Python 2/3 versus Perl 5/6

[Steven D'Aprano]

On Sun, 29 Mar 2015 06:36 am, Mario Figueiredo wrote:

> On Sat, 28 Mar 2015 21:32:31 +1100, Steven D'Aprano
> <steve+comp.lang.python at pearwood.info> wrote:
> 
>>The famous Perl coder Allison Randal writes about why Perl is not dead
>>(it's just pining for the fjords *wink* ) and contrasts the Perl 5/6 split
>>to Python 2/3:
> 
> A shame Allison doesn't frequent these groups. I would have a few
> questions for her.
> 
> Perl 6 is in fact a bit like Python 3. Perl 6 is an attempt to
> recreate the language, addressing (by throwing away or heavily
> changing) all the things that gave an indication the language would
> stiffle and die and bring in new ideas to address the demands of
> modern software design and implementation...
> 
> But that's not the thing that confuses me most about Allison's post.
> The whole conversation about Rakudo not being Perl was.
> 
> Perl 6 is Perl. It is part of the Perl family.

"Perl family" is not the same as Perl.

The Perl documentation is clear that they consider Perl 6 to be not just a
mere new version, but a completely different language:

http://perldoc.perl.org/perlfaq1.html#What-is-Perl-6%3f


My understanding is that Perl 6 is related to Perl 5 in a similar way to
(say) Pascal to Algol, Objective-C to C, or Scheme to Lisp.



> And Perl 6 has been 
> developed exclusively as a language specification. Which means it is
> not a language implementation. Rakudo is one such implementation, for
> the JVM, of the language specification known as Perl 6. Therefore
> Rakudo is Perl 6, which means Rakudo is also Perl.
> 
> I mean, we can all agree Jython is Python. Maybe not CPython, but
> Python. Maybe not pythonic in all its body, but Python. Rakudo is no
> different.

The Perl core developers think it is. I have no reason to disagree with
them.

That's exactly the point that Allison Randal is making: Perl 6 has split
with the Perl community, despite Larry Wall's interest in it. That's not
the case with Python 3. Although take-up of Python 3 has been slow, it has
also been steady, and it is still the same language.


> It's a shame(?) it is gaining no traction (and I think 
> Alisson optimism is either misplaced of whishful thinking). But Rakudo
> is a reminder to everyone about the design principles and motivations
> behind Perl 6. For good or bad, because no other implementation of
> Perl 6 exists yet that can produce working code, which after all this
> years can only mean that Perl 6 is in deep trouble.

I don't know about that. Is Perl 6 in danger of running out of developers?
Could you wipe out the entire Perl 6 development team with one out of
control bus? I don't think so. A language like Cobra is, I think,
maintained by one person. Frink is even worse: not only is it maintained by
a single person, but it isn't even open source, so if he is hit by a bus,
Frink dies. (Which is terribly sad, because Frink is amazing.)

What is the minimum number of users that a language needs to survive long
term? Ten? Ten million? Somewhere in between?

If you define "not in the TIOBE top 10" as "in deep trouble", then sure,
Perl 6 is in deep trouble. But I think languages can survive for decades or
longer in small and not-so-small niches, and continue to be influential
long after the language has ceased to be maintained. There are still people
using PL/I and APL and Tcl/Tk, even if they aren't being used for new major
projects.



-- 
Steven