Bokelskere.no

Beklager copy-paste av engelsk, men dette er fra en omtale jeg gjorde på Goodreads:

First a bit on the Kindle version:

One star is subtracted for the electronic conversion. I fully understand that they needed a fixed layout and font-size in order to keep the original two-column layout work with the code examples without overflowing, but this comes at the cost of losing all the benefits of the layout that Kindle provides. The font also looks a bit weird - possibly a pdf->mobi conversion with OCR. Reading on my regular Kindle is taxing, but it is ok. The screen size of the Kindle DX is probably a far better match, as is my computer screen, but I cannot stand to read using that for long. I managed to read it on my Kindle, but it was annoying. Just a word of warning.

On the content

As to the actual content of the book, I found myself enjoying it (after a while). I have to admit that there were some frustrations/questions that were nagging me for most of the book, but they were dealt with when just a tenth of the book remained. This was due to the author's wish to gradually build up a sense of how one ends up with the final version of the Visitor pattern, but I am unsure of the real value in showing so many less desirable versions of it first.

The style is very reminiscent of The Little Schemer (TLS), a book I have on my shelf, but one I never finished - mostly due to not knowing how to run the code. This book actually starts with a section on how one can test the code, so that is not a problem here. The other difference from TLS is that I read this book with a good working knowledge of Java. Thus I never felt the need to actually test any of the code.

One problem I have with the book is its class hierarchy and the names given to classes. The examples often use food classes and one of them is the shish kebab and its ingredients. There we have a class Shish and several ingredients like Onion, Lamb, etc that subclasses Shish. I find that very weird. I would think that a Shish HAS-A ingredent, not that an ingredient IS-A Shish. If the author had just somewhat refactored the classes to use composition I would be OK with the whole thing. The other examples have the same problem. Other than the issues with how the class hierarchy was composed, the code is quite nice and what you end up with in the end is very elegant code. I believe that the reviewers that most violently object to the code have not been exposes to much advanced OO design before and anything else than getters and setters probably freaks them out. It is not a shining jewel in my book collection, but it is interesting and absolutely worth reading. The last 20% percent of the book makes the initial frustration quite worth it.

In short, I recommend reading it, if just for the experience of the different teaching format