Python Vs C++ Vs Java! 

Good point. We should have briefly mentioned Python bytecode in the vid, probably towards the end. Thanks for pointing that out - it's an important part of how Python works under the hood.

@@ByteByteGo Very grateful for your high quality videos. But this one has an error. It is not at all clear that python has a bytecode and a vm from this video.

I was thinking the same thing... I also had an issue with the video saying "the JVM has a just-in-time compiler" and then in the next breath saying "Java can run on any device without recompilation." So is the JVM "compiling" or not? Further, the idea that a "compiled" language runs faster than an interpreted one is generally true, but not specifically true; the biggest slowdown to some modern languages is due to them not being strongly typed. Python suffers from this (to a degree), as does Perl and many others. But we can see that Python, even though it is JIT compiled (same with Perl), that does not speed the execution to that of C/C++ or even Java. All in all, a decent video here, just not super accurate for the pedants among us.

​@@MikeSwanberg Python is NOT JIT capable, at least not yet

@@theshermantanker7043 Okay, so what do YOU call the process of compiling to bytecode on-the-fly at run-time? I call that compiling that's just in time for execution... or JIT.

And also wrong and nonsense.

What about Jython?) Python frontend that gets compiled to JVM bytecode. Combinations are endless, covering everything will take ages

Yes, unfortunately it is misleading in regards to Python. See comments

What do you mean by "compiler level"?

c# is basically microsoft java, but slightt faster and more memory efficient. And in the end, what's the matter, code in the language you love (or hate if you do javascript)

@@potatoandpippen you are so nice to people, a real role model. Thank you 🙏

Compiled languages also use C bindings (or some other convention) to call library code. No difference there.

I noticed that too 😅 good eye my friend!

The lords language!

that stands for C++++

Freudian slip?

in very high performance machine? no. f35 fighter jet system is created using c++.

This is a sub-5 minute video.

But a process has to be managed by some OS so what's your point?

Python as a backend language for your web services is what I’d go for. Compatibility with your front end is irrelevant, but if you are worried about learning another language and want it to be Javascript-like then neither Java or Python are Javascript like. One could argue that curly braces and certain constructs like for loops look similar in Java, but it’s not really going to help. As for python back end support, you can write a Python web service using Flask or FastAPI in literally seconds once you know what you’re doing. (I don’t use Django, but it paves the way for relatively straightforward REST solutions). For most applications Python is performant enough and besides there are many scaleability options. I have plenty of experience in all these and for my money I’d stick with Python. The commercial user space is vast, as is the support, and as a language it is very easy to learn and get immediate rewards. For other uses apart from backend, well, obviously it has massive AI/data/machine learning tooling and user spaces, and it’s widely used for all kinds of glue and scripting applications. Learn Java too if you like but I think you’d end up having python as your go to language for more than you would Java. If you’re wanting to develop natively for Android, obviously that’s going to want Java (or Kotlin). Good luck. Check out FastAPI and get yourself a 5 line example web application running using uvicorn to serve it (or simply run it with the built in development server) and you’ll quickly see the potential.

c# lets gooooo

For a long time C# was not a valuable option for many developers because it was windows only. So a lot of people decided to opt against targeting one OS only. With other languages like C++ , java and python you can always target the big 4

@@r_v_t yes good point. But languages evolves. We are in different page now. That was a bad decision by microsoft. They are trying to route the ship other way. Its not easy but seems they are in good way.

@@r_v_t The big 4 ? I assume you mean Windows, Linux and MacOS.... what's the fourth ? Android (which is also Linux, but with different userland tools) ?

That’s because of Microsoft and their depreciation of Framework. There was a time where Core wasn’t up to snuff. I’ll never touch C# again

Same😂

The video brought me to this comment.

There is java.lang.ref.SoftReference to mitigate this security issue. It is a way to basically make the GC release the object in the next gc when it has not any reference about it.

cpu:how can I execute instructions slower than IO? (lol)

by the term database response , what I understood is 'querying of the db'. At the end of the day, the programming language you use to write the query command matters isnt?

Unless you are delegating to the DBMS the task of processing the query result (eg using stored procedures) then the language DOES make difference.

databases are usualy created using c++'s father too that is c 😅 like postgresql and mysql, so it already very fast for working with milions of data in just some seconds.

Why

@@potatoandpippen Your comment makes me so happy 😁

Not really. The Java compilation phase is done only once, not every time you run the code.

@@frogzie the python compilation is also done once, not every time you run the code.

C# is C++ times 2

What do you mean directly? Interpretation is also direct

Unfortunately this is a very common misconception. In fact "to compile" just means translating from source code format to execution format. It does not, however, mean to translate into (native) CPU instructions. There are many languages that compile source code into so called bytecode, which then gets executed by an interpreter, also known as a "virtual machine" (VM). For example Java, Python, JavaScript, Ruby all use this approach.

What if the byte codes are optimised in such a way that some parts of those byte codes are directly executable machine codes without the need for interpretation. I think that's what the video means. Bazinga 🙄

@@vijayvijay4123 t Good idea, yet that's not how it works. Bytecode always gets executed by a VM. Some VMs do "hot native compilation" which turns some parts of the bytecode into CPU machine code, making it faster. That's called just in time compilation, JIT for short.

ppl, everybody is ignoring the BIOS layer...

Would it be possible to build a language that depending on the scenario to choose either compiler, interpret, or virtual machine way??? If it is possible and no one do it, then it's not the problem of the programming language any more.

"United programming languages" is like saying "United transportation, I want a car that becomes a submarine whenever I want".

@@markdanielesplanaperilla There are already such kind of cars. It’s just expensive

Technically if you start from scratch each time you only ever write it once.

Probably just a new framework.

Its blazingly fast.

Same question

just guessing, but i assume it has to be the advantages with first compiling to bytecode, where javascript interpreter does not has this step.

JavaScript isn’t strongly typed unlike Java, so it might have to do additional checks which I think could make it slower

What does an interpreter do? Doesn’t it just read your code line-by-line and turn it into machine code? Does this make it like a JIT compiler or is what separates them the idea that interpreters don’t convert the source code into byte code vs. JIT compilers need byte code?

Java is compiled to byte code at the compile time (not execution time, as Python). The byte code is machine code for Java virtual machine - JVM (it's just an abstract computer). And then Java code is run and executed by JVM without additional compilation (almost the same as C++ does, but the code is executed not by raw hardware, but the software called JVM). It's faster than interpreting every line as Python does, but way slower than C++'s approach with pure machine code. At the runtime, JVM detects some places in the program that are executed often (e.g. loops and often-invoked methods) and compiles only those pieces of code to real machine code. The latter process is called Just-in-time (JIT) compilation and it expedites execution up to tens times.

Heard that Python is easier to learn than Java. But also heard that learning Java as a first language will make learning other languages easier. Not sure if I like this last reasoning. For me Java is an older more mature language that also contains more and more contradictions and unnecessary confusions.

I think you shouldn't look at programing language as something "primary". Learning new programming language might take a few months, but if you know another language, with GPT-like tools you can quite easily switch to another one. My main language is Python. I wrote a few very small programs in C++ and java. 2 years ago, it was a struggle - word by word, just like with new language. Today I could write a description and most likely generated code will be working. Today I guess it works mostly for Python, but in a few years these tools will be even better. You should focus on things that will be hard for computers. Unless they gonna replace us completely... Such things are design patterns and overall knowledge like database types, design pattern, where to use what. And search for skills required in your area - the hardest part might be getting the first job. I think AI will replace us first in creating simple web pages, and it'll happen quite fast. So landing job in this field will be hard. You might want to take this into account.

Did you just say that C is slower than Java nowadays ?

​@@Winnetou17yeah 😂

@@Maverick56912 Blasphemy!

Java is statically typed which makes compiling much faster, where python is dynamically typed where the interpreter has to spend much more time deciding types etc

@@Ryan-hk5yb then why isn't JS which is also dynamically typed also a lot more slower Js is slower than Java but not to the extent of python

as he said, python vm is an interpreter, which has to decode every line of code and execute it in real time. being dynamically typed, there is no guarantee what type of data goes into a variable and thus how memory should be handled at runtime. also, there are differences on how functions calls are retrieved and memory is handled at runtime. java is compiled at runtime, meaning that the vm don't need to decode anything, because it executes an assembly. so it takes some time to start, because the vm converts the bytecode into assembly, but when it starts, it's machine code. plus, java is highly optimized, meaning that the vm has some smart strategies to retrieve functions calls (stored in lookup tables) and handles memory very efficiently. the garbage collector is a slowdown though, but memory safety comes at a price.

@@MikiMaki76 what he said. there also differences in use case/engine/runtime environment

Python is crap.