Finally, well-designed modules will let you extend them beyond their initial use case to fit your specific workflows or data formats. For instance, the Arrow format has already been extended to represent PyTorch data frame (with #TorchArrow) and geospatial data with #GeoArrow.

That's it for today! To read the full chapter of the #ComposableCodex "Standards over Silos", it's at https://voltrondata.com/codex/standards-over-silos

I will continue soon with the next chapter on UIs for data systems.

🧵 17/...