Perl ranks among hackers' favorite languages — there's always another approach to a problem, always optimization to be done, and forever new techniques to try. Perl's a hoot. But the culture of tinkering that surrounds Perl has resulted in a large body of hard-to-understand legacy code. That legacy code has to be maintained, extended, and adapted to new conditions — often without the help of the person who originally created it. Perl Medic considers Perl from the perspective of a programmer looking at code written by someone else and trying to answer the ancient question: "What were they thinking?"

It's a creative approach, and one that makes good reading for someone well-versed in Perl programming (author Peter Scott makes the analogy of becoming fluent in a human language, such as French, then studying its various accents and dialects). He shows, for example, a kludgy piece of code that's meant to catch CGI form uploads, then indicates that the obvious replacement is the CGI.pm module. Elsewhere, Scott shows why symbolic references are bad, and how to avoid them by means of hashes. Some of the value in this book is in the form of documentation of the differences among Perl versions; other useful coverage deals with warnings and strictness control as debugging tools. Read this straight through to improve your own code; use the index to help decipher and improve what someone else has written. — David Wall

Topics covered: How to write good Perl code, read bad Perl code, and convert bad Perl code into good in less time that would be required to write an equivalent replacement program from scratch. Testing, debugging, documenting, replacing custom code with CPAN modules, and embracing features that became available in later versions of Perl are all covered. Overall, the author endorses and generally explains the principles of Extreme Programming (XP) for Perl work.