By consistantly using a system like this throughout your programs, there won't have to be question about whether a scalar contains something like a string, or if it's a reference to an array or hash.

7 years later, I have this to say ….

It seems to me that this offers exactly the benefit you describe, but that that benefit might not be what it seems at first glance. That is, I can tell by looking at a variable name ^* whether it is

• a string,
or
• a reference to an array or hash;

but I cannot answer the question of whether it is

• a string,
• a reference to a string,
• a reference to an array,
or
• a reference to a hash.

(For that matter, there are all sorts of other references (and, for that matter, other things to stuff in scalars, like numbers), too, like \\1.) It seems to me that this is a level of granularity that one would like, in many cases, to be able to maintain. I don't know a better solution than the pseudo-Hungarian $ima_href, though, and I don't think much of that.

^* Although it's important to note that I can't even necessarily do that—is $moose singular or plural?