic foundations relate to computational morphology research by groups at University of Groningen and University of Stuttgart, enabling decomposition and generation of word forms. Suggestion generation leverages edit-distance heuristics and phonetic matching similar to methods used in Soundex-based systems and in engines developed by Google and Microsoft Research. Compound splitting, affix combination, and forbidden word lists permit handling of productive compounding as in German language and orthographic conventions enforced by institutions like the Council for German Orthography.

Integration and supported software

Hunspell is embedded in numerous desktop and server applications: LibreOffice, Apache OpenOffice, Mozilla Firefox, Thunderbird, Google Chrome, Vim, Emacs, Gedit, Kate, Evolution, Pidgin, and LyX. It is available as a backend for text editors and integrated development environments including Eclipse, NetBeans, Visual Studio Code, and IntelliJ IDEA via plugins. Server-side and search integrations exist for Solr and Elasticsearch, while packaging and distribution are handled by platforms like Debian, Fedora, Homebrew and Chocolatey.

Language support and localization

Hunspell supports hundreds of languages with dictionaries contributed by communities, language academies, and projects such as Wiktionary, LibreOffice Localization Project, and national institutes including the Norwegian Language Council and Institut d'Estudis Catalans. Notable supported tongues include English language (multiple variants), German language, French language, Spanish language, Italian language, Portuguese language, Dutch language, Polish language, Czech language, Hungarian language, Finnish language, Turkish language, Russian language, Ukrainian language, Serbian language, Croatian language, Swedish language, Norwegian language, Danish language, Icelandic language, Welsh language, Basque language, Catalan language, and many others. Localization efforts interface with translation communities such as Transifex, Zanata, and Pootle.

Development and licensing

Development is collaborative, with contributions from individual linguists, open-source maintainers, and corporate engineers associated with projects like The Document Foundation, Mozilla Corporation, and various Linux distributions. Hunspell's codebase is licensed under a combination of GPL, LGPL, and MPL-compatible terms to facilitate broad reuse across proprietary and open-source ecosystems. The project governance model mirrors community-driven open-source practices found in Apache Software Foundation-hosted projects and coordinate with standards from the Free Software Foundation and the Open Source Initiative.

Category:Spell checkers