I'm currently reading Code Complete by Steve McConnell.
So far I'm not very impressed. I'm rather disappointed given the rave reviews and recommendations. He says a lot of sensible stuff, but it's all very basic to an experienced programmer.
There are a few howlers, like claiming Perl and Basic are acronyms.
What annoys me though is his repeated references to 7±2, the supposed magic number of items that the human mind can keep in short-term memory. Although this is a common belief, it is based on outdated science carried out when we did not understand the chunking mechanisms of memory. We now know that it is more like 4, or a bit less.
In McConnell's partial defence, the 7±2 figure is widely believed (and is more prominently mentioned on Wikipedia - at least until I change it!) However, it shows a lack of research for something he relies on so fundamentally.
Near the start of Chapter 7, McConnell uses the 7±2 figure to give an upper limit for the number of parameters to a function as 7. 7! That's hopelessly unmanageable and the sign of a very poorly written function. If he had used the correct(ish) figure of 4, or a bit less, I'd agree with him.
The main saving grace of the book is that he writes quite engagingly but new stuff is pretty sparse. I'm only 1/5th of the way through the book and I'm not sure if I'll make it to the end.
So far I'm not very impressed. I'm rather disappointed given the rave reviews and recommendations. He says a lot of sensible stuff, but it's all very basic to an experienced programmer.
There are a few howlers, like claiming Perl and Basic are acronyms.
What annoys me though is his repeated references to 7±2, the supposed magic number of items that the human mind can keep in short-term memory. Although this is a common belief, it is based on outdated science carried out when we did not understand the chunking mechanisms of memory. We now know that it is more like 4, or a bit less.
In McConnell's partial defence, the 7±2 figure is widely believed (and is more prominently mentioned on Wikipedia - at least until I change it!) However, it shows a lack of research for something he relies on so fundamentally.
Near the start of Chapter 7, McConnell uses the 7±2 figure to give an upper limit for the number of parameters to a function as 7. 7! That's hopelessly unmanageable and the sign of a very poorly written function. If he had used the correct(ish) figure of 4, or a bit less, I'd agree with him.
The main saving grace of the book is that he writes quite engagingly but new stuff is pretty sparse. I'm only 1/5th of the way through the book and I'm not sure if I'll make it to the end.
No comments:
Post a Comment