I still think, actualy, that symbol tables are magic hashes, and not another sort of entity

After some further investigation I've come to the conclusion that under the hood symbol tables are indeed just fancy hashes. So I now sit corrected :) I believe their special properties are as follows

• can't be lexically declared (due to the ::)
• they can be declared with a package declaration
• they hold the globs which hold the package level variables/subroutines/etc
• must have a :: appended to the name to qualify as a symbol table
• probably other bits and pieces which don't spring to my misty mind on a Monday morning

I guess this means I'll have to change a lot of the wording, but at least it'll be accurate this time. Thanks for forcing me to put my money where my node is :)

_________
broquaint