Like of alot of the other replies here, I have moved away from the C (C++/Java/whatever) syntax that alot of people start out using.

Im a unix admin, so my coding revolves around my responsibilities. The modules I use most are, Getopt::Std, Net::SSH::Perl, DBI*, and IO::Socket. Aside from that, I usually bang out my own solution, tweaked for the application at hand. I have built up a sizeable code base in my home directory, which I port around and tweak as necessary.

At first I was writing my code to be understandable to anyone with scripting/programming knowledge, but I have found lately that it leads to performance degredation in some situations. I value consistancy in my code almost as much as efficiency. I want someone who is familiar with perl to be able to read one snippet, then look at another and be reasonably certain it was I who wrote it. We have sh/ksh/csh scripters here, some Java, some Perl, but I think my 'style' is recognizable to whomever comes along.

Things Ive been doing lately:
Really starting to enjoy the power of map
Simple tests and assignments go on one line
Reducing redundant code by using subs ( maybe to a fault )
Matching via regex, as opposed to split()ing elements if possible
Moving to using arrays over hashes where possible/resonable
Optimizing loops for speed (preprocessing data, reducing testing inside loop, etc)
Passing around refs to data as opposed to data structures themselves

So my dialect isn't too off the wall, nor sexy, but its definately fast, which is ok by me. I haven't had the need to attempt playing inside of main's namespace tweaking special vars, or internal functions. I will leave that to wiser monks than I {at least till I need the functionality ;)}.



/* And the Creator, against his better judgement, wrote man.c */