[Rd] proposal for "strict" versions of subsetting operators

[Tony Plate]

I suspect that it would break too much existing code to make all 
subscripting behave in a "strict" manner.  However, I wonder if it would be 
possible to use namespaces and packages to make a particular package get 
strict versions of subscripting operators?  It would have to be done in 
such a way that functions called in a different package would use the 
appropriate subscripting behavior for their package.  This might be a way 
to manage a transition to strict subscripting.  However, I suspect that 
this would all be extremely confusing to users.

-- Tony Plate

At Tuesday 12:49 AM 10/28/2003 -0500, Gabor Grothendieck wrote:

>Having written a number of R programs and having been stung with
>this myself, I think this is a very good idea.  In fact, if the
>aim of R is to "to do it right" (as one person on the list claimed)
>rather than be compatible with S-Plus or prior versions of R then
>I would go even further and make all subscripting behave like this.
>(If that were to be the way R goes then, of course, some thought
>would be required on how to smoothly transition to it.)