WinDictator:   Dictate in Windows, get the text typed in Linux

The WinDictator application gives desktop users of the GNU/Linux operating system a convenient way to make use of speech recognition software that runs only under the Microsoft Windows operating system in their day-to-day work. With it, your dictation software (e.g., Dragon NaturallySpeaking) runs on a separate computer (virtual or real) that is running the Windows OS and networked to your Linux-based desktop.

One half of WinDictator is a server that runs on the computer you're using for dictation. The other half is a client, running on your Linux-based desktop, that establishes a TCP connection to the server and obtains the text you're dictating via the connection. The client "types" that text into whatever X window you're working with using faked X keystroke events.

WinDictator has to manage a lot of event-driven activity, including the client- server networking, generating X events with one process, and waiting for confirmation of each fake keystroke with another process. It meets that challenge by using the power of the Twisted asynchronous framework.

The application is working on my system, albeit with some rough edges that need smoothing out. You can download the unstable but probably-worth-a-try alpha release. See the download page for info on how to get the code.

---

Windows is a registered trademark of Microsoft Corporation in the United States and other countries. Linux is a registered trademark of Linus Torvalds. Dragon NaturallySpeaking is a registered trademark of Nuance Communications, Inc.

And, what the heck, WinDictator is a trademark of yours truly, Edwin A. Suominen.