Given Tcl 8.5's {*} expand feature, and an earlier thought experiment about adding a leading "$" to every command (can't recall what wiki page this was) ...
(KJN that wiki page is Getting rid of the value/command dichotomy for Tcl 9.)
What would a system be like if it worked as follows - this notation:
set obj {...}
set result [obj arg1 arg2 arg3]... would behave as if it had been written as follows in Tcl 8.5:
set obj {...}
set result [{*}$obj arg1 arg2 arg3]It obviously wouldn't be Tcl, since the variable and command namespace get mixed up when you start doing this.
But suppose you didn't care (Tcl 3000!) - would it be useful? convenient? worth looking into?
MJ - I am not seeing the use for the implicit $, is it so much trouble to write [$obj arg1 arg2 arg3]? Autoexpansion of the leading word would be very useful however (especially in OO type constructs and functional programming).
It seems you are suggesting unification of variables and procs, which sounds a lot like the concept of slots in Self, in that case $ will become unnecessary. Getting the value of a var is simply [var].
DKF: This idea ("auto expand leading word") has been known about and discussed (on and off, when we've nothing else to do) for years.
JJS: Isn't this a lot like [interp alias]?
interp alias {} obj {} ... ;# instead of set obj {...}
set result [obj arg1 arg2 arg3] ;# works like {*}$objNEM: The difference is that variables can be local to a proc whereas aliases can't. With apply you can now have procedures that are stored in local variables, but not commands in general, and the syntax could be better. I wrote some thoughts about ways to accomplish the same thing in a hypothetical Tcl 3000 here [L1 ].
RS: See Let unknown know how to tweak Tcl that it does auto-expansion of leading word... :^)