A clever name?
You can't get anything smaller than a Quark but you need three of them to make anything that matters.
(Coat: The one with lots of small things in the pocket.)
Having nominated the Internet of Things as key to its future strategies, Intel has added a super-cheap development board to its Quark lineup. At US$15, the Quark D2000 microcontroller development kit is Chipzilla's latest attempt to plant a flag in the cheap-as-chips breadboarding market. It features a 32 MHz low-power core, …
"You can't get anything smaller than a Quark..."
Strings (as in 'String Theory', or arguably 'String Theology'), if they exist, are many orders of magnitude smaller than quarks.
Beneath strings are, of course, recursively-supported turtles. So obviously, God programmed the Universe in Logo. Last-Thursdayism is thus constrained to post-1967.
The Arduino isn't the future of IoT. The soon to be released ESP32 is / will / should be. At least for the next year or so.
Its predecessor created a lot of buzz and even a few working projects. But if the boards based on this "Mk II" live up to expectations they should really start things moving. But as with all things IT, the success will only come if the software and manufacturer / user support is in place.
Arduino. I tried NodeMCU and the Python based equivalent but couldn't get either to work. (I think there was some sort of version incompatibility but I couldn't be bothered hunting it down.) On the other getting it to work with Arduino involved nothing more than adding a url in the settings. Running a webserver on a board the size of your fingernail is really quite impressive.
There are Uno-compatible (with an actual Atmega 328 on board) stamp-sized "Pro Mini" boards that can be had for 3-4 bucks and also work on 3.3V (down to 1.8V). This thing is nowhere near compelling enough to sway anyone - it is slightly better on most fronts, but only a little better, not enough to make a huge difference; if 16MHz is not fast enough for you, 32MHz won't be either, and if you can't fit into 2K of RAM, 8K is only marginally better. This won't be enough to overcome the massive inertia of "simply using an actual, ubiquitous Arduino"...
Er, I don't believe that if 16Mhz isn't enough, then double that won't be enough either. DOUBLE the speed? Of course it will be enough, for some projects.
Same with memory - 4 times the amount? That's a huge difference, not just marginally better. 400% better!
As to overcoming inertia - probably going to find it difficult. But not for the reasons you quote above.
I just bought a cheap Arduino UNO r3 clone from China for €3.29. The quality's not great, but they work fine. You can even get an Arduino Due clone (which has an 84MHz 32 bit arm chip in there as well as more I/O lines than you can usually use) for €12.49.
The comparison with the Intel is not fair, the Intel Quark is a 32MHz, 32-bit x86 microcontroller, so will be a considerable step up on the UNO. A better comparison is what it is like compared to the Due or better still the teensy 3.2.
Where the Arduino definitely wins (as well as the teensy to a slightly lesser extent) is the support that the community gives. I haven't seen that around the Intel.
Another point is that the clones are sent free from China, looking at the Intel, I can buy it from mouser for €13.53 but with a shipping charge of €20!
"[Arduino] clones are sent free from China, looking at the Intel, I can buy it from mouser for €13.53 but with a shipping charge of €20!"
Organizations like the Raspberry Pi folks, and here Intel, need to ship these gadgets in bulk to somewhere in Asia for distribution, where they can then be shipped anywhere on Earth for a dollar or two.
Please don't ask me why shipping in one direction is nearly free and shipping in the other direction always costs $£€20. I have no idea. It's the exact same people and vehicles, they're just facing in the other direction. This asymmetry in shipping fees only exacerbates the Balance of Trade issue. It'd be well worth looking into this mystery.
Free and can run off a watch battery.
As Intel knows only too well: it won't really be the chip or the board that decides this but the surrounding eco-system of devices. Like the Rapsberry Pi, the Arduino profits from being the first to the market in significant volumes.
As for cost: Intel cannot afford to undercut the cheapest Chinese ARM chips, which keep getting better and better.
The subtle but critical thing that the Arduino folks did was loading the 'Blink' program into the production hardware by default.
Making an LED blink is the PRINT "HELLO WORLD!" of embedded software.
Arranging to have an on-board LED blink upon simply powering up the board, and providing the source code for 'first steps' IDE experimentation, saves at least a day (or perhaps twelve) on the initial learning curve for noobies.
It's very subtle, but hugely important. Whoever thought of doing it is a subtle genius.
This post has been deleted by its author
How is Intel's board an ARM killer if it is taking aim at an AVR-based Arduino?
I also don't see what any of this has to do with IoT given neither board has any sort of networking capability as standard.
If the ARM-based C.H.I.P. delivers what is promised, then that'll be vastly better suited to IoT applications (built-in Wi-Fi and Bluetooth 4.0) and will supposedly only cost $9. Even the BBC micro:bit is more relevant, in the IoT space, than the Arduino due to its built-in Bluetooth - another ARM based board.
I don't have anything against Intel - Very happy with all my i7 machines, but Intel blew their chances in the embedded market when they sold off the XScale range - I fully expect this Quark nonsense to go the way of the i960.
I use a lot of Arduinos and I can see no reason to switch to this Intel "thingy".
The reasons are many but, for starters, the Arduino is open source and available in many shapes and sizes to suit the project requirements. There are many vendors. They are very cheap with nanos' less than $3 from China with no carriage costs. The wide power source voltage is nice too. The cross-platform Arduino programming environment, along with the vast collection of libraries, makes development fast regardless of what you are doing with it.
In fact, since Arduino, the other loser in my company is Microchip. Where a PIC chip would go in the past there now sits an Arduino Nano socket. Unless size is an issue using Arduinos saves us R&D budget.
Having left the comfort zone of the wintel project and witnessed ARM clean up in the low power sector where to go next? The Raspberry Pi and Arduino projects are far too entrenched to offer a viable alternative in my view. Even IOT will not be easy since the massive rise of the ESP8266 has already made a large dent in the market.
So, Intel, you may be trying to do a land grab but you are a bit late to the party.
I have one Arduino board... but dozens of atmega 328s that were all programmed in that board, then popped out, connected to a 16MHz resonator and a handful of relevant bits to do specific jobs. Total cost is about £6 per 'unit'. Intel aren't even in the same country, let alone ball-park.
...manufacturing the $2 SOC that'll be the high performance PC embedded right into the HDMI connector.
They'd better focus on self-driving cars, since those will need TFLOPS of processing to deal with the last 5% of the total real world driving problem space. E.g. Russia, India, Bhutan.
What a truly insane thing to produce - take a widely reviled instruction set (x86) which is only still around because of the desire for binary backwards compatibility on very large fast chips (which go to great lengths to dynamically recompile it in hardware into something that doesn't suck) and then put it on an embedded CPU where nobody has the slightest interest in binary compatibility (..not that it even is!) and in a field where there are several very long established and highly efficient alternatives (ARM, AVR, etc) that work better in every way. IN-FRICKIN-SANE. This thing is already deader than a dead dodo & absurd to the level of finger-pointing mockery.