Book review: IBM Rational ClearCase 7.0: Master the Tools that Monitor, Analyze, and Manage Software Configurations

The book: IBM Rational ClearCase 7.0: Master the Tools that Monitor, Analyze, and Manage Software Configurations by Marc Girod and Tatania Shpichko Published by PACKT Publishing ISBN 978-1-849680-12-7 I am reviewing the eBook (PDF download) version of the book. Reading this book will not make you a ClearCase administrator, or user for that matter. It [...]

The book:

IBM Rational ClearCase 7.0: Master the Tools that Monitor, Analyze, and Manage Software Configurations

by Marc Girod and Tatania Shpichko

Published by PACKT Publishing

ISBN 978-1-849680-12-7

I am reviewing the eBook (PDF download) version of the book.

Reading this book will not make you a ClearCase administrator, or user for that matter. It is mainly a manifesto for the authors’ views on why base ClearCase is the SCM tool and how you should be using it.

This book is part technical presentation, part dogmatic manifesto. On the one hand the authors have evidently found a method of using ClearCase that works for them, on  the other they are fanatical in their approach, brooking no descent from their chosen methodology. Any other approach to SCM or use of ClearCase is simply wrong; a position repeatedly and unambiguously emphasised throughout the book.

This is not an easy book to read. Not because the subject matter is particularly sophisticated but because the prose is often difficult to decode. Throughout the book you will encounter things like the following two samples (chosen pretty much at random):

The opacity of tools, as well as that of sources, is of course only a matter of viewpoint: the boundaries are only surfaces of separation between realms of distinct coherence.

Attributes are the preferred alternative to comments. They are preferred because their typing allows to avoid depending on parsing informal text.

Unfortunately this abstruse prose style means the authors’ ideas are often obscured. They really make the reader work hard to figure out what is being said. The book would, in short, have benefit immensely from being professionally edited by a native English speaker.

Another feature that I would have found immensely helpful would be a few well chosen diagrams. The authors are big advocates of the command line, but extending this bias into a book intended to educate, leaves readers with the additional cognitive task of imaging many of the situations based on text output where a simple diagram would make everything immediately obvious. Imagining the situation being described is not always a simple task when the output is from unfamiliar tools (and given the authors make extensive use of some Perl utilities there is plenty for even existing ClearCase users to find unfamiliar).

That said, the book is not without merit and does contain some interesting ideas (if only one takes the time to wade through the authors’ style).

The book does contain plenty of examples and these are used to discuss the application of the authors’ approach. The authors’ evidently use the approach that they advocate, so this is no theoretical approach but one that obviously works for them.

So, who would benefit from reading this book? With the caveat that any serious practitioner should read everything they can: if you want to use base ClearCase (no UCM, the authors are very dismissive of this innovation) and you are content to follow the prescriptions of the author, then this book may be of value to you.

If you are looking for a book that will teach you about ClearCase or a book showing anything other than base ClearCase, then you will have to look elsewhere. Multisite is dealt with, but mostly to highlight the difficulties in using it—this in itself is quite useful information.

The authors principal concern is with the leveraging of ClearCase’s derived objects and ClearMake system to make builds more efficient (although there is nodding acceptance that the advantages are less apparent when using systems such a Java with its compiler’s built in dependency management).

I can see some benefits to the authors’ derived object maintenance approach. However, there are many simpler, and better, ways to improve build reliability and performance (proper system decomposition and leveraging parallel builds, for example). Anyway, this is a review of the book, not a critique of the approach, so I shall defer my observations on the correctness of the authors assertions about the superiority of managing derived objects and simply say that the book contained no substantial argument that I found persuasive in favour of their approach (although it does contain much assertion of the superiority of the approach).

The approach to version control described in the book is fundamentally quite simple, but the explaination makes this simple system appear more complex than necessary. In summary, the proposed approach is, in essence, branch per change. The “novel” feature, according to the authors’, is that they promote the idea that there is no need to merge changes back into an integration branch, preferring instead to use labels to pick up versions from the extant fix branches. I often found myself having difficulty seeing the real benefits of this system, or indeed any significant benefit of the system over more widely used approached. The main benefit seems to be that this approach maintains identities within ClearCase, making the authors derived object maintenance objective feasible.

The authors attempt to dismiss alternative approaches (such as use of integration branches), but I found their explanations of the weaknesses in these approaches seemed most often to be based on misunderstandings of the mechanics of these alternative approaches rather than fundamental problems. (Although I defer to their superior knowledge of ClearCase’s shortcomings when they make specific comments on the problems with these methods when using ClearCase.) Perhaps I misunderstood their case. It was honestly difficult to tell sometimes simply because of the writing style and dismissive tone in some sections.

This is not a book with wide appeal, and this may explain the high price. Is it worth the £20.39 (eBook price), £30.59 (printed book price)?

At about 320 pages (excluding the index and front-matter) the book is brief enough to read comfortably on a wet Sunday. Unless you are specifically looking for a prescriptive method of applying base ClearCase to your organisation I would have to say, ‘no, this book is not worth the purchase price’. There is simply not sufficient content of general interest to the SCM practitioner, or even of interest to those looking for a general ClearCase book.

Trackback(0)

TrackBack URI for this entry

Comments (0)

Subscribe to this comment's feed

Write comment