mdsc (multi-dictionary spell checker) is a library for checking the spelling of individual words, using more than one dictionary. It is designed, initially, to be a component in an identifier name convention checking system that is a product of the doctoral research of the author, Simon Butler. See http://www.floss.nu/research for more details.
mdsc allows the caller to spell check individual words with a set of dictionaries and to receive information about the dictionary in which the word is found or that provides alternative spellings.
mdsc is not designed to be used for spell checking documents. If you are looking for a library to spell check documents and blocks of text in a Java application (GUI or otherwise) then Jazzy (http://jazzy.sourceforge.net/) may meet your requirements.
v0.2.1 of mdsc introduced an API to access the wordlists for use in other applications.
Releases of mdsc are available from maven central. For gradle users just add 'uk.org.facetus:mdsc:0.2.4' as a dependency to your build file. Maven users should add the following to the pom.
<dependency>
<groupId>uk.org.facetus</groupId>
<artifactId>mdsc</artifactId>
<version>0.2.4</version>
</dependency>
The core spell checking functionality of mdsc is based on the Jazzy spell checker code base (http://jazzy.sourceforge.net/), in particular the com.swabunga.spell.engine package. A list of the developers of Jazzy can be found in the file docs/JazzyAuthors.txt
mdsc was initially developed during Simon Butler's doctoral research and consequently The Open University owns the copyright of work on mdsc prior to and including the commit https://github.com/sjbutler/mdsc/commit/2d4320145488219ebf2015311c00a641d5a12647
The copyright of revisions made in the commit https://github.com/sjbutler/mdsc/commit/2ff0d6c862af7ea977a9101841acf53f1d0f5864 and since belongs to Simon Butler.
mdsc is released under the terms of the GNU Public Licence (GPL) v3 with the 'classpath' exception. See the files COPYING and LICENSE for details.
mdsc requires a Java 8 JRE to run. It has not been tested with Java 9 or greater.
SLF4J is used for logging. The slf4j-api-1.7.x.jar file should be on the classpath along with an appropriate SJF4J jar for the application's logging system.
mdsc supports two types of dictionaries, and is, as far as is known, independent of natural language, so long as the characters can be encoded in UTF-8. The dictionaries must be in the form of a word list (i.e. a plain text file with one word per line).
A default set of dictionaries derived from the SCOWL word lists (http://wordlist.aspell.net/) and Simon Butler's research are provided. Temporarily, we include lists of abbreviations derived from Emily Hill's AMAP project (currently available from http://users.drew.edu/ehill1/AMAP.tar.gz). mdsc is also designed so that new or alternative dictionaries can be added through the API and used in isolation or groups.
Since v0.2.4 a German word list has also been included. The JavaDocs explain how to use the dictionary.
The DefaultDictionaryManager is a convenience class that provides the default set of dictionaries. The AbbreviationDictionaryManager provides a set of abbreviation dictionaries. The class should only be used to identify abbreviations, not to try to correct them.
mdsc uses SLF4J (http://www.slf4j.org/) for logging. Logging is minimal and intended to report initialisation errors.