I'm interested in Factor as a serious replacement for all the programming I used to do on my HP-48GX in RPL. (Those were the days!)
Like RPL, Factor is a stack-based language: all of your code shares a common stack of data, rather than passing data around as variables. I had never thought much about the implications of this, but I did read up on the philosophy of concatenative languages. One of the things they point out is that we usually spend a lot of bytes on naming our data, when what's really important is what you're doing with the data. With Factor, you pretty much can't name your data, and you have to break your code down into very small chunks—factors—that perform simple operations.
Check out the code. It's worth a look if you're not used to this kind of language. It won't look like anything you've seen before.