Given I have recently been casting covetous glances in the direction of Dragon Speaking, that is a very good question.
As far as I know, SIRI is almost entirely a server based system where both the command text voice recognition is performed at a server farm and the resulting "commands" transmitted back to the iPhone over the internet. So I would surmise that at the moment, one barrier to porting SIRI to OS X would be the load placed on the SIRI servers by the new iPhone 4s. Apple does not want to do anything to degrade the performance of the iPhone with its much bigger market than that of Mac computers.
Once Apple has a good handle on that then they will be in a better position to gauge what additional load would be created by adding in all the Mac computers — with lots more command and action options — and how more server farm capacity would be required to support the additional load without degrading the performance on the iPhones. Then, and only then, will they port SIRI to the Mac. Integration of voice recognition into OS X could well be a big enough feature to serve as the lynchpin of OS X 10.8.
So far SIRI on my iPhone 4s works well provided:
- There is a low level of environmental noise
- Either there is no one else around me or I don't care if I annoy the h*** out of anyone and everyone around me as I speak commands.
- I have at least a decent WiFi or other data connection
- The network is not too heavily loaded
- My voice is not currently hoarse from a cold or allergies.
FWIW, I find SIRI
very convenient when I can use the Bluetooth connection in my car (I wish my car had a better microphone) and I don't have a car load of passengers. But I choose not to use it in public places for privacy reasons as well as avoiding public censure from those around me. SIRI is decently accurate where small chunks of data are being used, but as far as I am concerned the jury is still out on whether it is accurate enough for large amounts of data input. Even with short commands such as
Call home I have a fair share are
Cancels due to recognition errors, but that could easily be a result of my speech patterns and rhythms or my Texas drawl.
