The *TALŌ speller/correctors are conceptually different from other ‘spellers’ They do not spell check single words, but handle a text buffer such as a sentence and search for mistakes in multiple and single words. The morphological principles applied to their mechanism stem from our vast experience with hyphenators. Reliability is ensured by the principle of fuzzy distance: an alternative is presented that really resembles the misspelled word, including reversals (e.g., teh instead of the) and missing (initial) letters. Each European language has a very distinct character of its own. Therefore, each speller has its own language model, describing its characteristic behaviour, guaranteeing its speed. The most recent spellers go beyond the word boundary leaving the fundamental principles unchanged (e.g., the error an house, see also the section Languages Continue ...).
The spellers follow the rules and exceptions of spelling governed by leading academies. Abbreviations, punctuation marks, fixed formats like dates and numbers (financial data), and Internet addresses are being controlled during the spelling process. The speller has a learning capability but forgets infrequent, unimportant words. Spelling reforms, from old to new and vice versa, can be executed automatically. Spelling changes (for example between leading academies or between American and British English) are an option too. You can create and implement your favourite style for communication with the external world.
Try a search with wild cards in the large lexicons for interesting and surprising information!
Although the spellers have been designed and built for expert users who value their own style or are working in a pre-press environment, they are extremely easy to use.
The *TALŌ™ speller/correctors are also useful in an OCR environment, where individual letters of correct texts are often mutilated (for instance replaced by numbers or punctuation marks or by blanks).
The existing user interface has the usual speller functions but can be adapted to your own environment and requirements on request.
We have spellers for financial, scientific, industrial or graphic
arts usage and they can help in many other business categories as well.
You can count on *TALŌ™
The automatic learning capabilities and automation of text correction contribute
to a smooth progress of spellcheck sessions, principles normally absent in spellcheck tools.
So you can count on *TALŌ's speller/correctors to save you time and money.
*TALŌ speller/correctors have been implemented as a Unicode shared library
available for different platforms (Windows x86 & x64,
Mac OS X (source version), Linux Intel x86 & x64
(view Linux wxWidgets C++ example) , see also our download menu.
For some platforms a source package of the library will be required. Please send requests for a Linux XML demo version to email@example.com.
The library (current version 18.104.22.168/5, January/July 2023), now supporting more then 96 language modules, has become a tool that focuses on spellchecking crossing the word boundary. Today Asian and African modules are included in our system too.
visit download page | view *TALŌ's Language Book, ed. 2023 (PDF 5.2MB) | view erroneous assumptions on spellchecking