Show HN: Geomatic – A command-driven geometry studio enabled with autodiff
www.tinyvolt.comAll commands have the format `output = \func inputs` or just `\function inputs`. Points and scalars are built on the fly. Eg `\line a b` to an empty canvas creates points `a` and `b`, and joins them with a line.
One can use broadcasting semantics similar to NumPy and PyTorch in a visual setting (imagine creating a list of circles where one dim corresponds to radius and another to the center). One can also use backpropagation, run gradient descent or visualize vector fields. Almost everything is reactive so changing a variable updates all of the downstream geometry. It also allows anyone to write and load their own visualization, which can be broadcasted and differentiated through.
I liked this one, but it took me a while to understand, I think this page is a much better intro: https://www.tinyvolt.com/geomatic/examples/getting-started
We've added that link to the toptext too. Thanks!
Bug: https://www.tinyvolt.com/geomatic/examples/broadcasting-in-g... Click "Draw a single n-star" Results in a message popping up "unknown command: n-star"
You have to load the extension, the instructions are in the text above the button.
Clean implementation. One thing I always look for: how does this degrade when things go wrong? Good error handling is what separates weekend projects from tools people actually use.
As an adult, I now write clean error handling first thing.
The person it benefits the most is the author, when they are building it and the errors-per-use are as high as they’ll (hopefully) ever be.
- When I have an example open, I can't type any commands.
- When I open an example, I expected to actually... see an example. I'm not gonna read the wall of text. I don't even understand what this is yet, that's why I tried to see an example.
The example was great for someone like me: I'd rather read through and follow excellent instructions by clicking through commands to understand the value of this command-driven tool.
If you want to look through a video or something, then this tool is likely not for you! (nothing bad with that, that's why it's great that there are so many different tools we can choose from)
Pretty cool. Curious, why a one time payment? Why not, say, a smaller monthly payment?
Not OP, but for me personally I’m tired of subscriptions. I’m grabbing the one time payment before OP changes his mind.
Not a fan of subscription hell myself, I plan to use one-time payment for all my products. Implementation wise one time payment is much simpler than setting up smaller payments that cap at a fixed amount.
Cool idea but not very mobile friendly it seems.
What is autodiff?
Automatic differentiation. For any DAG with a scalar output, it allows calculating its partial derivative wrt the input parameters.
I don't understand what the use automatic differentiation for in this context.
Some use cases and examples are provided here: https://www.tinyvolt.com/blog/bringing-autodiff-to-geometry