2020ok  Directory of FREE Online Books and FREE eBooks

Free eBooks > Computers & Internet > Programming > Introductory & Beginning > General > Perl Medic Transforming Legacy Code

Perl Medic Transforming Legacy Code

by Peter J. Scott

Download Book

If you are the author or the publisher, and would like to link to your site here, please contact us.

About Book

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.

Book Description

This book is about taking over Perl code, whether written by someone else or by yourself at an earlier time. Developers regularly estimate that they spend 60 to 80 percent of their time working with existing code. Many problems of code inheritance are common to all languages, but the nature of the language makes Perl especially tricky. The reason why is that Perl is similar to English - bursting with irregular verbs, consistent only when it's convenient, borrowing terms from other languages, and providing many ways to say the same thing. In fact, Perl developers have a motto with the abbreviation TMTOWTDI: There's More Than One Way To Do It. While this flexibility is one of the language's strengths, it also makes it extremely difficult when you are faced with an existing piece of code. There are millions of lines of Perl code being used all over the Web; much of it was built on an ad hoc basis, the creators never imagining that the code would still be in use months or years later. This book will be the resource all Perl programmers need to understand someone else's code, even when it's bad; repair it; convert it to a better style; upgrade it to the latest version of Perl; maintain it; and find and fix its bugs.

From the Back Cover

  • Cure whatever ails your Perl code!
  • Maintain, optimize, and scale any Perl software... whether you wrote it or not
  • Perl software engineering best practices for enterprise environments
  • Includes case studies and code in a fun-to-read format

If you code in Perl, you need to read this book.—Adam Turoff, Technical Editor, The Perl Review.

Perl Medic is more than a book. It is a well-crafted strategy for approaching, updating, and furthering the cause of inherited Perl programs.—Allen Wyke, co-author of several computer books including JavaScript Unleashed and Pure JavaScript.

Scott's explanations of complex material are smooth and deceptively simple. He knows his subject matter and his craft-he makes it look easy. Scott remains relentless practical-even the 'Analysis' chapter is filled with code and tests to run.—Dan Livingston, author of several computer books including Advanced Flash 5: Actionscript in Action

Bring new power, performance, and scalability to your existing Perl code!

Today's Perl developers spend 60-80% of their time working with existing Perl code. Now, there's a start-to-finish guide to understanding that code, maintaining it, updating it, and refactoring it for maximum performance and reliability. Peter J. Scott, lead author of Perl Debugged, has written the first systematic guide to Perl software engineering. Through extensive examples, he shows how to bring powerful discipline, consistency, and structure to any Perl program-new or old. You'll discover how to:

  • Scale existing Perl code to serve larger network, Web, enterprise, or e-commerce applications
  • Rewrite, restructure, and upgrade any Perl program for improved performance
  • Bring standards and best practices to your entire library of Perl software
  • Organize Perl code into modules and components that are easier to reuse
  • Upgrade code written for earlier versions of Perl
  • Write and execute better tests for your software...or anyone else's
  • Use Perl in team-based, methodology-driven environments
  • Document your Perl code more effectively and efficiently

If you've ever inherited Perl code that's hard to maintain, if you write Perl code others will read, if you want to write code that'll be easier for you to maintain, the book that comes to your rescue is Perl Medic.

On the Web Site

Download all of the book's sample code from <www.perlmedic.com>.

About the Author

PETER J. SCOTT runs Pacific Systems Design Technologies, providing Perl training, application development, and enterprise systems analysis. He was a speaker on the 2002 Perl Whirl cruise and at YAPC::Canada, and he founded his local Perl Monger group. A software developer since 1981 and a Perl developer since 1992, he has also created programs for NASA's Jet Propulsion Laboratory. Scott graduated from Cambridge University, England, with a Master's of Arts Degree in Computer Science and now lives in the Pacific Northwest with his wife Grace, a cat, and a parrot, at least one of which also uses Perl. He is the lead author of Perl Debugged.

Excerpt. © Reprinted by permission. All rights reserved.


Worldwide, there are well over 200 billion lines of software that are fragmented, redundantly defined, hard to decipher, and highly inflexible… organizations run the risk of being mired down by a mountain of legacy code.—William Ulrich, Legacy Systems: Transformation Strategies

Congratulations! Let's say you just graduated with a computer science degree and now, bucking the economic trend, you've landed a job at a prestigious company with a large information technology department. You're going to be replacing Bill, a programmer who won the lottery and was not seen or heard from again, save for a postcard from Puerto Vallarta two weeks later. Your coworkers warn you not to mention the postcard to your supervisor. You sit in Bill's cubicle throwing out pieces of vendor advertising left in the center desk drawer, thinking about how you're going to apply the elegant principles and sublime paradigms that professors inculcated in you at college. Just then, your supervisor arrives and, leaning over your shoulder, taps at your keyboard, bringing up a file.

"This is the last program Bill was working on. We think it's almost finished. We're behind schedule, so see if you can get it done by Thursday at the latest."

As he leaves, you look at the program's tangle of misindented lines and cryptic variable names, searching for comments, but the only ones you can find read, "XXX-Must change" and "Kludge!-But should work." You wonder whether this is a corporate hazing ritual, but your instinct tells you otherwise.

Welcome to the real world.

In the real world, you're lucky if you get to spend all your time developing one new program after another. Much of the time you'll have to deal with someone else's. In the real world, programmers take over responsibility for programs written by people they might not know, like, or agree with. Even if you're fortunate enough to avoid this situation, you'll still have to maintain your own code; and one day you're going to look at something you wrote two years ago and ask, "What idiot wrote this?" Thereby arriving at more or less the same situation as the less fortunate programmers.

This book is about taking over Perl code, whether written by someone else or by yourself at a time when you were less wise about maintainability. Many problems of code inheritance are common to all languages, but I have noticed them especially in Perl.

Why does Perl tend to foster maintenance issues? The answer to this is the dark side of Perl's strength and motto: "There's More Than One Way To Do It" (enshrined in the acronym TMTOWTDI). Perl provides so many ways to do it that someone else quite possibly picked one that wasn't your way, or might have used several different ways all in the same program.

The medical metaphor for this book stems from the rather drastic nature of the work we do as maintenance programmers. Often we must perform triage, deciding what code is worth saving and what is beyond redemption. Frequently we only have time for first aid, applying a field dressing to a ruptured program. We also have a hard time explaining our bills to the client. There may not be a Hippocratic Oath for programming, but it wouldn't hurt to come up with one.

I wrote this book because I kept finding myself telling my students, "I'm going to teach you how to program Perl well, but I'd have to teach you a lot more before you could take over a program that wasn't written well, and you wouldn't appreciate taking that much time away from learning how to write good programs of your own." So I've written a book to fill that need.

Perl is to computer languages as English is to human languages: bursting with irregular verbs, consistent only when it's convenient, borrowing terms from other languages to form a great melting pot of syntax. Most computer languages are described in terms of some kind of functional niche (Pascal: teaching computer languages; FORTRAN: numeric analysis; Prolog: rule-driven expert systems; etc.). Perl's makers simply describe it as "a language for getting your job done." Perl hosts a fantastic conglomeration of syntactic devices that allow a programmer from virtually any background to find a familiar foothold for learning the language.

The full picture isn't as chaotic as this might imply: Larry Wall and others have done a brilliant job of tying together these eclectic devices into a framework that has an essential beauty. Therefore, just as the British speak of a "BBC English," while many people program Perl with, say, a LISP or C accent, there is something approaching an "accentless Perl" style that leverages the language's features to their best advantages. I will show how you can "speak Perl like a native" in order to optimize the maintainability of your programs.

It's true that you're officially allowed to program Perl in "baby talk" and Perl gurus have promised "not to laugh." (See the same preface.) But by the same token, while aviators call any landing you can walk away from a good one, what I'm doing in this book is helping you avoid having your pilot's license revoked.

Perl is like those people behind the travelers' help desk in airports; it's very good at understanding you no matter how poor your command of their language is. Because there are so many ways to write a Perl program that is not only syntactically correct (Perl makes no objection to running it) but also semantically correct (the program does what it's supposed to--at least in the situations it's been tried in), there is a wide variety of Perl programming styles that you might encounter, ranging from beautiful to what can charitably be described as incomprehensible.

The savvy among you will take that information and ask, "Where do my programs fit on that scale?" Because someone else may end up inheriting your code, and you'd prefer that they not end up sending it to authors like me as bad examples to go in books like this. See Chapter 5 for more advice on avoiding scorn.

If your experience or image of Perl is limited to short, mundane scripts, this book will appear to be overkill. I want you to know that Perl can quite easily accommodate large--as in tens of thousands of lines of code, multiple modules, and multiple programmers--projects. Projects of the size that demand rigorous requirements, documentation, and testing. If you're used to Perl programs escaping that sort of attention, I believe that is partly the result of a misperception of the role and power of Perl.

For example, if a C program is written to fulfill some requirement and turns out to be 1,000 lines long, then the common reaction is, "This must be serious … we'd better have code walkthroughs, acceptance testing, operational readiness reviews, and static code analyses. Oh, and don't forget the Help Desk training and documentation."

But if a Perl program that fulfills exactly the same requirements weighs in at 100 lines (and 10:1 is a typical compression ratio for C code to Perl), the reaction is more likely to be, "Ah, a simple utility … and in a plebeian scripting language to boot. Just plunk it in the delivery directory and get on with the next task."

When a Perl program reaches the 1,000-line mark, however, the honeymoon is probably over. Much of what I have to say addresses large programs. Chapter 3 in particular will show you how to get the respect of development teams who are used to putting everything through regression testing.

Please also see my earlier book with Ed Wright, Perl Debugged (Addison-Wesley, 2001) for more advice on good practices for developing and debugging Perl programs.

Perl or perl?

When you read this book and other works about Perl, you'll see an apparent inconsistency in capitalization: sometimes it's written as "Perl", and others as "perl". There's really no inconsistency; the authors are referring to two different things. Perl is the language itself; perl is the program that runs Perl programs. There is only one Perl, but there are many perls (one or more for each type of computer).

Sometimes this distinction gets a bit blurred: For instance, most people will write, "Perl objects to mismatched parentheses" when it is arguably the program that's doing the objecting and not the language. Just don't write "PERL"; Perl isn't an acronym, it doesn't stand for anything. (Well, aside from standing for diversity of expression, freedom from artificial constraints, and the right to have fun in your work. But we'll get to those later.)

Obtaining Perl

It would be remiss of me to tell you so much about Perl without telling you how to get it, although these days it's hard to avoid; you probably already have it, especially if you have any flavor of UNIX or Linux. The easiest way to find out whether you have it is to get a command prompt and type:

	perl -v

and see if you get a response. If not, try:

	whereis perl

which on a UNIX system or similar will look around for Perl. If that doesn't work for you either, here are brief instructions on how to download and install Perl:

  • For Microsoft Windows machines, get the free ActivePerl distribution: <http://www.activeState.com/ActivePerl/download.htm>.
  • For Macintosh: <http://www.cpan.org/ports/index.html#mac> (That URL is for pre-X versions of the OS; Perl comes with Mac OS X and builds fine on it, too.)
  • For binary distributions for all other machines: <http://www.cpan.org/ports/>.
  • For the source of perl itself: <http://www.cpan.org/src/>.

The source file you want is called stable.tar.gz. The file devel.tar.gz is for Perl developers or testers only, and the file latest.tar.gz is the same as stable.tar.gz for complex historical reasons. Anything mentioning "Ponie" will be for developers only through 2004 at least, and any perl with a three-number component version with an odd middle number is likewise a development version.

Building Perl from source on a supported UNIX architecture requires just these commands after you downlo...



PLEASE READ: All comments must be approved before appearing in the thread; time and space constraints prevent all comments from appearing. We will only approve comments that are directly related to the article, use appropriate language and are not attacking the comments of others.

Message (please, no HTML tags. Web addresses will be hyperlinked):

Related Free eBooks

Related Tags

DIGG This story   Save To Google   Save To Windows Live   Save To Del.icio.us   diigo it   Save To blinklist
Save To Furl   Save To Yahoo! My Web 2.0   Save To Blogmarks   Save To Shadows   Save To stumbleupon   Save To Reddit