KindleStrip > an amazing script...
Partly to answer my own question above, I've just now run the recent Kindlestrip python script against my huge Mobi file, stripping out the original ePub. Holy @#$!! what a difference. I have to say this is a sweet and simple little script, and it reduced the original 14.5mb Mobi file down to under 5mb.
For those who care, that amounts to a savings of $1.50 per book if Amazon charges the 'transfer fee' based on a file uploaded to them and left at the original 'Previewer' converted size. Serious money to say the least.
On another point, which is significant to me, if anyone happened to get the original Mobi file, they'd be able to strip out the ePub and have that format as well...basically for free. Stripping it out is to the author's advantage I suppose. Amazon may strip it out, I don't know, but it made me wonder whether having all versions in there is done for other reasons. Maybe they want access to the ePub to offer that in the future.
So, I want to say a big thank you to Kovid, DiapDealer, and definitely KevinH for all the help. My vote would be to incorporate the script into Calibre somehow since it would definitely make Calibre valuable not only for what it already is, but the potential cost savings described above.
(Last thing: I noticed in testing with the new Kindle Previewer, that output looks great in the Fire preview, but apparently all other 'devices' still present the old (K7?) version, even the Kindle for iPad--no blockquotes, all right aligned, etc. That was disappointing since I might have hoped Amazon would let things look as good on the iPad as the do on the Fire...but maybe they have a different agenda.)
A big thanks you to all you guys.
btw, I'm starting a small publishing biz, for reasons more than just my own book. I think I'll write an article on what I've learned, and I'll definitely speak highly of Calibre and all of you.