Same with C. 🤣
I’ve learned being productive is the best I’m gonna get. Feeling like I know what’s going on is probably beyond reach.
Discussion
The trick to mastering C is to know an assembler very well. Once you understand the metal, C becomes pretty transparent. That's not true of C++, and is definitely not true of Java, C#, Ruby, Python, JavaScript, Clojure, etc. The farther you get from the metal the more things start to look like magic.
From: TheSameCat<-mazin at 06/16 09:11
> Same with C. 🤣
> I’ve learned being productive is the best I’m gonna get. Feeling like I know what’s going on is probably beyond reach.
What’s funny is the differences in compilers. C++ is my primary language for over 20 years now. I still feel lost 90% of the time (especially with complex templated things).
With C#, I never felt that lost, because Roslyn does such a phenomenal job, snd because C# actually has proper reflection - something C++ still doesnt have (maybe in 26 🤔).
I started on DOS using Turbo C++. I miss those days. Beautiful simplicity to single threaded applications on an OS that did very little. No going back though.
I walked away from C++ twenty years ago just as the template craze started. It felt _wrong_ to me. The more of it I see, the _wronger_ it looks. I think the damage done to the language since then justifies my feelings.
By the same token I hate most of the generic stuff they put into Java and C#. I think simple generics are fine; but for some reason they can't stop themselves from trying to make the generics turing complete in their own right.
I also don't much care for the attributes/properties (@/[]) in Java and C#. I can see _some_ small utility for it (as in Junit) but for the most part I think it does (or at least tempts) more damage than good.
I feel like a lot of the stuff that would've been most useful to the language has not been added, while many things that cover edge cases have not been. The exception are smart pointers and the STL, both of which I whole heartedly love.
What are your thoughts on Zig as a closer-to-metal language? I've been looking at it very closely lately.
From: TheSameCat<-mazin at 06/16 10:00
> I feel like a lot of the stuff that would've been most useful to the language has not been added, while many things that cover edge cases have not been. The exception are smart pointers and the STL, both of which I whole heartedly love.
It was the STL that convinced me to leave the language. I thought we could have done much better. In general I think languages should not evolve much. The first few years should be a period of adjustment and evolution; but after that the language should stabilize. Unfortunately many languages evolve to the point of being top-heavy and useless.
>
> What are your thoughts on Zig as a closer-to-metal language? I've been looking at it very closely lately.
Never heard of it. I'll take a look.