Would really, really appreciate the opportunity to meet and slap the
marketing idiot who came up with the names for speech models and
vocabularies in version 4.0 - they are confusing to talk about, even
more confusing to write about and expect everyone to understand.  One person writes "I use BestMatch on my XXXX and it works great" - next person says "You can't use BestMatchIII on a XXXX"  They are talking about TWO DIFFERENT THINGS but you have to read carefully to know that. There are 50 zillion words in the English language - Dragon, could you pick a couple additional ones?  Anyway- please read the following chart - VERY CAREFULLY - and note the subtle differences and also please note THERE IS BOTH A SPEECH MODEL AND A VOCABULARY MODEL to choose from when
creating a new voice file:

SPEECH MODELS:
Standard Model              for systems with less than 48 MB RAM
        BestMatch               for systems AT LEAST 200 Hz with 48 MB RAM
        BestMatch III         for systems AT LEAST 450 Hz with 96 MB RAM
        Student                   (like Standard but adjusted for kids voices)
VOCABULARY MODELS:
        US General English Standard             less than 64 MB RAM
        US General English BestMatch            at least 64 MB RAM
        Us General English BestMatch PLUS    at least 96 MB RAM (found only in Pro
                                                                  -has more speech modeling info 
                                                                   resulting in better accuracy)

The above was taken from the NatPro 4.0 help that is available by
clicking on HELP when in the New User Wizard.  Which means that,
according to Dragon, it is possible to run Nat, version 4.0 using the
BestMatch Speech Model and the BestMatch Vocabulary on systems that are only 200/64 MB RAM.  Please also note that BestMatch Speech Model is a different thing than BestMatch III Speech Model - and both are different things than BestMatch Vocabulary or BestMatchPLUS Vocabulary.  Running Nat, with the recommended Speech and Vocabulary on a minimum requirement system, demands that you dictate only in Nat's text editor and that you turn off all the resource-eating-stupid do-do and run your system clean and mean.  I am sorry for yelling and ranting but when even those who darn well should know better are confusing the already confusing terminology issue even further, I get even thornier!
To Comment on this rant, go here

"No matter how hard I try Nat won't get above about (fill in blank here
- varies from 0 to 99)% accuracy for me - what soundboard or microphone would you recommend?"

"Nat is a (fill in blank with favorite "bad word" -sucky) piece of software - I have tried very hard to make it work - I've spent (fill in blank here - hours/days/years/lifetimes) working on it you are *((*&^&**&% idiot if you think that I'm going to shut down any of the programs that load when I turn my system on - I have to have all 27 of them - I may want to know what my horoscope is in the middle of using TimeSlips and WordPerfect and Word while surfing the net and running the scheduling program for my 260 part time employees while playing my favorite Enya tune on the CD.  For your information I use Win2000 with all the latest and greatest patches from M$ and I haven't rebooted ran ScanDisk or Defrag in the last 124 days - it is a point of great personal pride with me that I never turn my system off - it is hard on them as you would know if you knew anything.  And - Ms no-it-all - I don't have time for all that Vocabulary Manager crap and I DO correct most of the time."