Feb
19 Book Review: The Pragmatic Programmer
Well, I don't expect these to be
comprehensive reviews, but just a few notes about the books I have
finished.
I have begun a regimen where I read a chapter out of a book every day. This keeps a constant velocity, and lets me get through books that have been on my shelf for years. I have enough varying interests that I am reading 4 book chapters/day. This takes some time away from coding right now, but I am learning so much that I accept the cost. Eventually I will get disciplined enough to do my reading in the early morning, before the world awakes. I expect that will be later, though, when the log house is complete and I can stumble over to my office and sit in a comfortable chair.
The Pragmatic Programmer was written in 1999 by Dave Thomas and Andy Hunt. It has stood the test of time, mostly because it avoids too much specific technology and focuses on the principles of effective programming. I can see the influence of this book in the book I'm working on now, "Agile Software Development: Principles, Patterns, and Practices" by Robert Martin. "Pragmatic" has become a code word and has been branded as such in the industry. The "Pragmatic" authors have taken their own advice and started learning a new language every year, which is profitable for them, since they often get to write the book about said new language. They jumped on Ruby fairly early, if I recall. It makes sense: this book has a lot of Perl in it, and Perl is so 20th Century.
Metaprogramming is a big deal for them - getting scripts to write code. This has become standard in the industry, and .Net and Java annotations do this as part of their internal structure. However, that, and DRY (don't repeat yourself) are the big takeaways from this book.
They are fumbling about the agile way, but XP was just being articulated at the time, so things like Unit Testing and short iterations were still somewhat controversial. Now they are not controversial, just not done in practice. Hmmm.
Don't Repeat Yourself is a huge deal. If I could get my code to that level, I'd be ecstatic. Redundancy is a real enemy. It is the one thing that killed EJBs, I think. Spring is just so much more terse. It is also the reason I don't like Hibernate. I have a database schema that already exists; I don't want to repeat it in awkward XML. I also don't want to give a single application power over a database schema - often the data is much longer-lived than the application. So I want to work from the data out. That's what led me to the design for DBDB.
Getting good at an editor! What a good idea! I used to think in Emacs keystrokes, but I've been away from coding for long enough that the Mac way is more natural now. I don't know if that will stay, or if I'll get back to Emacs, which has so much more power available without lifting my fingers off of the keyboard. With this laptop, I have a touchpad not too far away, which makes things somewhat easier, but still, keeping the fingers on the home row is definitely a win overall.
I really appreciate the wisdom in "The Pragmatic Programmer," but I am nowhere near that level yet. I will keep working towards it - it feels like I may have to put this one in a rotation, and read it again in a year or so, to keep the target in front of me.
In the meantime, I have plenty of other books in front of me.
What did you think of "The Pragmatic Programmer?" What, you haven't read it yet? Or are you like I was, and felt so guilty about how you actually work that you never finished it?