I tried to do this, using DataViz, with a Palm M515 a few years ago.
The problem I ran into is that the spreadsheet needs to be loaded at your computer, but you never really know what you're going to run into until you reach the race course. At the time, all I had at my disposal was a desktop computer.
You don't know whether it makes more sense to measure the course backward or forward. You don't know whether there's going to be a turnaround point or not.
I suspect you could keep adding spreadsheets to account for various scenarios, but it just seemed that the whole thing was of marginal value. Another issue with the Palm was remembering to turn it off after referring to it, as the screen used the battery pretty seriously, and if I remember correctly it ran it all the way down over a long course.
I did use it on a validation ride of the Chicago Marathon course, and it was quite slick how it calculated the split points in a flash once I put in the calibration data and the starting counts. So there's no question I see the value in a robust solution. I just haven't been able to make it work.
Maybe it's time to revisit it with a notebook computer...