On Mon, May 09, 2005 at 11:34:10AM -0700, Larry Wall wrote: > On Mon, May 09, 2005 at 10:33:33AM -0500, Patrick R. Michaud wrote: > : After thinking on this a bit, I'm hoping we don't do this -- at least not > : initially. I'm not sure there's a lot of advantage of C< $1.1 > over > : C< $1[0] >, and one starts to wonder about things like $1.$j.2 and > : $1[$j].2 and the like. > > Or maybe it should generalize the other direction. We just got through > the great bracket shift to make $x<a> mean $x{'a'} so that we can recognize > constant hash subscripts easily. Maybe $x.1 is just the numeric analog > of that. And $x.$j.2 could just fall out of that, where the indirect > method dispatcher knows to turn a numeric method name into a subscript.
Hmmm, then would $x.$j.2 then be equivalent to $x[$j-1][1] ? Pm