On Tue, 2002-09-24 at 14:47, David Whipp wrote: > It seems that the fundamental problem is the dichotomy between > a scalar, and a list of 1 elem. Thus, we want
After the first couple of messages, that was really no longer *my* concern, but I can't speak for others. My concern was mostly that parentheses and brackets are now doing exactly the same thing with some very limited non-intersection around single-element lists. So my question was: why would we want that? Why would anyone ever need to have this in their code: C<(1,[2,(3,[4,(5,6),7],8),9],0xa)>. But, I'm quite content to wait for Larry to get back and have a chance to mull all of this over. He tends to have a very level head about these things, and I imagine he'll have a point of view colored by some concerns he has not yet put into writing. For what it's worth, I once started working on a cut-down, general-purpose-only version of Perl that I was calling Sand. The interesting thing was that I made a lot of these same decisions, but I went even further. There was no array or hash type at all. They could only be manipulated by reference through anonymous structures or scalar variables. I rat-holed on the topic of how to pass parameters without throwing away the benefit of Perl's sloppy-list-based stack that was always so useful. A lot of these "why do I need parens" and "how do you explode a list when you want, but not when you don't" sorts of issues kept coming up, and I just didn't have time to explore them. Now that Perl 6 is starting to take shape, I may go back and finish Sand as a Parrot front-end (though it was ultimately aimed at being purely compiled like C). -- Aaron Sherman <[EMAIL PROTECTED]>
