I have updated the latest stats on MineSweeper sales over the last 3 months. It includes 3 days of data from the v1.7 release and some of my thoughts on what it all might mean.
You can read about it here: http://codeatuts.blogspot.com.au/2012/11/interlude-14-update-on-minesweeper-sales.html
Thanks for this @Reefwing - intriguing to see how many downloads you got to date and the demographics, especially as there are a fair number of Minesweeper-esque games on the AppStore. I now know Minesweeper is not popular with Austrians and Malaysians! 
Thanks for the info @Reefwings, it’s always a pleasure to read your reports.
Thanks @Simeon - couldn’t have done it without Codea!
.@andymac3d - yeah what’s up with the Austrians and Malaysians?? As you say, I think the downloads aren’t too bad given it is the 50th app to come up when you search for MineSweeper in the app store. There is definitely a first mover advantage. Interestingly, Apple has featured it a number of times on the “What’s Hot” list, albeit fairly low down. More like “What’s Lukewarm”.
.@Jmv38 - thanks mate. I was going to have a play with 3D next and was looking to use your Sphere class (with appropriate attribution of course), if that was ok?
.@Reefwing fine, i’ll be honored. Do what you want with it, although i’am not quite happy with all that bunch of code (too much, too confusing, a few bugs still there, not evolutive enough, etc…).
Try also the post on mesh saving, it does less but is cleaner, simpler, and loading a 3d object with it is really simple. I think something like that should be included in Codea. I’am working on how to make a 3d project in Codea that would be for 3d an equivalent of what ‘paint’ is for 2d. so if you have suggestions let me know, maybe you could orient me. The difficult thing here is to find the good approch: how to make it simple, clear, for the user and the developper…?
In the Sphere project, i wanted a ‘procedural tool’ meaning the object would be created from a series of actions, that would be saved and re-done to re-generate the object. Then it had to be real time compatible, etc… And it became too complex, because there are so many different ways to get the final mesh data…
So i think now i was wrong from the beginning: i should just focus on saving the final mesh data only, with no memory of how it was created, a bit like in the ‘paint’ program. With ‘paint’ only the final result counts (the image), no matter the intermediate steps. This is why i restarted my program from the mesh save/load utility core. I have added the normals computations, and the shadowing too, and also multiple copies of same mesh (not published yet). Here is a short video:
http://www.youtube.com/watch?v=Dgc9wMu6ctU
Make a recorder of your motions, then re-render to reprint your work?