I wonder how the computation time compares (in the article, he only talks about optimisation time, and that’s not really fair as Excel’s optimiser is more of a toy, while the DAG computation engine is pretty good, I think).
I know an investment bank that had tools to transpile Excel sheets (with specified input/output named ranges) into a nice DAG in C, and expose it as a function, so they could run Excel models on a compute farm and prevent manipulation of the formulae. (Yes, that means they had to reimplement all Excel functions.)
That was more than a decade ago, wonder whether they still use it.
It would be nice to run this on the following to see what comes out.
https://github.com/amzn/computer-vision-basics-in-microsoft-...
https://news.ycombinator.com/item?id=22357374
I know an investment bank that had tools to transpile Excel sheets (with specified input/output named ranges) into a nice DAG in C, and expose it as a function, so they could run Excel models on a compute farm and prevent manipulation of the formulae. (Yes, that means they had to reimplement all Excel functions.)
That was more than a decade ago, wonder whether they still use it.