What are the limits of Python? [closed]

Answer 1

They're not used for the same kinds of things, in general. Python is very useful for prototyping since it is so simple, but isn't always the best choice when speed is an important factor, since it is an interpreted language. C++ can be heavily optimized, but as you said might be much more difficult to write and understand.

If you want to do systems or applications programming, you should learn C++. If you are programming for fun and leisure, or to solve problems like at http://projecteuler.net/ then Python is a good choice too.

I personally also recommend learning C/C++ first if you want to get very serious about programming, because it teaches you the important concept of memory management, which might be overlooked in managed languages.

Answer 2

From what I've been told, 1% of learning C++ is learning C. 1% is learning the extra basic features. 98% is learning to use the features in a safe, maintainable way, and coping with the dark hairy corners of the language.

Learning python will teach you to write code that is safe, and maintainable. I think that if you learn python then go back to C++, then you will be able to write good C++ code. Of course, that won't mean you will understand bad C++, or C++ code that was written in a non-pythonic way.

Limits to python?

• It's interpreted, so you have to ship the source and the interpreter; and processes will take much longer to start up.
• It's not C++, so it won't play with existing C++ code.
• It's a bit slower (even if you wrap the hot loops in C).*
• It encourages you to be "pythonic", and some problems are easier if you are not "pythonic".

*Python might be faster:

• Automatic GC. C++ is only faster if it doesn't leak too much.
• Dictionaries. Lots of code runs in O(N plus a bit), rather than O(N^2) if you use a dictionary. Sure, you can use a hash table in C++, but not everbody does.
• Memory management - the python interpreter caches some of the basic data structures' memory, then reallocates them, rather than hitting the system for new memory. This reduces system calls, which is a very good thing.
• Profiling new algorithms is waaaay easier on python. In lots of problems, a better algorithm is more important than a linear speedup (which is what C++ gives you).
• If you are making a program that "only runs once" (scientific analysis, data migration, etc), then the compile-build-test cycle should be faster in python. That's what really counts ;)

Answer 3

One thing that people often use C for, and never (to my knowledge) Python, is low-level code such as operating-system kernels and embedded software.

C has a lot of of constructs that make it very easy to, for example, convert an arbitrary machine address into a pointer and dereference it, or tell the compiler that an address used for memory-mapped I/O might change even though this program doesn’t change it, or to specify the exact layout of an object in memory. It’s designed to run as fast as possible, with as little wasted memory, rather than be safe.

That’s what the comments about “letting you shoot yourself in the foot” are mostly referring to.

Answer 4

if you are trying to find out whether you are still going to be employed some time later using C++ or Python, don't concern yourself with a single language's longevity. Learn to program. Don't learn to program in <insert a language here>.

Here is an analogy: If your car is running fine (gets you where you're going, has good mileage, cheap to maintain, relatively safe), there is no logical reason to trade it in for another one. None. Whatsoever. Drive it to the ground before you even consider what make or model to get next. But if you are already looking around thinking what car to get, just go get it, stop asking everybody you know whether you should do it. If you need to ask, you need to change it. It's as simple as that.

Answer 5

Why don't you ask the converse question? Unlike C++, Python can give you antigravity and summon souls via its import command. On the other hand, C++'s 'equivalent' -- #include -- only allows you to get some boring I/O and math libraries.

Seriously though.. C++ allows you to do more low-level stuff e.g. kernel programming, and allows you to write programs that run much faster (approximately ~20x). You can use it to create real threads that can take advantage of multiple-cored processors, while Python (due to its design) can generally only run its simulated threads on a single core.

IMHO you should learn both; Python for web development and quick-and-dirty scripts, C++ to write systems code, desktop applications, as well as to have a better low-level understanding of the computer. If you're just starting out, then Python; it's much easier to begin with.

More concrete measurements on the speed difference can be found here.

Edit: Seems like my information on multi-threading is out of date; see John Paulett's comment.

Answer 6

One significant difference not so far mentioned is the difference between a language like C++ that builds to native code, and a language like Python which by default puts a VM between you and the hardware. For doing low-level work, like coding against the OS kernel, the native language will be the preferred option.

In practise, though, when you're working in that context it usually means dropping all the way down C (in its role as portable assembler) rather than being able to use C++ (and its runtime libraries), for much if not all of the code.