I'm nearing the end of my adventure with Perl, my field has taken a turn in direction and Perl hasn't kept up, but it remains my go-to prototyping language and my all-time favourite.

Perl taught me (at least) 3 things:

1. Source code is king; but the programmer is king maker.

   And clearly laid out, concise --even when highly complex and idiomatic -- code, is -- assuming a willingness to achieve good familiarity with the language -- far easier to read, understand, modify and maintain than verbose, prosaic, generic code that avoids idioms and specialties in an attempt to "be readable at first glance". And, with a modicum of effort, it takes less time to produce, runs faster and contains less bugs.

2. Understanding and using the full power and nuance of the language can save 80+% of the kloc to perform a given task; and 90+% of the time you wait for the results.

   Think about that! That means it takes 1/5th the time to write and runs in 1/10th time.

   Those are numbers that make learning the language -- rather than writing COBOL, Pascal, or C in Perl, well worth the effort.

   None of the other 10+ languages I've become competent in, nor any of the other 20+ I've explored, has taught me so much, nor enabled me so well. A big part of that is the concise thoroughness of the documentation. (Almost) Anything you need to know about Perl is defined and catalogued in the documentation. It may take several readings to find it; and two or three more times to recognise it, but most everything you need to know is there. Even the hairy dark corners.

3. Code that runs clean (strict & warnings), but is functionally incomplete, is worth more than any amount of code, no matter how theoretically, functionally complete it purports to be, that doesn't run, and run clean.

Perl (5) is pretty much complete. Despite a world that wants to orthogonalize everything, banish special cases, objectify and box everything into a corner, and thus reduce the role of the coder to that of a billion monkeys tapping on typewriters; Perl demonstrates that generality costs: that concentrating effort in those areas that are used -- even at the cost of the loss of orthogonality and generality -- pays dividends over and over.

There is no point in having every operator handle every possible combination of parameters, if 90% of those combinations are never used.

Perl5's design and implementation recognises that simple fact and expends the effort where it is most effective. And much -- most, if not all -- of that is down to the pragmatism over vision of the original designer. L.Wall, I salute you!

(Just a shame you got side tracked.)