Slashdot Mirror

You now have to type them into a computer to actually use the search results - so it's slower, more expensive and worse than typing the query.

Actually I would presume that Yahoo! is building some sort of VoiceXML application.

The list of items you've provided is pretty big; however a lot of these are built-in to Opera. In addition, you get some very interesting capabilities like the ones I outline below, including a tight js debugger, and voiceXML support in websites.

First off, I worked for the group that partnered with Opera to create this, but the Voice plugin for Opera is also free!

Check http://www.opera.com/voice for information about the XHTML + Voice standard that is implemented here. X+V is an open standard - go read the spec for information, and Opera's voice development page for code introductions.
For those that downloaded the new 8.x build, go to the Advanced tab in your Preferences and tick the enable box for the "Voice" option. This will download a 10 megabyte add-on that allows voice interactions with Opera. In addition, you can control websites that support X+V! See some basic examples.

More information from a technical perspective can be found at IBM's Multimodal Software Group.

Also, see the WebDevToolbar for an INCREDIBLY handy toolbar for web developers debugging complicated interactive web apps. You get trace features for your javascript, the ability to inspect the DOM for a given page and many other introspection features. The handiest feature is a javascript shell for tight debugging of applications inside the browser.

~ Mike

Designing interfaces for visually impaired users was a topic of my own research for a couple of years at ISEF and affiliated fairs while I was in high school. As far as interfaces go, I experimented with a braile keyboard (tried and true) and voice recognition (ammount of success varies depending on possible input vocabulary). One tool I found particularly useful for designing interactive voice interfaces using both speech synthesis and recognition was VoiceXML, which defines a markup language with basic scripting logic for quickly building voice interfaces. During my junior year of High School, I implemented a perl module for automagically generating voice interfaces (my ultimate goal was to create a replacement for the integrated cgi modules for live internet-aware voice apps), and showed it off by designing a basic newsreader app that pulled data from NewsBlaster (this was back in the day before Google News).

My research experiences taught me several things. Firstly, it is important to offer auditory feedback for blind users. With voice recognition on a limited vocabulary, this isn't really a problem, as the user always knows what they said. With a broader range of input vocabulary, or with keyboard input, it's important to verify what has been entered at least every sentence or so, as there is naturally no way to provide visual feedback. Secondly, you must realize that all data being transmitted to the user is necessarily in a totally linear format. That means that, in any interface you are designing fresh, you should keep the interface as slim as possible. When you are reading out information that was originally intended for sighted users, some sort of adaptive content filtering is a must. If a blind user goes to slashdot, chances are "image, image, image, image, image, image, username, preferences, subscribe, journal, etc..." is not the first thing they want to hear.

With VoiceXML and ENUM every POTS device becomes an internet access device. Does this means that every mail order retailer that currently collects sales taxes (due to local point-of-presense sales tax laws) can stop collecting those taxes?

I suspect that the senate found it rather hard to create a clear demarcation between commerce based on "internet access" versus commerce based on traditional, taxed categories of custmer interactions.

Regarding the technical issues - I'm thinking of something like VNC and Voice-over-IP.

And the w3c didn't *create* this standard. The working group started off as the VoiceXML Forum, including Lucent, Motorola, IBM and AT&T, who were interested in standardizing the API used to create voice site deployments (which was already at that time big business).