Tag: Book Review

I believe this is the third time I read Design Patterns: Elements of Reusable Object-Oriented Software  – by Erich Gamma, Ralph Johnson, Richard Helm and John Vlissides. This is the first time I bother to review a book I have read before, but this is such a classic it is hard to avoid.

The Wikipedia definition of a design pattern is a “general reusable solution to a commonly occurring problem within a given context in software design”. The point of the book is cataloging the most generic and useful patterns. Most patterns also have little analysis of details on ways you could implement them.

Every time I read this I appreciate it more, and come up with better ideas on how to improve the design of my classes.

Some people seem to think it is too heavy and technical. Personally I like it, but if you want something lighter, check out Design Patterns Explained: A New Perspective on Object-Oriented Design (which I am stunned to realize I never got around to reviewing), which is a simpler review of design patterns.

Clean Code: A Handbook of Agile Software Craftsmanship, by Robert C. Martin is a book about how to keep your code readable, flexible and naturally bug free.

It reminds me, of course, of Code Complete, which I really liked when I first read it (1997). Clean Code is much weaker on the research side (i.e.: there is a big difference between say “this is bad because I say so” and actually showing references to research that prove that a specific practice causes more bugs).

However, as far as I remember the coverage of OOP in Code Complete was much weaker, and this is where Clean Code shines. There are plenty of interesting and useful ideas on how to clean your classes, when and how to refactor them, and much more. There are also some practical refactorings of  real code, which I found somewhat useful – mostly because I read the book on an iPad, and switching around to keep track of what happened is somewhat painful in comparison to flipping a page.

Other than that, I strongly recommend the book for coders everywhere – just keep in mind that all the samples are in Java! I haven't used Java in years but I could still keep up with everything.

I have just finished looking through (and trying out) the recipes in R Graphics Cookbook, by Winston Chang. First of all, I'm very much a beginner in R, but it was possible to follow the examples.

For those wondering what R is – it is an interesting language for statistical computing and graphics. Usually I am more of a Delphi guy (although I have used many other languages in the past, and still use PHP when adequate), but R has a bunch of features and libraries that makes it easier for that area, and it is also used on a few courses – thus my interest.

One thing I really liked about the book is that the example data comes straight from a package. So you install and download them with a simple command in R when you start the book, and you just have to do a library(gcookbook) each time you use an example in a session. Very nice (vs downloading, setting a folder or using the full name for each file,etc).

The book seems to cover the material well enough, and the discussion section usually present useful options. For example, on the recipe Labeling Points in a Scatter Plot the basic solution is presented, and then it shows how to shift labels so they don't crowd the data points, or how to add only a few relevant labels. The examples I tried almost all worked (except for one that needed an extra library, which only took a install_packages call), which unfortunately is not true for many technical books.

Since my background only include the most common graphs, I also learned about a few other graph types, such as correlation matrix, dendograms, vector fields and choropleth map.

Overall, pretty good.

It is interesting to note that this was a free review copy – fortunate, too, since I had almost bought it a few days before getting an offer to review it.

You can get it directly from O'Reilly or other places. The advantage of O'Reilly is that their e-books are DRM-free (and usually available on mobi, PDF and ePub formats).

Killer Headlines for Web Content, by Nick Osborne is an e-book by WordTracker.

It is more of a report than a book, at 65 pages. This is a problem in the sense that it limits the amount of content it covers, but it also makes it a quick read.

The book proposes to cover a few approaches so that you can write better headlines.

It does have good suggestions – mostly the classic approaches (numbered lists, fear, how to, etc) – but it does feel quite oriented for news and generic articles, rather than business content. It does mention this in the Introduction, and says that these headlines work, and it is your job to fit it to your pages and articles to make them work.

I feel that the approaches were reasonably well covered, and it did give me a few ideas. I might even look at my notes when I'm looking for a good headline.

At the current discounted price – US$19 – I feel that the book is worth its price, but I wouldn't recommend it at its full price.