Valid points Xymox, but then you have to ask what is the actual argument here? Is it "true hardware" only (I'm talking about ttl or cmos logic chips or even discrete components only), versus a mix (which is what Crestron and the one I mentioned are) or software/firmware/hardware. In simplest terms, if I want to make an LED blink; I can used a resistor and capacitor, or I can use a 555 timer and a transistor or two, or I can use a PIC and some assembly, a BASIC stamp and supporting circuitry, a purely mechanical timer, use an ebay-crestron, or tie the LED to an I/O port on a PC and write some code in my favorite language to make it blink. I could even tie it to a diode and current transformer and put it near Upstatemikes’s driveway lights.
Obviously, the final design choice depends on the devices application – the right tool for the job.
You tout Crestron as being hardware-only but make no mistake, without software; it would be a very expensive paperweight and nothing more. It is more than just a grouping of hardwired logic gates and I/O drivers, it too is loaded with software. Yes, you are correct in that Crestron gear, like the devices I mentioned are purposed-built for a particular task, and that adds an intrinsic amount of reliability to the mix. Crestron “hardware†is of no better quality (on a electronic component level) than high-quality PC components – the same part from the same manufactures are on both boards. Yes, they’ve a single business model that allows them to concentrate their efforts on doing one task and doing it well, but that specialization comes at a price. A price to the consumers in the bottom line, and a price to Crestron as it tend to make them inflexible and more resistant to change. The exact argument was made for big-iron mainframes when mid-range Unix boxes came into picture. When the last time you saw a mainframe running a company?
Yes complexity adds additional points-of-failure, but also adds opportunities for innovation. Yes, a simple, true hardware-only system will always perform better (although I just had to buy a new toaster) for single tasks than one comprised of hardware/firmware/software, it will also always only do one thing. People here choose PCs for HA because of availability, supportability, and reliability (sorry, true). PCs also give us the ability to innovate and not to feel constrained by a particular set of logic that may not always apply to our situation or meet our needs. Generally I think you’ll find that people here also do not like to be told that they cannot do something, I know I don’t like it. Crestron (and others) seems to be known for telling people “no†to all sorts of questions and that directly opposes the stance of most HA DIYers, as you have stated. PCs in their generality give us the option to explore, they don’t always say “yesâ€, but they never say “you cant do that†and almost always say “maybe, how bad do you want it?†and that is enough for most of us. Crestron will always have its place, as long as there are people with more money than time, they’ll survive. Me, I’m on the other end of the spectrum – life gets in the way of my hobbies – That’s why my house runs on linux, misterhouse, Xlobby, and zoneminder, all free, on a few yard-sale Pentium III PCs. The only time they have been down is due to extended power outages.
Soooo, to sum: If you are talking “true†hardware vs a mix (HW/FW/SW). Then I agree; hardware is more reliable and less flexible. If you’re comparing a mix vs PC tech, then your back on the Ford vs Chevy thing and you’ll never win that (neither side will).
Fly by wire is a great example of hardware based engineering. The W2K box does not directly control critical systems for a reason.
Ever been to NASA?