>Dev: How many eggs do you put in the cake?
>GF: 3 or 4
>Dev: Ok, I'll put 7.

Help.

I'm a hardware guy. If I do software, it's bare-metal (almost always). I need to fully understand my build system and tweak it exactly to my needs. I'm the sorta guy that needs memory alignment and bitwise operations on a daily basis. I'm always cautious about processor cycles, memory allocation, and power consumption. I think twice if I really need to use a float there and I consider exactly what cost the abstraction layers I build come at.

I had done some web design and development, but that was back in the day when you knew all the workarounds for IE 5-7 by heart and when people were disappointed there wasn't going to be a XHTML 2.0. I didn't build anything large until recently.

Since that time, a lot has happened. Web development has evolved in a way I didn't really fancy, to say the least. Client-side rendering for everything the server could easily do? Of course. Wasting precious energy on mobile devices because it works well enough? Naturally. Solving the simplest problems with a gigantic mess of dependencies you don't even bother to inspect? Well, how else are you going to handle all your sensitive data?

I was going to compare this to the Arduino culture of using modules you don't understand in code you don't understand. But then again, you don't see consumer products or customer-specific electronics powered by an Arduino (at least not that I'm aware of).

I'm just not fit for that shooting-drills-at-walls methodology for getting holes. I'm not against neither easy nor pretty-to-look-at solutions, but it just comes across as wasteful for me nowadays.

So, after my hiatus from web development, I've now been in a sort of internet platform project for a few months. I'm now directly confronted with all that you guys love and hate, frontend frameworks and Node for the backend and whatever. I deliberately didn't voice my opinion when the stack was chosen, because I didn't want to interfere with the modern ways and instead get some experience out of it (and I am).

And now, I'm slowly starting to feel like it was OKAY to work like this.

Hey, fellow dRs.
I may need some help regarding this circuit.

It is about an 8 Bit multifunctional shift register with a 2 Bit instruction set.

These work:
If "Sel" is 0x00, the circuit loads data into its 8 D flip flops.
If "Sel" is 0x03, the circuit replaces all data with 0s.

These don't work (they do, but they do it in a very strange way and end up filling 1s instead):
If "Sel" is 0x01, it is supposed to ringshift bitwise to the right.
If "Sel" is 0x02, it is supposed to ringshift to the left.

As mentioned above. They do it in a very strange way. If we have 10000000, it will shift to the left or right depending on the "Sel"'s binary value. Let us say we want to shift it right. The output will turn from 10000000 to 01000000 (which is what we want), BUT after that it adds another 1 to it: 01100000. It keeps doing it until every single Bit is a 1. Same thing happens with the left ringshift.

I will include the other circuits used in this circuit in the comment section.

What’s one thing (big or small) you still don’t know or understand in software development and still don’t give a fuck.

Mine: Bitwise operation

Just had a thought: Instead of LLVM modeling and optimizing an IR and then backends having to optimize again for actual machine code lowering, wouldn't it be possible to unite both under one unified system?

If you model everything as one huge and complex state machine with a bunch of predefined "micro ops", couldn't you write an optimizer which lowers to the mathmatical presentation of the target platform's instructions?

I.e. the actual identities of the instructions don't matter. What matters is that the input ir is `(x + 3) & 0xff` and the optimizer tries to fit a sequence of instructions to that so that it "solves the system". It doesn't know x86 `andb`; it knows that `andb` takes an input, maybe truncates it, does a bitwise or, and stores the output into a reg

That way you wouldn't have to write complex target dependent backends. Just declare the sequence of actiosn each instruction does and llvm would automatically be able to produce very high quality machine code

I think there's a phd worth of research here but helllll no I'm not touching compilers again lol

Why do bitwise operators (and, or) have a lower precedence than equals?!?! What is the reason?😨

I have checked it in Java and JS...

Need opinions: When your knowledgeable colleague backend-developer chooses 1,2,4,8,16 as enum values instead of 1,2,3,4,5 (for roles associated with permissions, which may be cumulatable) in order to be able to do bitwise operations, is it a sound decision for this scenario? Is it a best practice, just as good, or pedantic?

I want to master bitwise but have a hard time grasping such operations as quickly as logical ones.

[ADVICE NEEDED]
I'm just going to graduate, and I got a job as S/W engineer(trainee) in a small (500 odd employees) company, which uses salesforce, SAP and sharepoint technologies! They are most probably gonna put me in salesforce or SAP. Is it good enough for `me`(read my background), I'm kind of confused, should I go for higher studies?

BACKGROUND:
very average student, but swift at learning technologies, never really got interested in competitive (otherwise I had a real good chance in top companies), I kind of have good IT skills - proficient in python and angularJS, but recently I have got into ML and done some projects!
Okay here's the part, I know it's important for a fresher to be good in data structures, I'm indeed good in parts which I have used! I haven't used AVL tree in any of my projects so I don't know, nor I have ever used bitwise ops!
I think I want to get into roots of ML (some people say I'm fickle but IDC), I think if I take the above job I may loose my interests or may not have time, Please advice.
(sorry for the tags but I need advice from people for all these fields)