A key feature about data-flow programming that seems too often missed is that it is (or can be) hierarchical.
Define a subgraph of atomic nodes as itself a node with its ports formed from as-yet unconnected ports of its atomic constituents. Compose yet higher subgraphs of subgraphs and atomic nodes. Package all this in some way.
This is directly analogous to syntactic programming where functions aggregate other function calls and all that packaged into a library with an API.
This is amazing work. I've used GPIOZero in my RPi projects before. In the future, you can introduce conditional operators in the flow. It can become a complete coding platform.
I don't think this is for people that would just write code. Maybe you need that pressure release to get some edge cases, but for the most part, if you're using a visual programming tool, you are trying to avoid the code part for some reason.
So, I don't see that limit as something that's all that bad.
It's a good onramp to building a programmatic intuition, but you're correct that it very quickly gets left behind once the 'magic' is understood. If you do a quick image search for , e.g. Rhino3D Grasshopper, Blender Geometry Nodes, Vectorworks Marionette, Autodesk Dynamo, ComfyUI, you'll quickly see the implementations of the real world procedures users are crafting become unwieldy rather fast.
I can't tell you why an 80 column wall of characters is easier to reason about than a visual graph, but I assume it has something to do with the reason Mitch Hedberg does not like arrows[0]; look, a function.. let's got that way.
I had a similar thought while building a visual workout programming system for bodybuilding.
One thing I keep running into is that workout programs are essentially state machines progression rules, deloads, conditional branches, different exercise substitutions, autoregulation etc...
That made me wonder whether a visual programming model would make these systems more approachable. I'm still unconvinced, though. Unlike automation workflows where the graph itself reveals the logic, I'm not sure what the right visual abstraction is for training programs without making them even more overwhelming.
I think this looks like a great idea for the kind of automation people often use GPIOs for. I wonder what the complexity would be to make gpiozero that is compatible with Pi Pico 2 or ESP32 devices? A full Linux OS can be overkill for just controlling some GPIOs. That said, a Pi Zero 2 W is only marginally larger than a Pico 2.
i think there are two patterns of programming: control flow, and data flow.
this visual programming tool is actually a data flow approach.
the good part is, it will be easier to express simple logic. as a tradeoff, it will be awkward to express complex logics
15 comments
[ 0.25 ms ] story [ 7.6 ms ] threadDefine a subgraph of atomic nodes as itself a node with its ports formed from as-yet unconnected ports of its atomic constituents. Compose yet higher subgraphs of subgraphs and atomic nodes. Package all this in some way.
This is directly analogous to syntactic programming where functions aggregate other function calls and all that packaged into a library with an API.
See JSONLogic UI for similar implementation: https://github.com/GoPlasmatic/datalogic-rs Disclaimer: I am the maintainer of the datalogic-rs project
https://github.com/flux-doctrine/awesome-fbp
Are there any exceptions?
So, I don't see that limit as something that's all that bad.
(maybe more context would help)
I can't tell you why an 80 column wall of characters is easier to reason about than a visual graph, but I assume it has something to do with the reason Mitch Hedberg does not like arrows[0]; look, a function.. let's got that way.
[0] https://youtu.be/EI1DBRz3JLk?si=DCLNkbeCGWyw-Ceu
One thing I keep running into is that workout programs are essentially state machines progression rules, deloads, conditional branches, different exercise substitutions, autoregulation etc...
That made me wonder whether a visual programming model would make these systems more approachable. I'm still unconvinced, though. Unlike automation workflows where the graph itself reveals the logic, I'm not sure what the right visual abstraction is for training programs without making them even more overwhelming.
So far i've a workout programming language and bunch of sample programs and an app which can help you execute these workout plans, all free ofc: https://symbiote-studio.macrocodex.app/?builtin=gzclp