David, it was my typo (in the stuffed var names and yes,
[subst $y]
does work, thanks Don. I was testing this in a proc in /tcl and
this process is slow because aolserver must be restarted
each time.
I was also thinking that giving users access to vars was
perhaps a bad idea. It seems like undefined vars don't throw
an error if in a /tcl/ proc. But Michael has made a
*most* convincing security argument. String maps look like a great solution and maybe I'll use something other than a $... In another thread Jerry Asher suggested this format ::name::. I was thinking maybe [name] would work as something simple... that would be [name] in the string map. That's one backslash depending on how you see this message!
-Bob
