Every once in a while you run across an open source project that has the potential to make a lot of lives easier - without making anyone richer. Open source is a good fit for this sort of altruistic project: giving stuff away and reaching out in practical ways just seem to go well together. One such project is WebAnywhere, a new take on the speech-enabled (and thus accessible to people with limited or no sight) browser.
There are, of course, plenty of screen reader applications out there, including open-source projects like ZipSpeak and oralux. These are fine when you're setting up software for your own computer, but what do you do if you're blind and want to surf the web from some random public computer or your friend's house? It's not always possible to install your own screen-reading software.
That's where WebAnywhere comes in. Written in a combination of JavaScript and PHP (and with the source code available under a BSD license), it works as a web site that proxies requests to other sites. The site being browsed to both appears on the screen and is delivered as a series of small (100KB) sound files, which can be handled by a wide variety of client-side applications. The result is somewhat similar to Google translations, with the translation in this case being text-to-speech.
Although the WebAnywhere code is open source, it's currently not running on an open server (the project page at the University of Washington has details on requesting access to their alpha release). I wonder whether anyone in the open source community will step forward to host a public WebAnywhere server?