"CPython"

[David J W]

The main motivation for a Python virtual machine in Rust is to strengthen
my knowledge with Rust which currently has some gnarly bits to it but
otherwise is an impressive low level language.   Rust's future is looking
very bright as even Linus Torvalds agrees with most of its design choices
and is allowing it to be used as a linux kernel module language.

Skipping ahead to the subject of names, Rython was chosen because "Python"
is trademarked by the PSF so anything with the complete word Python in it
is out.   A close runner up would have been Camelot but that is already
taken.

Going backward to the issue of use and audience.  Making Rython a real
virtual machine that passes the CPython unit-tests is the only goal.   I am
actively following the faster CPython fork that Mike Shannon, GVR, and
others are working on with the intention to try and incorporate what they
discover into my project but I don't think Rython will be dramatically
faster than Cpython because I am going to implement the same PyObject
reference counting garbage collector and unless faster CPython creates a
JIT component, Rython won't have one either.  Additionally Ryhon won't have
the must have killer libraries like numpy so it's a moot point if my
project turns out to be dramatically faster.

To sum things up, I've been retired for over a decade so I have plenty of
free time.   Initially I thought I might invest time into becoming a core
python developer but looking into it further, all I will say is that
doesn't feel like a very appealing use of my time.



On Thu, Jun 23, 2022 at 9:42 AM Avi Gross <avigross at verizon.net> wrote:

> David,
>
> I am curious why you are undertaking the effort to take a language already
> decades old and showing signs of being a tad rusty into a language
> that suggests further oxidation.
>
> More seriously, I am interested in what this can gain and the intended
> user
> base. I studied Rust for a while and it has it's features but have had no
> opportunity to use it. Is it expected to make a faster version of Python,
> or enable better connections to libraries and so on?
>
> What I mean is that if you are planning on making it pass all tests for
> python functionality, are you also adding unique features or ... ?
>
> My preference is to have names that fully include what they are about.
> So the name "python" would be left intact rather than mangled, even if
> the name itself happens to be totally meaningless. So may I suggest
> something like """rustic-python""" ?
>
>
>
> -----Original Message-----
> From: David J W <ward.davidj at gmail.com>
> To: python-list at python.org
> Sent: Thu, Jun 23, 2022 10:29 am
> Subject: Re: "CPython"
>
> >> Let's say they reimplement "reference python" CPython in Rust. What is
> >> better? Change the "reference python" CPython name to RPython, for
> >> example, or let it as CPython?
>
> >The C implementation would still be called CPython, and the new
> >implementation might be called RPython, or RustyPython, or whatever.
> >The names are independent of which one is currently blessed as the
> >reference implementation.
>
> I am at the pre planning stages of making a Rust implementation of the
> Python virtual machine and to avoid ambiguity I've been working with Rython
> as the name.  I tried looking for a Monty Python themed name but the good
> ones seem to be taken.
>
> Otherwise as for a timeline, solo I figure it's going to take me a couple
> years to get something that actually passes cpython's python unit-tests.
> --
> https://mail.python.org/mailman/listinfo/python-list
>