Comment by wheaties

3 years ago

This is why adding language features needs to be carefully thought out and explored to great lengths. Sadly C++ never got that memo. Here's hoping Python remains "simple."

17 comments

wheaties

Reply
Waterluvian  3 years ago

I’m not sure I understand the “this is why” part. Python wants to let you override almost everything. It’s a feature and a powerful and occasionally helpful one. It’s also a feature you never ever have to know about or touch.

  • jesushax  3 years ago

    This debate could be summed up as:

    There's more than one way to do it *vs* there should be one, and preferably only one, obvious way to do it

    • DangitBobby  3 years ago

      > there should be one, and preferably only one, obvious way to do it

      Loops, list/dict comprehensions, if statements, and ternary expressions would all like a word

      1 reply →

  • jamesfinlayson  3 years ago

    Agreed - I mostly use Python for small scripts and it makes that use case very easy. I know it has a bunch more features for more niche stuff as well but my throwaway script to download images from a webpage doesn't need pattern matching.

  • Blackthorn  3 years ago

    It's like python's metaclasses. You rarely need them but sometimes they really are just the best solution to the problem. Those times, you're really glad they're available.

  • kazinator  3 years ago

    > It’s also a feature you never ever have to know about or touch.

    The "never have to know or touch" argument applies only to the lone hacker working on a completely new project with no inherited legacy code.

    • zqlark  3 years ago

      Indeed. Most Python code bases I have seen are maxing out all obscure features like those in the article. In practice, Python today is one of the most unreadable languages in existence.

pansa2  3 years ago

It’s quite clear from the design of Python, both of recent features and of fundamentals like `+=`, that its designers never got that memo either.

I’m baffled as to how it retains its reputation of being a simple language.

  • shepherdjerred  3 years ago

    I think it’s a pretty simple language for how powerful it is. But it’s not simple compared to some other languages like Go

  • defen  3 years ago

    Wait, what's the deal with `+=` ?

    • pansa2  3 years ago

      `a += b` sometimes does the same thing as `a = a + b`:

          >>> a = b = (1, 2)        >>> a = b = (1, 2)
    >>> a += (3,)             >>> a = a + (3,)
    >>> print(a, b)           >>> print(a, b)
    (1, 2, 3) (1, 2)          (1, 2, 3) (1, 2)

      Sometimes it does something different:

          >>> a = b = [1, 2]        >>> a = b = [1, 2]
    >>> a += [3]              >>> a = a + [3]
    >>> print(a, b)           >>> print(a, b)
    [1, 2, 3] [1, 2, 3]       [1, 2, 3] [1, 2]

      IMO if that behaviour had been "carefully thought out" by the language designers, it should have been obvious that it's a bad idea.

      Failing that, the implementation of that behaviour is convoluted - so the designers should have paid attention to the Zen of Python: "If the implementation is hard to explain, it's a bad idea".

      Failing that, if the behaviour had been "explored to great lengths", the designers would have understood how it interacts with other language features - in particular, nested mutable and immutable objects.

      Python's designers failed to do any of these things, so we've ended up with an operator with unpredictable behaviour and a long FAQ entry [0] about how its possible for an operator to both succeed and fail at the same time:

          >>> a = ([1, 2], 4)
    >>> a[0] += [3]
    Traceback (most recent call last):
      File "<stdin>", line 1, in <module>
    TypeError: 'tuple' object does not support item assignment
    >>> a
    ([1, 2, 3], 4)

      [0] https://docs.python.org/3/faq/programming.html#why-does-a-tu...

    • patrec  3 years ago

      It can mean two completely different things. `let x = x + 1` (rebinding) or `x += 1` (mutation).

          In [5]: def cap(x):  
       ...:     xs = [] 
       ...:     for i in range(3): 
       ...:         x+=1 
       ...:         xs.append(x) 
       ...:     return xs 
       ...:      
       ...:

    In [6]: cap(numpy.array(0))                                                                                                                                          
    Out[6]: [array(3), array(3), array(3)]
    In [7]: cap(0)                                                                                                                                          
    Out[7]: [1, 2, 3]

    • CGamesPlay  3 years ago

      The obvious x = x + y and the also-works x += y. There is more than one way to do the operation. The conceptually more simple way would be the former as it does not require the programmer to know an extra operator.

petters  3 years ago

I like Python, but it really isn’t “simple” (at least not for more than a decade).

pizza  3 years ago

Is it better for Python to be simple than for Python to be simple for simple problems?

pjmlp  3 years ago

All languages got the memo, including C++.

Either you get Scheme, or languages with features.

Even C isnt' as "simple" as people take it to be.