Softpanorama (slightly skeptical) Open Source Software Educational Society

May the source be with you, but remember the KISS principle ;-)

Learning Perl on Win32 Systems

By Randal L. Schwartz, Erik Olson & Tom Christiansen

1st Edition August 1997

1-56592-324-3, 306 pages, $29.

This is a pretty weak introductory book. Don’t buy it with your own money (corporate funds are probably OK if the book is discounted 50% ;-). The authors definitely do not understand the Win32 environment well enough. The language is explained as an isolated phenomenon without reference to existing scripting tools (VBA, VBScrit, REXX, etc.). The main use of Perl in the Win32 environment is CGI programming and in this area the book does not have enough value for beginners.

Generally this book looks like an attempt to adapt the original Randal Schwartz text   to the Win32 environment (second edition for Learning Perl  which is oriented mainly on the UNIX audience also is no longer the best introductory book available, but at least the first edition was one of the first books on Perl). The adaptation attempt failed and the results are pretty disappointing. IMHO Cross platform Perl by Eric F. Johnson is a better deal.

Neither the examples, nor style are very suitable to Win32. For example Chapter 1 the subtitle "Standard Perl distribution" (page 6) contains the following recommendations:

• nmake                 ( Build all of Perl)
• nmake Test         (Test your distribution)
• nmake install    (Install to the target dir. in the Makefile)

These recommendations can be considered a joke for Win32 beginners, who get used to prepackaged software.

Chapters are very uneven in length and quality.

Chapter 1 is disappointing. The secret word guesser introduced in Chapter 1 is definitely far from the best introductory example for Win32 audience. The number of unexplained concepts introduced here would scare novices away. Probably more simple examples like a couple of UNIX-style filters (head, tail) or generation of HTML for the famous phrase  "Hello world"  would be more appropriate.

Chapter 2 is OK.

Chapter 3 does not provide a clear description of Perl syntactical and semantical achievements and pitfalls. For example, Perl considers a negative subscript on the array as a count back from the end. So each element has two indexes (from the beginning and from the end). The last element has an index -1, which is really very convenient. And what is more important Perl uses the same approach in the substr built-in function. 

Chapter 4 seems to be written by "structured programming diehards" and is very weak. Rich Perl control structures are not covered. Neither continue, next, break operators, nor the short-circuit semantics of || && operators in if statements (not apparent for people who never programmed in C) was not discussed. IMHO continue, next, break operators really simplify construction of non-trivial loops and in this respect Perl looks superior to Java.

Chapter 5 (hashes) is only 4 pages long for such an important subject as associative arrays. Strangely enough it does not mention the %ENV -- probably the most important associative array for novices.

Chapter 6 (Basic I/O) is very short. Perl after all was designed to process logs and such a chapter definitely should be more than four pages long and should contain more examples.

Chapter 7 fails to recognize the different mentality of the Win32 users. Basically there are two approaches to the regular expressions in the Win32 environment. One is to use regular expressions only when necessary and the second is to use them whenever possible. The first camp (partially represented by former REXX users and VB users) prefer a more procedural way of text manipulations and tend to rely more on substr, index, length, split and join functions, while the second is the regular expression addicts. The authors definitely belong to the second camp ;-). IMHO non-greedy (minimal) matching is often more convenient than a regular one, especially for novices, but such a possibility was never mentioned in the book.

Chapter 15 is probably one of the few things that were done right in the book. The index and substr built-in functions are covered adequately. I would like to mention that the authors cover a very important use of the substr in the left part of the assignment statement (the idea probably borrowed from the PL/1).

Chapter 16 is just OK.

Chapter 17 is a joke as DBM is the Unix-only beast. Only the last page ("Win32 Database interfaces") can be considered relevant to the Win32 users. It's really depressing to see such a blunder in an O'Reilly book.

The chapter about CGI is somewhat superficial. At the same time it is the main application area for Perl, especially in the Win32 environment.

The only subject covered in this book that was not covered in the Cross Platform Perl is the coverage of OLE2 (Chapter 19). It uses a rather interesting example of MAPI.

Win32 is not UNIX. Strangely enough for such a book the discussion of pitfalls of Win32 environment is completely absent. For example the authors fail to stress the difference between chop and chomp functions. Even the most annoying difference between UNIX-style and Win32 style directory paths syntax (with "\" instead of "/") was never mentioned.

I believe O'Reilly should do a better reviewing and editing job, if it wants to preserve its reputation of a publisher of consistently well written and edited books.

- Nikolai Bezroukov

Standard disclaimer applies.

Last modified: February 28, 2008