It seems like our customers keep on using gadgeteer for prototyping, we do internally as well! So here is a thought, why don’t we as a community rethink what would be the right move forward. GHI e-blocks from 10 years ago, GHI gadgeteer, seeed grove, sparkfun qwiic, adafruit stemma, dfrobot gravity, makecode jacdak, mikroe click… this is fragmented and depressing! How do we bring it all together?
What is important to you? Pick one or more…pick all!
Unified board size
Discoverable at runtime
Connect through embedded high-level protocol interface (no need to read datasheet)
Connect to sensor directly (need to read datasheet)
Connect in a daisy chain
Connect through a hub
The first reply to this post a a lit of potential modules…
My answer would be very different depending on whether this is targeting hobbyist projects that never proceed to commercialization vs rapid-prototyping that might proceed to commercialization.
High-level protocol interfaces and runtime discoverability make one-off projects faster, but aren’t practical for direct commercialization (due to increased BOM cost, increased current cost, decreased performance, etc.) - you’d have to re-write all your software.
So, tell me who the audience is and I will tell you what that side of my sw work cares most about.
The point of the matter is Gadgeteer or something very similar is the ultimate next step in device design and customization. It is the equivalent of high level languages or frameworks, it opens device design up to so many more people. What a modular system like Gadgeteer also enables is automation of PCB/device manufacturing where a system takes the Gadgeteer prototype and using the modules and pins used and creates an optimized PCB ready for mass production.
Gadgeteer was the best electronics educational tool that I’ve ever seen or used. In one class I could cover the how a device worked, including the interfaces (I2C, SPI, etc) and have the students construct, code, test and experiment with the device. If a device required multiple classes it was easy to store the uncompleted projects till the next class. Managing modules is pretty much infinitely easier to manage them individual components like resistors, wires etc and in all the classes I taught Gadgeteer proved to be pretty much invincible to student mistakes (ie no fried components or fried students ) In short the world needs Gadgeteer.
Gadgeteer allowed me to get into hardware, I could code my brains out, but hardware wasn’t something I had a lot of experience with (I was the idea guy behind a number of startup companies so being able to think and breath in code was essential). Gadgeteer enabled me to ‘warm’ up to hardware by using my coding skills and some easy ‘framework’ like module interfaces and build and code devices. Once I was confident with that, then I could dive deeper into the hardware and start building my own drivers. This also means that Gadgeteer is the perfect tool for teaching kids about coding and hardware at any level. I had students who didn’t really care about how SPI worked, just that it worked and enabled their code and device to work, great!! I also had students who really wanted to learn at the chip level, which Gadgeteer could also support and modules where wonderful example of how interfaces and chips worked. Combined all of this with an easy to use and full powered development tool like Visual Studio, make helping students a breeze for the teacher. As you can tell I miss Gadgeteer a lot, it was/is a wonderful tool and piece of technology.
As for a bright future, i’d suggest the following:
For fast and fun prototyping and ‘mass’ hobbyist / education market … make sure GHI has TinyCLR on SITCore hardware available as first choice MCU’s of the various ecosystems. If anyone wants to use some Qwiic modules, make them want to use TinyCLR and SITCore boards. Then repeat for all ‘prototype’ ecosystems (boards like pybd and teensys are missing this boat). This means great hardware, in various formfactors/powersupply/networking options and a huge driver repo. The extensibility of TinyCLR should make all the difference here. Currently the market is fragmented, if you want to prototype some Groove module, then you need their boards. If you want to combine a Qwiic module with a Click board to demonstrate something? GHI can fix that.
For core hardware product development, you’d need great chips, SOM’s, security, IP protection, high processing power, low power etc. Reliability, resilience … perhaps even more into the industrial side of things. Wider range of temperature, etc. Extensible development platform.
Focus 100% of development on these 2 tracks for a year, and SITCore will soon need extra pnp machines to keep up with the pace.