transluSense is a piece of software - a tool - an application. It is designed to interpret language and word relationships within language. Essentially transluSense is a Natural Language Parser (NLP), however it it ignores rules that are generally associated with language (ie: grammar). The software uses word relationships to build its knowledge base for different genres of text so that it could (a) assist with improving the coherence of texta, (b) be utilized for differentiating between language-use patterns, and potentially (c) for pattern-based word prediction.
The theory goes that if any speaker is exposed to certain language use patterns, they will be able to identify the "good" language usage from the "bad". Basically, anything that doesn't sound right would potentially be considered bad language usage. Everything else that does in fact sound right would be considered "coherent". The big question is: Can computers differentiate between good language usage and bad language usage? Some would say Never- or maybe most would say that.
(b) Domain Identification
As most native language speakers can agree, language is used differently in different domains, and what makes the difference is the vocabulary (word types used) as well as the word-usage (fancy lingo vs. colloquial jargon). So one thing this software could be useful is to automatically differentiate between one domain from another. Something that humans generally have difficulty realizing is that individual authors generally have a unique word-usage "fingerprint"; this falls in the field of forensic linguistics. Depending on the sensitivity and accuracy of the transluSense software, it could potentially be able to distinguish one author to another.
(c) Pattern-Based Word Prediction
Based on the knowledge gathered observing inter-word relationships, algorithms could be developed to suggest potential word replacements or offer suggestions for upcoming words. While monitoring user input, transluSense, (based on heavily seen word patterns) could suggest upcoming words or potentually phrases. Word patterns applied to t9 as opposed to just letters.