Apologies to all of you for the delay in book reviews – however, the office movage left me without a scanner for some time…
Ever wondered how a dating web site manages to generate automatic “matches”? Ever wanted to find out how to group elements programatically? Ever looked for ways to predict stuff? If you answered any of the questions above with a Yes, Toby Segaran’s book should be for you. But can it stack up?
Programming Collective Intelligence is subdivided into multiple, loosely connected chapters. usually, you can start reading any part of the book without issues – this is definitely not a structured text book.
Toby Segaran did a good job with the choice of topics – things like classification, matching and optimization are covered in significant detail. However, the book also covers more “exotic” topics like genetic programming: in case anyone of you ever wanted to write a “self-modifying” program, this definitely is very interesting.
Each subject is treated thoroughly, although useful things sometimes are spread out upon multiple, not-connected chapters of the book. Generally, the tome presents at least two different approaches for each problem – it is very likely that one of the algorithms given is well-suited for your application idea.
An appendix covers all algorithms in a somewhat structured fashion – I am puzzled why the book itself was not arranged in a similar form.
The author did an excellent job describing the different algorithms and problems and even added graphs where necessary. Unfortunately, non-Python folk are hit hard with understanding the examples(folks – this is NOT a C/Java or C/VB style difference).
For me, the Python examples manage to completely and effortlessly destroy an otherwise well-written book. Yes, there are many people who can do Python, but (especially in the Palm OS world), there are many who cannot. Of course, I can always reconstruct the algorithm if I need to…but this is an unnecessary step that the author could have spared his readers. People with a faible/need for the algorithms and problems covered in this tome (and who have no problems with Python) will find the 27$ a good investment – everybody else is recommended to stay far away…
Related posts:
Content, RSS
Twitter
I don’t know, Python is the ideal language for this book. It is basically a higher level C type language which allows complex data structures and manipulation to be implemented quickly and easily. Put it this way, if you have trouble picking up Python is an hour or two you are going to have trouble with any software engineering in general.
Hi Allen,
thank you so much for talking back!
The problem is not that I can’t learn Python. The problem is that I wanted to learn about algorithms, not about Python.
Also, the book doesn’t contain a Python intro – assuming that everyone in the world is a Python head is – um – …
Just my 2 cents
Tam Hanna
Hi Tam,
Actually having read a few pages of the book I think you’re right. While Python is an easy language to learn describing the algorithms in Python syntax is a bit silly. A better strategy would have been to describe the algorithm in basic pseudo code and then have a CD or download containing Python code which the user could play around with.
Hi allen,
.
thank you so much for talking back! I am happy that you agree with my opinion
Best regards
Tam Hanna