My team wants to have the ability to add information about dictionary entries such as gender, number and part of speech. Maybe having this data will increase the accuracy of TM results. We have company-specific acronyms yet we may not generate enough text in each language for the MT to guess what the gender of the acronyms is.