Making Programs Faster

Pseudo-Hashes

For 5.005, someone had an interesting idea:
Suppose an object was declared from a certain class, like this:

        my Critter $self;

Suppose Critter objects are based on arrays instead of hashes
And suppose Critter.pm declared its fields at compile time, like this:

        package Critter;
        use fields qw(NAME TYPE size hearts likes_cookies 
           ...
                pelagic tentacles is_tasty 
           ...
                );

Then when Perl saw $self->{TYPE} it could pretend you wrote $self->[2]
You would get all the benefits of both!