So I don't want to rag on something that had a huge amount of effort put into it, but... (and these buts apply to other systems as well as the C.H.I.P.)

1. The Allwinner chipset was a bad choice. It's non-free, flaky, underpowered (performs worse than a Pi Zero in my non-scientific tests so far).

2. The use of full fat Debian may contribute to 1. I think the AllWinner SBC is more suited to Android phones than general purpose SBC use, and using full fat Debian as a primary OS is potentially throwing away resources on things like a full fat libc. Yes there's Angstrom Linux as an option, but Angstrom's not really well maintained. Again, not limited to CHIP but definitely including it.

3. The PocketCHIP as interesting as it is, is a ballache compared to using a Pi Zero. For a Pi Zero I can install retropie and be up and running in under 10 mins. For PocketCHIP I have to hand configure each emulator, not all the debian packages work (weird segfaults and errors ahoy) and so on. Again, the PocketCHIP is not the only thing that experiences this, but once you've played Pico-8 for a bit, you start to realise that this is the only game in town for this platform unless you're willing to expend serious effort on work already done elsewhere.

4. The flash mechanism. Good god, where do I begin. Firmware flash updates require Chrome to be installed. That's right, the flash updater is a Chrome extension - a web browser extension. This is a ludicrous choice. A web browser is for browsing the web, not flashing firmware onto devices. There is an alternative route buried deep in the docs, but this is pretty convoluted. To flash the firmware in order to get hardware accelerated CPU support, I had to install chrome on my mac, fail several times thanks to the GUI interface clearing errors after they appear, try again with different combinations of USB hub, swear at the damned thing, then repeat the entire process on Windows.

5. I've alluded to this elsewhere, but the Debian build is broken in places. Various packages break because they don't take the architecture into account. Again, this isn't limited to the CHIP, but includes it.

The bottom line is that a $9 computer has a lot of competition depending on target use. Against a Pi Zero it loses to the incredible Pi Ecosystem. Against a Beaglebone it loses to a completely open design. Against a ZSun Wifi reader or plenty of cheap WRT devices, it loses on performance because of the full fat libc and the tightly specialised nature of the underlying distro.

At best (and without some serious work) the CHIP is destined to be a footnote amongst closed and partially broken Pi clones. At worst the PocketCHIP will be forgotten about in a couple of years. It's a shame, from a hardware perspective PocketCHIP has a lot of potential, being pretty hackable. Sure, the keyboard is terrible, the screen's pretty crappy and the peeling plastic is a nightmare, but it's got huge potential. I just don't see the resources there to bring it up to the same level as the Pi ecosystem.

Again, this isn't just limited to the CHIP - we'll see the same thing with the Onion Omega2, we've seen this with the Pine64 and various Pi type devices.

The fact that this post is about a port of a Pi-specific device sadly says more about CHIP's future than anything else.