Re: Tailcall

[Jay Carlson <nop@...>]

On Wednesday, December 26, 2001, at 06:55 AM, John D. Ramsdell wrote:

> The problem is that people have an erroneous mental model of
> computation in which invoking every procedures involves pushing
> arguments on a stack, saving a return address, and then jumping to the
> body of the procedure.  This is not how a tail call works.  There is
> no need to put a return address on the stack.

This mental model of computation is correct for most language 
implementations, and what's more, most languages are taught this way.  
("Hi, my name is Jay, and I was a BASIC programmer."  "Hi, Jay.")

A significant audience for Lua is the casual programmer.  We're both 
highly appreciative of how people with any kind of procedural 
programming experience can pretty much guess how to write small bits of 
Lua code, given suitable examples.  So breaking the expectations of 
casual users is not something to be done lightly.

On the other hand, the two places people will encounter this is in 
stack-introspective code (like MOO's callers(), which mostly is used to 
talk about the most immediate caller) and in stack dumps ("how the hell 
did the code get *there*??").  So it's not something that the *really* 
casual user will run into.

> If you misrepresent the concept of a tail call, you will produce
> confusing in error messages.  Treat tail calls as gotos, and error
> messages will enlighten users.

In the context of Lua, it bothers me to have such an important 
distinction made solely by statement position.  Would treturn or tail 
statements help?

Jay