A question about efficiency...

[JLX]

When creating a command like: "Set Number [0..999]" I get a warning that this will create 1000 dynamic commands which may impact performance.

I get no such warning when I use: "Set Number*"

The former ([0..999]) knows to listen for a number (thus the distinct number of entries) which seems to improve recognition and also act as a form of input validation. The latter (*) has to listen for anything and while mostly accurate (with numbers) it's not uncommon to get "to" or "too" instead of "2 (two)" for example.

Of course this came up as I'm looking for reliable ways to recognise such things as: "Tune to 222 too" ;-)

Does anyone know enough about the internal workings of the speech engine to know if there is a significant difference in processing overhead? What is the best practise here?

Any insight would be appreciated. Cheers.