This morning there were a few blogs I caught via Technorati on the fact that Vista's Windows Speech Recognition (WSR) requires a minimum of 1024x768 screen resolution to run. First Craig blogged about it on his blog, and then Rob picked it up on gottabemobile.com. Here's what they had to say:
The speech reco team built the interface with a screen that cannot be resized or minimized, and uses a lot of white space in the screens. And it displays the above error when you try to run it with an LS800 running 800 x 600, or any other device with smaller resolution than 1024 x 768.
I guess the Speech Reco team didn't attend the Mobile Partners conference where Microsoft taught on building applications that are mobile friendly: screen real estate issues, etc.
NOTE: BTW, Rob, I actually worked on the Tablet PC team for v1 of Windows XP Tablet PC Edition on some of the speech integration as well as the inking platform components so I'm well aware of how to optimize UI for Tablets and mobile environments. I understand how a lot of people aren't aware of the guidelines for building applications that are mobile friendly, and thus why you assumed we might not be familiar with the guidelines.
Don’t you think that perhaps small screen devices are exactly the kind of device that would benefit from natural language input like speech?
As I said I have run through this wizard before. I’ve seen the interface and there is a stack of white space in it. But don’t take my word for it – I dropped the resolution down to 1024x768 on my Toshiba M400 and kicked off the tutorial. It runs in a full screen window that you can’t resize. Here it is.
I understand the feedback ... and it was a painful decision to have to make ... but we had to go up to 1024x768 minimum resolution for a bunch of the screens in the Tutorial that Craig didn't show you. We tried doing that in 800x600, but it just didn't work; especially in the "Dictation" and "Working with Windows" sections of the Tutorial.
You see, in the tutorial, we try to make the system look as much like an actual running Windows Vista PC as possible, and also give the user instructions off to the right. We ask them to interactively try things (in a directed manner) so when they're done with the tutorial, they'll have a great idea of how to use WSR in the real world.
We may revisit this decision in the future (for Vista + 1), but for now, you'll have to run the Tutorial on a system that supports 1024x768. However, users can still use the old style Training window on lower resolution screens.
Windows Speech Recognition itself doesn't have a minimum resolution that's any different than the core OS itself. You can run it on any size desktop you want.