Hashslices are useful when you want to get/set the values for more than one key at a time. So, instead of

my @values;
push @values, $hash{$_} for qw( key1 key2 key2 );
[download]

my @values = @hash{ qw( key1 key2 key3 ) };
[download]

Some people feel that hashslices are annoying to read. Others, like me, love them. No-one says you have to use them. If you like them, great. If you don't, great. :-)

I think the issue you're running into is that if CGI::Carp traps an error and attempts to output the message to the browser before your CGI script has had a chance to output proper CGI headers, you'll get that old familiar problem of attempting to generate output without proper headers.

The solution is for your CGI::Carp handle_errors() subref to handle the outputting of headers too. You may find it helpful to keep a flag that lets you know whether or not you've already output a header. The error handling has to be established in a BEGIN{} block so that compiletime errors will properly be handled too.

Regarding hash slices, consider the following examples:

@hash{ @keynames } = @valuelist;  # Assign values to a bunch 
                                  # of keynames at once.

@valuelist = @hash{ @keynames };  # Grab a bunch of values given 
                                  # a list of keynames.

( $address, $phone ) = @{$people{ $name }}{ 'address', 'phone' }; 
                                  # Grab individual fields
                                  # from a record.
[download]

The usefulness of a language feature is just a matter of need and imagination.

For a code example see Re: Essential CGI Security Practices by gellyfish.

--
Do not seek to follow in the footsteps of the wise. Seek what they sought. -Basho