Hey.
I've only scan read the thread, but I'll make a few points:
1) The current CGM I've used with my pump (Medtronic Paradigm) is generally pretty rubbish. Quite inaccurate and very painful (compared to the infusion sets anyway). I'd much sooner stick with a BGM machine - particularly something as advanced as the Contour Next, which is an amazing bit of kit.
2) The 'expense' comes principally from the companies that make the pumps and CGM machines needing to make profit. I'd estimate that there is no more than about £250 worth of components (and that's assuming a lot of custom stuff!) in my insulin pump, yet they're valued at over £3500. Obviously you have to take into account the development, manufacturing costs and target audience (only 260,000 type 1 diabetics in the UK, and around 2.8 million worldwide - we make up around 0.4% of the UK population), but even then I think they could be cheaper.
3) Your friend is correct - engineering apps to talk to CGM's is easy enough if you know the communication protocol. There is a remote control available for my pump, but unfortunately I think it uses radio as opposed to Wifi/Bluetooth, so I couldn't even 'hack' a connection to it.
4) The ultimate goal for CGM would be like the watch in 'Panic Room' (the movie with Jodie Foster. Or even better, the upcoming artificial pancreases, which render the whole BGM concept moot.
I'm also of the opinion that a lot of failings of current mechanical treatments for diabetes (and indeed other conditions) are down to the fact that they're driven by people who don't necessarily know the technology they're dealing with. My doctor knows a lot about diabetes and endocrinology, but she couldn't compete with me on Maths and computer programming. I think that a combination of the two is required to get to a point where everything works together.