Unless it only takes you an hour or two to whip up the entire network stack including the crypto I think it's pretty cost-effective to splurge on the processor that can run linux.

If they were interested in saving money for it's own sake, they wouldn't have bothered using sterling silver for the metal parts.

> A 16-bit microcontroller (< 5 USD) would pull it off, but you'd need to custom code most of the software to power that.

So it's an engineering trade-off: You "don't need" something complex, not at all, nope (but if you use the "simpler" solution, other parts become more complex as a result).

I've noticed this multiple times before, and it annoys me: "Simpler" is not always better; in fact, it isn't even always simpler!

I don't mean to jump on you in specific, but this idea does come up again and again, often in places like Hacker News.

I completely agree - except this is a hobby project, not something serious in the first place. If you're hand-tooling the metal parts for the grenade, CAD designing the parts, and machining them or 3D printing them, it would seem that novelty is the factor.

If you can save 100-150 bucks while increasing the cool factor and having a more fulfilling rewarding experience (like you said, this is Hacker News, after all), why not?

$150 does not seem worth whatever effort would be required to custom code most of the software. Especially for a one-off art piece. I think the goal is to make it functional and aesthetically pleasing, and that's what makes it a fulfilling experience for the artist.

But I still think an ARM processor running Linux is way cooler than a silly micro-controller! We should vote.

So what you're saying is that you should save the 10$ from the 15$ you pay for a Cortex and instead spend many many days to end up with a outlandishly slow solution?

I take it this is supposed to inspect network traffic, at some point you need some raw throughput.

I don't know where you got those numbers from. The savings difference (after buying the needed hardware for networking, etc. for the non-comprehensive solution) would be along the lines of a hundred dollars (if not closer to 150).