Hacker News Reader

I think this theory is too simple. For example, I find both well-written Perl and Haskell to be beautiful.

It's a bit like saying some people only like minimalist writers and some people only like maximalist writers. But a discerning reader is interested in quality over the style. That is, they recognize that there is good maximalism and good minimalism, and appreciate both.

And that's true of reading, but notice that you refered to "minimalist writers" and "maximalist writers" rather than minimal and maximal works. Because most people do choose one over the other when creating rather than consuming.

I disagree with the suggestion that Perl is not built with a consistent philosophy. That philosophy isn't "build from a minimal set of orthogonal features", but it's a coherent design philosophy:

http://modernperlbooks.com/books/modern_perl_2016/01-perl-ph...

Many (but not all) of the questionable decisions in Perl come from its origins in Unix philosophy and the desire to provide a comfortable migration path for system administrators away from command-line tools such as awk and shell. Where Perl's struggled in recent years to borrow features from Rakudo, it's because those features assume a different underlying philosophy.

I'd also suggest the purpose matters. What languages I like to use for my own personal work vs. what I want for professional work are quite different.

Perl 6 may have a theoretically more consistent philosophy behind it than Perl 5, but as that philosophy would likely lead to even more complicated code than Perl 5 already does, I would live in fear of anyone at work adopting it for code I then have to maintain. Except that there doesn't seem to be any chance of that, precisely because everyone else around me feels the same way.

People make the same criticisms about Scala, but many people use Scala in production. My point is that we're talking about opinion here, not fact.

I hope it's permissible to contribute to an explicitly "opinion" subthread. Given that premise, here's my 2c:

I'm not a lover of Scala, in fact I tried it and gave up on it because I didn't like its ratio of cleverness to simplicity (again, a personal opinion). But that said, and also being reasonably familiar with Perl, I would say that Scala is still much cleaner, more consistent, less idiosyncratic than Perl. I've seen some beautiful, expressive, masterful Perl code but I find that Perl really doesn't encourage me to write that way, while Scala works a bit harder to nudge you to write elegant code.

Many things are simpler in Perl 6 than in Perl 5. My guess is you are thinking of the original Perl family.

Absolutely! I hadn't looked at Perl 6 before the OP's presentation.

But isn't Perl 6 a full superset of Perl 5? That would mean the full clutter is still underneath, and some people will continue to code Perl 5 in Perl 6.

> But isn't Perl 6 a full superset of Perl 5?

Nope, it's a new language but with a similar look & feel. Interoperability is currently supported through embedding a Perl5 interpreter.

> isn't Perl 6 a full superset of Perl 5?

That depends on what you mean.

Perl 6 is carefully designed to be able to smoothly use other programming languages and their compilers and/or libraries. The most advanced implementation of this concept is (re)use of Perl 5:

    use Inline::Perl5;
    use DBI:from<Perl5>;
    my $dbh = DBI.connect('dbi:Pg:database=test');
    # etc.

The first line happens to be valid Perl 5 syntax as well as valid Perl 6. But the next two are Perl 6 code that calls Perl 5 code behind the scenes. In principle those who write such Perl 6 code don't need to know any Perl 5 whatsoever. (In practice current API docs assume use of Perl 5 so there's that.)

The same is true for Python code etc. (though the sugar isn't yet as sweet):

    use Inline::Python;
    my $py = Inline::Python.new();
    $py.run('print "hello world"');

So Perl 6 is either a superset of Perl 5, Pythons, Java, Lua, etc. or it's not a superset of any of them, just good at working with them, depending on how you look at things.

It is worth noting that like C++, Scala use in production tends to come with guidelines in terms of what subset of the language people should use.

Scala has additional considerations going for it with its integration into the JVM ecosystem.

That said, may I observe that the hype seems to have died down for Scala. Everyone I personally know who was into it has dropped it. None of them would have shipped anything into production. Anecdotal, sure, but that's all we ever really have. My guess is that its own complexity is indeed strangling it. Only time will tell, but I'm not sure Scala is really that great an argument.

And, also, for consistency, yes, absolutely, I wouldn't want to suddenly be handed a Scala system to professionally maintain either.

The simple theory I use for explaining Perl preference is that it correlates to your preference and tolerance of postmodernism in general. http://www.wall.org/~larry/pm.html is a good elaboration on the postmodernism of Perl, I think.

I started with Perl in the 1990s and well it was the only game in town for the things I needed to get done. Now as a semi-professional programmer (Temporary work and side jobs at work) I have learned to love Racket (Lisp Family), awk/sed and R.

Since I only do this less then full time my brain can't hold everything in it for Perl or similar languages. I can sit down with a Lisp and just work it out with the parts I know and have to look up things far less.

Interesting you mention awk. Do you use that professionally? For string manipulation I use almost exclusively pipe output to bash then format/grok with awk/sed, and wish that more (any?) languages implemented them fully!

I run several awk scripts to format csv files for my reports. awk is a great programming language but I use it for tiny 20 line scripts at most.

  > There are some people who appreciate games that are very abstract and with
  > simple rules from which complex behavior emerges, such as a game like Go.

I love Go, but I despise perl. Perl is not built on "very abstract and simple rules. For instance, this code is quite hard to reason about:

    1: my @to-infinity-and-beyond = 0, 2 … ∞;
    2: say @to-infinity-and-beyond[1008];
    
    # OUTPUT:
    # 2016

This is self-indulgent mucking around. An inference procedure which goes from "0, 2 … ∞" to "the even-nonnegative integers" is bound to lead to unpleasant and hard-to-debug surprises.

The next sentence says:

> These people tend to be drawn towards programming languages that are built out of small pieces with a consistent philosophy from the ground up, such as lisp or Haskell.

If you like go and dislike perl its to his point.

A TLDR (because apparently); People who like games that are complex because of complex rules like perl. People who like complex games with simple rules dislike it.

What you consider "self-indulgent mucking around" others may call "declarative programming". I can appreciate that you don't like it, but I think you've missed my point.

I'm confused. What's so hard to reason about that? It's very obvious that it prints 2016. Maybe "infinite lazy lists" isn't something that most mainstream languages consider, but honestly, they're very handy and hardly "self-indulgent mucking about."

(Though I'd probably use an asterisk instead of the infinity symbol)

Indeed, if we're going to hold up Haskell as an exemplar of simplicity—which, in its philosophy if not always in its implementation, I think it is—then we can hardly ding Perl for including lazy lists!

I have no problem with lazy lists. It's the "0, 2 ... \infty" expression which offends me.

GP was emphatically not saying that Perl 6 is the abstract and simple one. GP was saying those qualities apply to Go and are what make some personalities like Go more than P6. You like the Go qualities and not the P6 ones; QED.

Actually I was saying that those qualities apply to the _board game_ Go, not necessarily to the Go programming language.

He was talking about the GAME Go as compared to a complex game like Chess.