Thanks. It is still crazy to me that this was 15 years ago since I still have nightmares about completing it.
I used the approach to design a PDP-11 Floating Point Unit commercially, but haven't really seen more Spreadheet-to-FPGA work or RISC emulation -- it is clearly my thing to do.
Microsoft just added LAMBDA to Excel but didn't copy my approach of capturing a table as the formula--Like most people if I start with a big data table to analyze, I start by spreading my formulas across a row with short intermediate results, and then test it on a few rows before applying it to the whole table. My lambda would let you select your input cells and define your output cells and capture the dataflow graph between them, with any external references as globals and then spare you the results in intermediate cells.
I spent some time trying to sell this to Wall Street folks in 2007-08. I remember presenting on this at a supercmputing conference on Wall Street exactly today 13 years ago, and the Lehman guy on the panel didn't show up because they shut down that day.
Your step-by-step approach would be vastly more user friendly than the current LAMBDA implementation (where the user needs to put the entire nested function in the Name Manager).
I used the approach to design a PDP-11 Floating Point Unit commercially, but haven't really seen more Spreadheet-to-FPGA work or RISC emulation -- it is clearly my thing to do.
Microsoft just added LAMBDA to Excel but didn't copy my approach of capturing a table as the formula--Like most people if I start with a big data table to analyze, I start by spreading my formulas across a row with short intermediate results, and then test it on a few rows before applying it to the whole table. My lambda would let you select your input cells and define your output cells and capture the dataflow graph between them, with any external references as globals and then spare you the results in intermediate cells.
I spent some time trying to sell this to Wall Street folks in 2007-08. I remember presenting on this at a supercmputing conference on Wall Street exactly today 13 years ago, and the Lehman guy on the panel didn't show up because they shut down that day.