in reply to Puzzlement in Splitsville

Id say you have got yourself into a situation of functional fixidity. Instead of using the list form of regexes, use the /g modifier with while. Its intended for exactly this type of scenario. Also, for your first question about splitting into triplets or what have you its probably more efficient to use unpack if you can. Im not sure but its possible 5.6 doesnt support the pattern grouping in pack formats.

D:\>perl -le "print $1 while 'aaabbcddddaee'=~/((.)\2*)/gs" aaa bb c dddd a ee
D:\>perl -le "print for unpack '(a3)*','atgactaatagcagtgg'"; atg act aat agc agt gg

Note the difference in approaches between using for and while. For expects the list to be preconstructed, while expects the list to be constructed as we go.

UPDATE: Apologies to thundergnat, I didn't notice his comment before i posted.

UPDATE2: To answer your question about eliminating every $N th element id say you have the right approach. I would do it via:

my @filtered=do { my $i=1; grep $i++ % $N,@unfiltered};

---
$world=~s/war/peace/g