$a=ord$/;;;for(483315457,3759149058,3883269635,3794874370,1647390720,
1644171264,1912787457,4277106435,4100403459,608567554,608436482){$c=0
;for(split//,substr unpack("B*",pack"L*",$_),6){$o[$c]+=(ord d,ord 2,
$a,$a,$a,$a,$a/2,$b=$a-9,$b,$b,$b)[$i]if$_;$c++}$i++}print chr for@o
[download]

Now that it's been unraveled here's the spoiler.
Each "long int" represents a bitmap. Each bitmap is the same width as "Just another Perl Hacker.", so one bit corresponds to one letter. However, there are about eleven rows of bitmaps. The first row corresponds to Roman numeral 'C' (100). The second row corresponds to 'L' (50), then a few rows of 'X' (10), one of 'V' (five), and a few of 'I'. Add the rows up by vertical column within the bitmap, assigning proper value to each row, and you get a number that corresponds to an ascii value in each column.

Anything so straightforward probably shouldn't be an obfu, but I thought the method of storing the roman numerals in sequences of bits carried with it some intregue.

That's it for the spoiler. Back to work...