Re: extensions to settor variables?

[friedman@cli.com (Noah Friedman)]

>I must admit executing a function every time a variable is referenced
>seems a bit overwhelming to me.  This creature doesn't feep, it galumphes!

It's a terrible idea.  There is already a perfectly good mechanism for
returning arbitrary, computable values from a reference.  They're called
functions, and they make the volatile nature (including the possibility of
side effects) of such references explicit.

I like settors just the way they are.