Larry,
No need to respond to individual points, because you
are so clearly wrong. ;> But I would appreciate an
overall response of something like either "this ain't
happening, so give up" or "it remains a possibility,
but I'm not yet remotely convinced". Thanks for your
continued forbearance.
> [mental model of ordered and unordered is important]
Yes. But I'm not sure that:
# ordered
@array = (1, 2, 3, 5, 8);
# unordered
%hash = (Fred => 22, Jane => 30);
is more or less typical than:
# unordered:
@array = ('England', 'France', 'Germany');
# ordered:
%hash = (Name => 'Ralph',
Street => '2512 Essex Place',
City => 'Nashville',
State => 'TN',
Zip => '37212',
Country => 'USA');
-----------------
> [normal people want different operators for strings than for numbers]
I realize that perl is not a democracy (thankfully), but
it always helps to ask the people, as it were, so I've
created a survey about these issues, subtly slanted
in favor of voting for the status quo.
11 "normal people" (perl beginners) have responded
to this so far. 5 people say they prefer [] for both named
and numbered subscript parens, rather than [] for one
and {} for the other. 4 people prefer the status quo, and
these four hold this opinion much more strongly than
the ones voting for change. 2 don't care.
So, that doesn't help at all. ;>
You have noted the parallel between array/hash accessing
and function calls. The principle that people want this sort
of difference seems to suggest:
foo[1,2] # call function with numeric args
foo{'fred'} # call function with string args
which is just weird.
-----------------
If you deprecated %foo{bar}, you would simplify a, er,
key aspect of perl and eventually free up % and {} for
other duties.
