1 & 2) Do you know whether continuous command recognition is even technically possible via the official dragon SDK (and therefore also KnowBrainer)?

3) Well, Vocola as well as Dragonfly already work with WSR. The problem is that WSR isn't that great. I'm not aware of any speech recognition solution that's on par with dragon.

It might be possible to split it down the middle. You might think about the possible advantage of using Vocola in WSR for commands only. That is, after all, Vocola's purpose. If you could limit WSR to Command Mode, you could use Dragon for dictation; both speech engines simultaneously. We don't know if this is even possible but thought we would throw it out there. As to the other question…

Even if we could utilize Natlink/Vocola in KnowBrainer, we would be violating existing Nuance agreements. Having said that, Nuance wouldn't likely mind because Vocola can only improve Dragon. We only mention this because Natlink is utilizing Dragon code while other third-party command utilities, such as VoiceComputer and KnowBrainer, don't borrow Dragon code and only hook into Dragon, which is legal. When Nuance eventually stops selling dated versions of Dragon (Home Edition 13 and Premium 13), they may no longer care because beginning with Ver. 14, all versions of Dragon include the ability to create Advanced-Scripting commands. In other words, end-users will eventually stop purchasing Premium 13 for $99 so that they can pseudo-equivalent Professional 13 by adding Vocola. We suspect that Nuance looks at all third-party command utilities as potentially making Dragon better but we wonder if they will feel the same way about these utilities supporting other speech engines. Vocola and VoiceComputer already support WSR but we suspect WSR is not really a Dragon threat. However, the Google speech engine and Cortana seem to be getting closer.

Forum Mission Statement
Trial Downloads
Dragon/Sales@KnowBrainer.com 
(615) 884-4558 ext 1

I never used Vocola to create a Pseudo-equivalent to professional. I use it because it's an order of magnitude better than the pseudo-basic they shipped as an extension toolkit. I would gladly use the built-in extension tool if I had the ability to:

use the first row and column of the CSV file to dynamically create a grammar and the values within the CSV file to create the results returned by the grammar

Invoke an RPC call (xmlrpc/soap/pyrpc)

Query mongodb

Get information from an Emacs buffer, process it through a local filter (which uses a database) and then put the result back in the buffer

Have extension grammars independent of your user/speech model (helps survive upgrades/moments of frustration were you delete your user and start training a new one and only then realize you lost all of your macros)

And last, write extensions using NaturallySpeaking itself without a whole lot of editor support.

As you can see, these are things I used Vocola for and quite frankly nuance's macroenvironment is "inadequate". 

Google recognition is quite good. They've just added some editing by speech commands but I don't know if they have the ability to add more commands. The Google recognition engine is so good, my mom, at age 82 started using speech recognition because her hands hurt too much to type. Since she's running Linux (started learning it at age 80) she could not use NaturallySpeaking so I had her try Google and it works really well for her. 

The problem is as much as NaturallySpeaking plus vocola is a bag of parts, building a new environment out of Google speech recognition is an even bigger bag of parts.