Announcing project Wordtip

Today I made the first release of Wordtip. It’s very alpha, but it aims to be a text classification service over D-Bus. What’s that good for? Well, it will allow applications to potentially be a little bit smart and do some tasks instead of the user – in this case automatically sort anything that contains plain text (notes, documents, blog entries, emails…) into categories, for recommendation or organizing purposes.

It is a bit amazing that only a few kilobytes of C++ code implementing a naive Bayes classifier and a little bit of D-Bus glue, with no stemming yet, can actually provide reasonable results. Hopefully I’ll be making frequent snapshot releases. I actually started the whole project as a little bit of foreplay before diving into the work for my master’s thesis, which will hopefully be about Support Vector Machines.