C++ books that should remain in the past
I tend to read a ton of books just to forget most of the valuable information. Here are a few books I've recently dug out of my basement. I'm pretty sure they should remain in the basement.
BartVandewoestyneIf you ever wanted 42 reasons why not to use C++, then "Effective Modern C++" has all 42 of them. Only parents are worse at advertising parenthood and I wonder why C++ lets you do so many stupid things. Then again if you are not a parent nor C++ developers then it is hard to explain the joy of it - although there are notions of self-harm in both. I've read it several times and I cannot recommend it.
Next on my "not so useful anymore" is "Modern C++ Design" and it doesn't feel modern after 20 years. I must say that I'm not a template fanboy, there are other ways to do it - like code generation for better performance and a lot of the design problems can be solved with functional programming paradigms. The books had a well deserved place in any curious C++ developer.
It's almost unfair for me to review this book, I've read the first chapter and some of the last chapters. It lost when "loop unrolling" was mentioned as a performance optimization techniques - really? In 2022? The table of contents looked really promising. I would rather recommend "Optimized C++".
I guess not much needs to be said about Microsofts C++ AMP now when they have deprecated the header files in Visual Studio 2022. "Maybe next time".
And lastly Intel TBB, highly promising parallelization library. It can do mostly what C++ Concurrency can do for you. Someone at Intel knew 15 years ago what the future challenges developers would face with multi-core. Not sure TBB fixes any of them, but it is also surprising given the time we had to prepare how unprepared we are to this date.