Python and the need for speed

Mikhail V mikhailwas at gmail.com
Mon Apr 10 08:25:16 EDT 2017


On 10 April 2017 at 02:21, Gregory Ewing <greg.ewing at canterbury.ac.nz> wrote:
>
> ....
>
> My take on the idea of making Python less dynamic in order
> to improve speed is that you'll end up with a language that,
> while it may superficially resemble Python, doesn't
> really feel like Python.
>
> Boo is an example of that. It has a Python-like syntax, but
> to get any speed advantage you need to add static type
> delarations, and then it feels more like programming in
> C# than Python. At that point, you wonder whether you might
> just be better off writing your program in C# to begin with.
>
> That's not to say this kind of approach isn't worth pursuing,
> but like the JIT attempts mentioned in the article, it has
> also been tried before, with varying levels of success.

Agree. python is python and I suppose that performance issues
has much to do with types and OOP.
When I first started with python I thought - no, it is not possible
without types, if I'll write something more complicated than
hello world, it will all break at some point.
But hell, this works and works good.

Still I miss some old school features in Python, e.g. "goto" statement would
be very useful in some cases. I know it is considered bad style
to use goto, but in some cases it is just most natural thing to use.

What I am (and probably many people) missing is a good tool for
performance middle- and low-level applications.
For me it would be a coding tool, sort of minimalist IDE, with simple
readable syntax which generates compilable C code.
And it would not be necessarily python-like syntax,
but I tend to agree that for today python's syntax is
most readable.

The problem that many are overlooking still is that the
possibilities for syntaxes are very limited in pure text-mode
presentation.


More information about the Python-list mailing list