Spelling errors may make your story difficult to read and comprehend, and even alter its meaning. In any case they undermine the quality impression. If a newspaper, for example, is sloppy with its spelling, you may start to question its journalistic quality as well. A spelling checker cannot guarantee that you write excellent content, but it can help you avoid typos and misspellings that draws the reader's attention away from what you actually want to communicate.
Lingsoft's language checkers are able to recognise words in their inflected forms in a snap, as well as compound words that didn't exist yesterday. Why wouldn't you join Microsoft, Infomaker, Anygraaf and others, and take advantage of Lingsoft's unique language technology.
A sound balance between recall and precision
We have paid special attention on avoiding a common spelling checker plague called masking, meaning that a misspelling of a common word coincides with a correct spelling of a less common word. Inflating a spelling dictionary with a gigantic amount of words in an uncontrolled manner most certainly causes masking. A sound balance between recall and precision is one of the most important design objectives for a good spelling checker.
Lingsoft has delivered its core language proofing tools to several generations of Microsoft Office for more than a decade. There could not be any better refinement opportunity for a well-balanced spelling checker than the rigorous testing routines required, combined with the response from millions of Microsoft Office users. When you apply Lingsoft's spelling modules in your application, or when you choose to use our Adobe plugin or Quark extension, you get a mature and well-maintained tool because of this. But there's more to the uniqueness of Lingsoft's spelling checkers.
Not just simple word lists
Each Lingsoft spelling checker is based on a comprehensive model of the inflectional, derivational and compositional morphology of the according language. The inflectional morphology provides recognition of the correct inflected word forms. The derivational and compositional mechanisms allow for new words to be formed based on words known to the model. These generative mechanisms have been restricted to increase precision and avoid masking, meaning that not all morphologically acceptable inflected, compound or derivative words are recognized and accepted.
The language models cover the central vocabulary, including abbreviations, acronyms, proper names and numerals of the according language. Such a model does not contain "words" as such, but lexical entries called lexemes. These lexemes can be thought of as building blocks for actual words, the building being steered by morphological rules. The amount of lexemes for instance in the Swedish model is some 250,000, but this number should not be compared to a simple word list in a traditional spelling checker. In reality our Swedish speller can recognize millions, if not billions of different word forms, and recognize totally new compound words formed by combining existing words.
The basic suggestion mechanism in Lingsoft's spelling checkers suggests all recognized words with the editing distance of one (one-letter addition/deletion/transposition, except for the first letter of the word). More wide-ranging and more specific suggestions are given based on common spelling errors.
The checker module generally avoids suggesting words that may seem awkward or incomprehensible for the user. In particular generated compounds and derivatives are only suggested based on segment-specific correction rules. We also endeavor not to suggest words that may potentially seem offensive for the user. If no suitable found, no suggestions are given. Particularly with common spelling errors, only the typically appropriate corrections may be given.
Compact, fast and flexible
The lexical content and the two-level rules of the language model are compiled to an extremely compact and fast finite state transducer, which along with additional data is included in the spelling module file. The typical size of that binary file is only some megabytes, which makes it an easy load both for workstation and server implementations. Yet the module delivers amazing speed, which enable demanding server-based spelling checker implementations even for large user groups.
Lingsoft has put much effort in making it easy to integrate the spelling modules with LSPROOF, a common programming library for Lingsoft's spelling and grammar checkers, available for Windows, Linux, Mac OS X, and more. Java and .NET wrappers are also available e.g. for spelling and grammar checking in web applications. LSPROOF supports Unicode character encoding.
Product specifications | How to buy