Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?

Pragmatic Project Automation 69

twelve71 (Alan Francis) writes "Apologies in advance for overuse of the word 'pragmatic,' but Dave Thomas and Andy Hunt together form a company called The Pragmatic Programmers, and published a book I'm sure many of you have read, titled The Pragmatic Programmer: from Journeyman to Master. The Pragmatic Programmer (or 'PragProg' as it is usually referred to) is a wonderful grab bag of 'good old common sense,' but its main strength (covering a very broad range of subjects) means that the authors have left a few holes around some important details. To plug some of these holes, and provide a good grounding for those just starting out, they have recently published 'The Pragmatic Starter Kit' - a set of books covering in detail some of the basics mentioned in PragProg." Read on for Alan's review of the latest book in the kit.
Pragmatic Project Automation
author Mike Clark
pages http://www.pragmaticprogrammer.com/bookshelf/index.html
publisher Pragmatic Bookshelf
rating 9
reviewer Alan Francis
ISBN 0974514039
summary Learn to use common, freely available tools to automate build, test, and release procedures.

The first 2 (or possibly 3) books are Pragmatic Version Control with CVS and Pragmatic Unit Testing (which is available in Java and .NET flavours).

Pragmatic Project Automation is the latest book in the series and, interestingly, this book wasn't actually written by either Dave Thomas or Andy Hunt, but by Mike Clark (contributor to the 'Bitter EJB' book, editor of the JUnit FAQ, and responsible for the JUnitPerf and JDepend tools). Mike does a great job of ensuring this book fits in with the overall style of the other books in the series.

Up front, in case you're a "cut to the chase kind of guy," this book (and the others in the series) are must-haves and as a consultant with ThoughtWorks I'll have a set ready to distribute whenever I start working with a new client team.


The book helps the reader build a Java project slowly over the chapters, starting with a manual build-and-deploy process and automating a new aspect of it, chapter by chapter, until by the end of the book the software compiles, archives, deploys and configures itself multiple times a day. In addition, the build tools notify you of success or failure in a variety of interesting ways including email, SMS messages and different colored lava lamps.

The first chapter provides a good introduction to the different types of automation available to projects. It also introduces the acronym 'CRISP' to help the reader remember the desirable characteristics of an ideal build process: Complete, Repeatable, Informative, Scheduled and Portable.

Chapter 2 gets you Repeatable by using ANT to bring together all the various steps you currently perform on your project into a single, one-click build. Chapter 3 works to turn the one-click build into a no-click build using tools as simple as cron as well as more complicated tools like the ANT scheduling tool CruiseControl. By the end of these two chapters your software can be compiling and testing itself automatically each time changes are checked into the version control system. But this is only the beginning.

Chapters 4 and 5 address the "Complete" and "Portable" portions of the CRISP model discussing how to include packaging, release management and deployment into your scheduled build. The last chapter addresses "Informative." How to monitor the build for success or failure, how to notify members of the project team using email, SMS, RSS or even the red and green Lava Lamps I mentioned above.


This book, and the others in the series, provide a much needed set of manuals for getting a good set of basic practices up and running at the start of a project. Unlike the Unit Testing book, there's not a lot of programming in this one, but it's a worthwhile read for any programmer, regardless of experience level.

Many people are becoming interested in eXtreme Programming and Agile methods for software development. These books help to support some of the key ideas of those methods - extensive unit testing and continuous integration.

The main flaw in the book doesn't affect its usefulness, only its readability. Of all the files used in the sample project, the only one covered in any detail is the build file. The source and manifest files that we're writing the ANT file to build are never discussed or described - we are left to guess at what the sample project might contain (unless we download the code from the website). While this didn't change what I was learning (how to manipulate the project files with ANT), I like to understand all the details and this omission did occasionally leave me a little irritated.

You can purchase Pragmatic Project Automation from bn.com. Slashdot welcomes readers' book reviews. To see your own review here, carefully read the book review guidelines, then visit the submission page. If you are interested in the Pragmatic Programmers, see also this interview linked earlier from Slashdot.

This discussion has been archived. No new comments can be posted.

Pragmatic Project Automation

Comments Filter:
  • What if your an idealist??
  • by Anonymous Coward
    Squeegee Robot with sign saying: "Will Automate For Food"
  • by tcopeland ( 32225 ) * <tom@thomaslee[ ]eland.com ['cop' in gap]> on Tuesday August 31, 2004 @03:06PM (#10121401) Homepage
    ...I've integrated it into our hourly build [ultralog.net]; it's handy info.
  • by Colonel Panic ( 15235 ) on Tuesday August 31, 2004 @03:08PM (#10121426)
    The Pragmatic Programmers seem to be doing a great job with their new publishing efforts. They're able to get new titles out much more quickly than some of the more established publishers by following their own advice about automation and applying it to publishing. I recall reading on one of their blogs that they were able to go from idea to actual book in under six months.

    I can't wait for their upcoming Ruby [pragmaticprogrammer.com] book.
    • I assume you're aware that this will be a second edition. The first was good, but I could do without the printed reference material. I think if they cut that out they could bring it up to the standards of their other books fairly easily.
    • For those that don't already know, the first edition of this book is available free of charge [rubycentral.com].

      It's a great book and I can't wait for the second edition.

      P.S. Send an email to mando.escamilla@gmail.com if you want some gmail invites.


  • nice book (Score:5, Interesting)

    by Anonymous Coward on Tuesday August 31, 2004 @03:13PM (#10121471)
    This is a nice little book. Don't forget the web site [pragmaticautomation.com] either.

    I'm not a Java programmer (mostly Ruby and Perl) but I found a lot of stuff in this book inspiring. After reading this book I got bitten by the automation bug and did stuff like this:

    * gave my big deployed apps RSS feed logging for errors (i.e., now I can track recent warnings and errors in apps deployed across the country, just in NetNewsWire) .. why didn't I think of this before??

    * wrote a ruby script to automatically run unit tests whenever the files change.. based on a script on the web site. this is really cool! You edit your source file, save it, and pause to glance at the unit test window. again, why on earth didn't I think of this before!

    * wrote an automated test framework using WWW::Mechanize to log into web apps and check for errors or anything else and send out emails

    * use a nightly cron to check latest code out of CVS, run all tests, and run web tests..

    Anyway "project automation" is like unit testing .. once you get a taste you start doing it all the time and your productivity goes up yet another notch.

    So even if you don't do Java you can find a lot of ideas in this book (like the lava lamps showing build status!)
    • Re:nice book (Score:1, Insightful)

      by Anonymous Coward
      What other wheels have you re-invented?
      Simple tools for nearly everything you mentioned already exist.
  • I found that Donald Knuth's "Art of Computer Programming" was the most visionary book in applying code to real-world situations. Although this slashdot story is a status of the first anouncment on the Pragmatic Programmer being composed a while ago, I'm thankful its status is made known. I bought it from Amazon.com and am only 33% into the book. It is verry enlightening, equaly so to "Art of Computer Programming." I think it doesn't bring enough introductory information to warrant purchase by a Novice.
    • My Book [cafeshops.com] is focused on kind of the same area, but focuses on the interface between the application and the operating system and hardware. While Knuth essentially assumes the complete lack of an operating system, my book teaches how it interacts with your program through system calls, linking, memory management, etc. It talks about calling conventions, addressing modes, and all of the other gritty details that programmers need to know that usually just comes in bits and pieces from other sources.
    • >I found the best recommendation from my shelf to be "Computer Hardware Theory" (1972 University of Illinois) and "Fundamentals of Signal Theory" (1960).

      Two books from 30 years ago on computer hardware and one is on Signal Theory?

      Unless I'm doing assembly, could you explain why these books would be good for me, a programmer?
      • I thought he was joking. Maybe he was and the moderators didn't get it. No way is "Art of Computer Programming" a book about "applying code to real-world situations". Interesting academic information, yes, but not exactly mainstream stuff.
      • Hmm..well Hardware still consists of the same things it always has, Logic, ALU, Memory, Cache, etc. It just works faster. Things like how to understand accuracy & precision from the number representation in the Hardware haven't changed. Signal Theory..very hairy Math... Mathematical Algorithms have been known for high/low/band/gap and other types of filters for umpteen years. All the theories underpinning the mathematics hasn't changed in some cases 100+ years. I don't recall the basic theories of Calc
      • quoth { "Unless I'm doing assembly, could you explain why these books would be good for me, a programmer?" } quoth Application and System layer programming was once united under one title "Programmer", but it is now separated. The books I suggested only contribute to the over-all understanding to establish pre-ponderance in your source-code that it may be of good portage and maintenance should another be granted its access. Today, Application Programmers are separated unto further hostile classification
        • the more you know to develop your own syntax within your C is the better; to build your own Abstraction Layer. ...Unix was not supposted to be an Operating System but an actual behavioral API for how code may interact with foreign code upon differing architectures.

          Use Java Language (cleaner syntax than C), with the Java Virtual Machine (abstraction layer) and the Java libraries (API that works on different architectures) and problem solved.
    • by SlowMovingTarget ( 550823 ) on Tuesday August 31, 2004 @04:47PM (#10122352) Homepage

      Comparing Pragmatic Programmer to The Art of Computer Programming is like saying that raisins are just like prunes because they both make you go to the bathroom if you eat enough of them.

      Don't get me wrong, I absolutely love Pragmatic Programmer and would consider it a classic, but TAOCP clearly has more, ummm, fiber.

  • ...is right here [pragmaticautomation.com].

    It uses Java to talk to the serial port... seems like a good application for ruby-serial [rubyforge.org], too...
  • by autophile ( 640621 ) on Tuesday August 31, 2004 @03:33PM (#10121645)
    Oh, I thought it said "Pragmatic Project Automaton", which would be real useful.


  • Helpful links (Score:1, Informative)

    by Anonymous Coward

    Probably most of us already know what these are, but for those who don't:

    Extreme Programming [wikipedia.org].

    Agile software development [wikipedia.org].

  • by AndyHunt ( 168956 ) on Tuesday August 31, 2004 @03:36PM (#10121690) Homepage
    You can also buy the PDF and paperback direct from the publisher (and we really appreciate that) at www.pragmaticprogrammer.com/sk/auto [pragmaticprogrammer.com]

    That's also the home page for the book with downloaded code samples and so on.



  • Good Start (Score:3, Insightful)

    by Derkec ( 463377 ) on Tuesday August 31, 2004 @03:41PM (#10121731)
    This book seems like a good one for all developers to read. I'm honestly only about half way though it, but I've skimmed the whole thing.

    The big thing that jumps out at me is that he promotes how cool CruiseControl is for automating builds, but uses cron for his release builds. Generally, I would want to use the same tool to create my continuous integration, nightly and production release builds. I was under the impression that there were ways to make that happen in CruiseControl.

    disclaimer: I help write a product that does do this, so maybe I'm just projecting biases.

    The section on diagnostics also looks quite interesting, but I'm not that far yet.
    • Re:Good Start (Score:3, Informative)

      by cubicleman ( 739204 )
      On my current project, we've used Cruise Control for continuous integration and to run our suite of Junit tests (2500 tests and counting now) for the last 20 months or so..on a largish J2EE app (800+ ejbs, 30 developers, 3 distinct client builds), it makes sense and helps keep in check the stability of the app..
      • Cool, do you have CruiseControl manage the release builds as well or something else?
        • No..that's still a human process. A release manager decides which builds are released to QA for testing on a schedule (and then branches those releases that QA passes off into builds to be deployed to the clients for their testing).
  • Following his success in writing a book on Project Automation Dave Thomas [davethomas...option.org] decided he'd share his knowledge of Pragmatic Patty Production with the world.

    In this tell all exposee of beefly manufacturing processes, he teaches the enthralled reader how to use common, freely available tools to automate build, test, and release highly beeftacular burgers at a fraction of the current cost!

    Way to go, Dave Thomas!
  • by Catamaran ( 106796 ) on Tuesday August 31, 2004 @04:00PM (#10121905)
    I am very interested in the ideas presented here (extreme, agile, automated), but my experience is that - even with CppUnit and C++ support for ANT - the fit is not very good. Most of the test and build automation that one hears about is targeted toward Java.

    On the other hand, there is Test Environment Toolkit [opengroup.org] that noone seems to use. And STAF [sourceforge.net] which requires a huge investment of time just to comprehend.

    So, question: what tools do people find useful for build/test automation with C++ ?

    • Most of the setups I've seen are running CppUnit or similar with Make for builds and tests. Another option is OpenMake

      Making that happen repeatedly on a controled server is the domain of Anthill [urbancode.com], CruiseControl [sourceforge.net] and a handful [google.com] of for money tools.
    • by stew1 ( 40578 ) on Tuesday August 31, 2004 @04:33PM (#10122214) Homepage Journal
      I did XP in extreme C++ for about a year (by extreme C++, I mean boostified [boost.org], Alexandrescu [moderncppdesign.com]'d C++). We used CppUnit for our test framework.

      I'm not especially satisfied with the currently available C++ unit testing frameworks. CppUnit and Boost's both have trade-offs. I suspect they'll both get better, though.

      The Ant-Contrib [sourceforge.net] project's cc task works pretty well, in my limited experience. I was playing around with it just this weekend. I've yet to set up a tinderbox build process, but I don't see why this wouldn't be easy with either Cruise Control or good ol' cron.

      In my mind, the two biggest hurdles with doing XP in C++ are build speed and developer prejudice. You can tackle build speed with a combination of ccache [samba.org], distcc [samba.org], good [www.gotw.ca] programming [att.com] principles [hillside.net], and cash [dell.com]. Tackling developer prejudice is harder. A lot of C++ programmers like to write low-level, unsafe, old school C++ code. Modern C++ mostly lets you discard unsafe coding practices without sacrificing efficiency. Whether you can convince an old C++ programmer of this is another matter; it depends on the person. I've had decent success taking Java programmers and teaching them modern C++ via pair programming.

      Using Boost helps, indirectly. You write safer code, which gives you a faster development cycle. All in all, I think the basic tools are there. It might be a little harder to get fancy lava lamp integration going with C++, but there's no reason why you can't have a good build process. It's just that a lot of C++ projects haven't evolved (I think this is due, in part, to the fact that the C++ community is late to the internet; a lot of C++ programmers just don't know what's [apache.org], out [apache.org] there [c2.com].


    • You should try cxxtest [sf.net].
      It's a great unit testing framework that's VERY flexible.

    • Use DamageControl. It is a rewrite of CruiseControl in Ruby and is completely build tool agnostic. It is also written by guys from ThoughtWorks that got fed up with CruiseControl requiring Ant. (Full disclosure: I'm one of them.)
      It's not 1.0 yet but it's been building most of the projects at Codehaus for half a year now.


      Another tip: Use something like Ruby/Perl/Python for your automation needs. Not only are they platform-agnostic, they are also waaay more powerful than Ant
  • The Pragmatic Programmer (or 'PragProg' as it is usually referred to) is a wonderful grab bag of 'good old common sense'

    So, the PragProg is a grab bag. Who knew.

  • (or 'PragProg' as it is usually referred to)

    I'm a huge fan of PragProg. Oh Yes, we Rushed though development of our last application. Only once did we lose Focus, but you Can easily avoid these issues. Since its Genesis, we've got nearly a Marillion hits a day. I can't wait Tull I pick up this book.

  • If you are looking for a more generic build/test runner dashboard than cruise control, take a look at DART [kitware.com]. Its free, open source, and makes very few restrictions on those projects which wish to use it.

    Here [kitware.com] are some example dartboards.

    Here [ohio-state.edu] is my dart board for the Mobius Project [projectmobius.org] I work on.
  • Interesting (Score:2, Informative)

    by Anonymous Coward
    I like this book and the review is good, but take the review with a grain of salt. The reviewer is a Thought Works consultant. The book spends some solid time on CruiseControl which is a Thought Works donation to the open source community. That he likes it isn't too much of a surprise - particularly since the book, in fact, is quite good.
    • This is a fascinating comment.

      It starts by suggesting the good review is supect because I have a vested interest in CruiseControl, and then goes on to say the review is good because the book is good.

      For the record, yes I work for TW. I don't however have a vested interest in Cruise. I joined TW well after Cruise became an open source project soI don't view it any differently than ANT - it's an OSS tool I use, rather than a TW tool.

      Again, for the record, I'd never really set a cruise intance up before a

I've finally learned what "upward compatible" means. It means we get to keep all our old mistakes. -- Dennie van Tassel