Chaotic Java

I’ve been watching an interesting video on Google’s Tech Talks about Dasher – A new method for the input of text. To give a short review on what it is, it would be a very neat way to input text using navigation (using virtually any input device, including one-dimensional ones, as the speaker demonstrates). Dasher learns from given input texts such as collections of books and your own writing to offer the best prediction of what the next letters you might want to write.

I’ve been actually practicing myself with it, but not being in the greater target market of the application (named by the speaker as Japanese/Chineese/Korean language writer, disabled in a way that prevents me from using a keyboard or testing this writing method on a portable device), it is much slower than actually using my trusty ol’ keyboard.

Two things have grabbed my attention though, one that he specifically speaks of and one that he doesn’t and might be a figment of my imagination:

1. Speech-Dasher: Instead of speaking to your computer and then correct the mistakes it has done using voice commands, you speak to the speech-dasher and then use the dasher navigation which, after being optimized using your own speech and some internal magic, is extremely magically fast. This would be great for writing short columns or blogs, for example.
2. Code-Dasher: If Dasher can learn from your own input while its working, why not have it a bit more contextual? Code-dasher could show syntax elements such as if, else, for etc., and even variables, parameters and class names when they are contextually appropriate!

I’m actually interested in opinions about this, so if you have the spare 50 minutes, please watch the video and write your comment!