2016.01.08 Perl 6 is Separate from All, Replacement for None

In response to the "Perl 5 and Perl 6 are mortal enemies" article.

We are definitely framing this whole thing wrong. We have two choices:

Most people outside of Perl 5 and Perl 6 go with Replacement just by looking at the names. Various insiders switch at random between Replacement and Separate, and I think that is the core mistake. This article has a very important sentence:

"First, a postulate: given the language similarities, the people that will find it easiest to learn Perl 6 are today's Perl 5 developers."

In reality there are SO MANY differences between Perl 5 and Perl 6, it is NOT a good postulate!

This is it -- this is the mistake where we are getting both messages at once, and the rest of the article paints a dark picture of a world (unfortunately most likely the one we live in) in which crossing these messages happens. The mixed message is a net negative for both languages.

I suggest that if you take a random [Perl 5, Ruby, Python, PHP, Javascript, R] programmer, the Perl 5 programmer might actually have MORE trouble than others due to misleading overlap and jarring differences. A developer that uses a lot of Moose might have less issues... or again might end up with the differences shocking rather than enlightening. I actually think the Ruby developer will have the LEAST trouble learning Perl 6, building on tons of overlap (everything is an object!), a shared history (sure, /foo/ is a regex, drop it in. Couple sigils here and there never hurt anybody), and similar culture (TIMTOWTDI is a motto for both). But the Ruby developer won't expect things to be mostly the same between Perl 6 and Ruby, unlike a naive Perl 5 developer.

I recommend you each shape your language to do two things. First stick strongly with Separate, second focus on the Polyglot nature and audience:

Perl 6 is a new and interesting language. It happened to be created by the creator of Perl 5 and an overlapping community, and has an unfortunate name. It is a Practical Research Language, aiming to be both cutting edge and useful. It is of interest to all dynamic-language programmers, and is especially interesting for the polyglot: you'll see familiar bits from Haskell, Perl 5, Ruby, Python, CLOS, ... plus lots of original ideas. Things that you can use, or at the very least learn and take back to your work. You can use Perl 6 now to solve small tasks, and join at the ground floor to build up the ecosystem and culture towards solving larger and larger problems.

One of the cultural things that I think can set Perl 6 apart is in being Module Level Polyglot. From Perl 5 or Ruby you can invoke Python code (both have libraries to support this that I've used) -- but you'd be hard pressed to get that put into production at your work. But NOBODY should ever implement pyplot again! If we can make doing "use matplotlib::pyplot:from<Python>" a socially-acceptable thing in Perl 6, we might be able to convince other communities to do the same... and do something amazing: revolutionize sharing cross-language modules.

Perl 6 isn't the research language for Perl 5. It is a research language for ALL languages to learn from. Perl 6 is the Borg of Languages, pulling in concepts and features to create a glorious monster. Maybe the first assimilated was Perl 5, but it clearly didn't stop there. We need to let go of its roots. Embrace that Perl 6 is Separate from All, Replacement for None.


Blog Blog RSS Feed



Follow @awwaiid



Wiki Edits Wiki RSS Feed


... more changes