Performance and multi-threading are niche issues in programming.
Why does Intel produce multi core CPUs then? Yes, you can use multi processing instead of multi threading but the cost of inter-process communication (i.e. serialization and memory copying) is much higher than the cost of thread coordination.
The vast majority of code today is web and business applications, which in most cases don't need cutting edge performance or multi-threading. There are some that do, of course, and in those cases, you might be better served with other languages (not Java usually either, which itself is not very fast), but for most programmers, Python is just fine.
Business applications are typically hosted in cloud and therefore hosting costs are proportional to resource consumption. Simple business apps are constrained by database performance so the choice of programming language may not be significant in many cases. As soon as you need to do anything more complex than storing and querying data in a DB, the choice of programming language directly translates to hosting costs. This will only get worse with increasing energy costs. And Python is the slowest mainstream language.
There are also studies indicating that website response time has direct impact on user retention (for example Google experimenting with the number of results in a single page, which in turn had impact on processing time).
Your claim that Java is slow is simply not true. Java is among the fastest JIT based languages. C/C++ are faster in some applications, mostly number crunching, but typically by not more 30%. Then again, Python is 2 orders of magnitude slower than C++ or Java in number crunching applications.
On Linux Mono is in fact slower than Java. I do not have opinion on .NET vs Java performance on Windows, but hardly anyone uses Windows for server side programming.
As far as stuff like strong typing, Python has annotations for that if you really need it. So your manager could enforce the team using annotations to "strong type" arguments. But ultimately, Python gives the programmer a ton of freedom, flexibility, and power, which means good programmers can make the most of it, and bad programmers can drown in their own mistakes.
Python does not even support multiline lambdas, unlike all other mainstream programmers.
If you look at mainstream compiled languages: Java (+ all other JVM languages), C#, C++, Go, Rust, all of them have native middleware libraries: database clients, message queue clients, message serialization libraries, math libraries. Python, on the other hand, does not even implement database clients, it just wraps C/C++ libraries. Even stuff like serialization is too slow for Python: Protocol Buffers library for Python has 2 modes: pure Python and C++ wrapper which is recommended for best performance (see here:
https://github.com/protocolbuffers/protobuf/issues/10051). Can you name another language that is too slow for simple binary serialization?
Heck, Python even needs C Numpy code to efficiently handle arrays.
Java is a lot more idiot-proof, as I mentioned in my original post here, and C# even more so, so yes, if you have low quality programmers/young programmers/outsourcing/etc, you would be better off with more restrictive languages, but if you are decent or better, Python will let you fly.
It is the opposite: Python does not even support parallel execution of threads: only one thread can acquire GIL and execute Python bytecode at a time. If you want parallel execution you need to resort to calling native functions (which do not require GIL), like with everything in Python.
Complex projects that process large amounts of data are typically built in Java, Scala, Kotlin, sometimes Haskell (in financial domain). You can therefore hire developers with proper experience. Python on the other hand is typically used by small software houses and interactive agencies as a competitor for PHP and Node.js. If you hire Python developers, on the other hand, you risk working with people whose only experience is building websites for florists. To be fair, there are also super-smart data scientists who use Python as a wrapper for C++ ML and data processing libraries, but they cannot replace programmers.
And if you claim that Python is not good for large projects, I would just list some of the behemoth applications out there using it as the main language (possibly with inserts from other languages where needed): Instagram, Quora, Youtube, DropBox, Pinterest, Spotify, Uber, Reddit, Netflix, Instacart, etc.
According to this page
https://en.wikipedia.org/wiki/Programming_languages_used_in_most_popular_websites Python is rarely used as the primary language. It is more suitable for glue code.
Pintereset, Quora etc. were startups with millions of funding, they could afford not to care about performance.
Uber was originally implemented in Node.js and is now transitioning to Go.
Godot game engine uses a scripting language which is basically exactly like Python but supports multiple threading.
GDScript is interpreted and therefore likely slow. At first glance there is no documentation for embedding GDScript in other applications so it is likely restricted to Godot engine.
GDScript = Python with multi-threading support, so you can just essentially code games in Python.
You have not addressed my points. GDSCript is interpreted so it must be slow and it is not designed for embedding outside of Godot, so you cannot easily use it with different engines.
Minecraft developers got lucky but you cannot deny that 95% of commercial computer games boast much better visual and audio fidelity. Java, being garbage collected, is also a bad fit for action games.
I was making a point that game design is much more important than the language you choose.
A writer is only limited by his/her imagination and can put anything on paper. A game designer, on the other hand, is limited by technology. If it was not the case we would have proper AI and interactivity decades ago.
Game design is way more important than the programming language, code in what you enjoy, vast majority of indie type games don't really need high end performance.
There are games that failed due to poor programming and resulting bugs so technology cannot be dismissed so easily.
Sure, but most games fail due to bad design, so...
In reality, graphics have more impact on financial success than design. I am not saying I am happy with that, but that is the reality.
Edit: fixed nested quotes.
