Mythbusters – Why I (still) Love Perl

At the very beginning, I should probably make it clear that this post is not a declaration of exclusivity in my relationship to Perl. I love programming languages. I first learned to program about the same time I first learned to read (English) and first studied French. My love for programming languages is very much akin (and I swear linked to the same part of my brain) as my love for human languages: they are all unique and beautiful in their own way. I love Python, I love C, I love Smalltalk, I love Erlang, etc, etc.

But Perl has taken an entirely undeserved beating in recent years, and so, in karmic balance, it deserves a round of outspoken championship, far more than others need right now. In pondering why Perl’s current reputation is so completely disjointed from the reality of the language, I’ve boiled it down to three “Big Bang Theory”-esque ideas: “The Cookie Slap! Effect”, “The Awkward Adolescence Fallacy”, and “The Singularity Paradox”.

The Cookie Slap! Effect

There’s a “Perl is Dead” meme floating around. It’s been around for a while, long enough that it’s picked up steam and plows along despite all evidence to the contrary. It’s so common, I won’t bother providing links, because you’ve seen more of them than I can count. It’s so common, I even heard it recently from a young MBA graduate, who was so entirely non-technical that he didn’t even know what HTML and CSS were, while working in sales for a web startup. But, he knew what Perl was, and he “knew” it was dead. Weird, but that’s how memes work. So, how did this meme start?

There’s a well-known gaming scenario often called “King of the Hill”, where one player has special status, and the goal of all the other players is to knock that player out of the privileged position and take it for themselves. There was a time when Perl owned the web. It was the duct tape that built the “Web 1.0 City”. There are a number of reasons why Perl succeeded so wildly, but most of them boil down to being in the right place, at the right time, with a fresh, dynamic take on what “programming” should be like, and what “programmers” should be like. There are many things that C excels at, but manipulating massive, variable-length strings is certainly not one of them. And, at the end of the day, no matter what abstractions you layer on it, Web development is essentially about dynamically building and pushing out very lengthy strings of HTML and CSS (and Javascript) to browsers that will decide how to render them. C sucked for that. Really sucked. Trust me, I’ve been-there-done-that.

So, Perl dropped into what was effectively an empty space, at a time when the demand for web services was sky-rocketing. Score! And established itself as the King of the Hill. Score! But then, as the dominant player, Perl also became “the one to beat”. Every upstart young programming language compared itself to Perl. “We’re better than Perl because…” And this is where the “Perl is dead” meme started. It was popular to put-down Perl, because Perl was “unbeatable”. Of course, it was never really unbeatable. The chances that any one language would continue to dominate the web are exceedingly tiny, effectively zero. The chances that any one language will ever again achieve the dominance Perl once had are equally tiny. Especially when you consider the fact that diversity is one of the single strongest cultural values of this miraculous, glorious, networked universe we now inhabit. A cultural value that it partially learned, BTW, from Perl’s TMTOWTDI (there’s more than one way to do it).

So, when you hear “Perl is dead”, remember this, the only reason the meme has strength is because Perl itself has strength. No one feels the need to loudly proclaim “Draco is dead”, because really, no one cares. And secondly, remember that at the root, the primary reason for declaring the untruthful “Perl is dead”, rather than the far simpler and truthful “Perl doesn’t have the domain dominance it once had”, is an insecurity about whatever language the speaker happens to love. Perl is an elder-statesman in a closed system. The shocking truth about the Tiobe Index isn’t that Perl has drifted down over the years (languages wax and wane, just look at C), it’s that all of the languages in the top 20 are OLD in technology-years. It’s like looking at Forbes’ list of billionaires from year-to-year: they shift in position, but once someone’s on the list they have an advantage over all the players who aren’t on the list (massive capital to invest in one case, massive numbers of users and lines of working production code in the other), which means they’re likely to stay on the list.

And, if you’ve ever said “Perl is dead”, my advice is to learn a lesson of tolerance, “live and let live”. Perl won’t plow your favorite language under the carpet, it’s not a threat to you, but I also guarantee your favorite language won’t plow Perl under the carpet. The best you can hope for is to be accepted as a member of “The Fellowship of the Languages”, so grant other languages the same respect you’d like to receive from them.

The Awkward Adolescence Fallacy

Around Perl’s 13th birthday, which just happened to also be the fiery heart of the dot-com bust, the child-prodigy suffered from a massive anxiety attack. As the Web 1.0 world went down in flames, the Perl community was quite literally tearing itself apart. There were a variety of factors, layers of complexity as in any human conflict, but one of the key factors was the nagging doubt that creeps into anyone’s head when things aren’t going as well as you’d hoped: “Maybe it’s my fault. Maybe I didn’t deserve success.” That line of thinking is generally not true, or at least wildly exaggerated, and generally not helpful.

One result of this early-life crisis was the birth of the Perl 6 idea, but I’ll punt that to the next section. Another more subtle (but also more powerful) result, was a growing obsession with things not being “good enough”. The calm confidence of Perl’s youth was replaced by fears that the syntax wasn’t good enough, the implementation wasn’t good enough, the community wasn’t good enough, the foundation wasn’t good enough, the license wasn’t good enough…

To a certain extent fear is a healthy thing, it drives you to push harder, to conquer the thing that you fear. And Perl did. The catastrophic flame-wars of the late 90′s were put to rest. The foundation was restructured and strengthened, and is now one of the most professionally run open source foundations I participate in, with solid, steady funding to run projects that are hugely beneficial to Perl. The Artistic 2 license was introduced as an improvement, but even the existing Artistic 1 license proved itself in court, in a way that no other open source license ever has, and in a way that benefited the entire open source community. The syntax, implementation, and libraries of Perl 5 have improved substantially, to the point that working with “Modern Perl” is really a very different experience than the Perl of 10 or 15 years ago, while still retaining the characteristics that make it a joyful language to code in.

But the bad side of that fear was an awkward, shy hesitation. Like a gawky teenager, Perl stood at the side of the room, afraid to dance because people might think he looks funny. So, I’ve got a news flash for you: no language is perfect, no syntax is perfect, no implementation is perfect, no community is perfect, no foundation is perfect, no license is perfect, nothing is perfect. Perl wasn’t perfect when it owned the web. Perfection is not the first step to success, it’s not even a milestone on the path. And if you really want to understand how irrelevant perfection is, pick any random language you admire, that you see as the pinnacle of success, and closely inspect its syntax, implementation, community, foundation, and license. You’ll find it’s flawed. Why? Because we’re all human, and we all produce things that are deeply creative, deeply wonderful, and yes, somewhat flawed.

It’s time for Perl to grow up. It’s not a teenager anymore. It’s time to accept what it is, accept what it isn’t, and walk on. And what it is, is pretty outrageously amazing. I’ve recently had the opportunity to help a wildly successful startup, in a domain that sorely needs the advantages of modern tech. Perl is the right tool for the job. If I explained the problem space you’d agree, even if Perl isn’t your favorite language. Perl is the right tool for a lot of jobs, all over the world, right now, stable and reliable, in production, with massive numbers of lines of code.

The Singularity Paradox

When Perl 6 was announced, it had a wonderful effect on the Perl community. It provided an “event horizon” to focus everyone’s attention on Perl’s future. It was an inspiration for new creativity, and a distraction from the flamewars to help kill them off. But over time, some things happened that we entirely didn’t expect. The most obvious is that it has taken rather longer that we anticipated. I remember a time when “6 months” was a completely reasonable project estimate for the Perl 6 production release. (Not “By Christmas”, but a real, project-planning 6 months, where I could map out what needed to happen each month.) 13 years later, that clearly didn’t happen. But the time factor is actually a side-effect of other things we didn’t anticipate. The single biggest thing we didn’t anticipate is that the “community rewrite of Perl” has, in fact, turned out to be a community fork. Perl 6 is not like Python 3, which really is a continuation of Python 2, with the same developers, same users, and same community values. (Sometime I’ll write about my interest and contributions toward the Python 3 migration effort, with its own unique successes and challenges.) What grew out of the Perl 6 idea is a new community, a new group of developers, and even a new identity, “Rakudo” rather than Perl (with a phase of “Pugs” along the way). The core Perl developers still work on Perl 5, and have little or no interest in Rakudo. Some of the Rakudo developers have a background in Perl, but many of them have a background in PHP, Java, C#, or other languages.

Rakudo is not an “upgrade” from Perl. It’s revolutionary and exciting, just like Perl was in 1987, but it is not Perl. Please note that I’m not commenting on the similarity or difference of syntax between Perl and Rakudo. If you take a long view over the history of programming languages, syntax is about as relevant to the success of a language as the color of the bike shed. And if you really, really get down to the nuts and bolts, the syntax and functionality of Perl, Python, Ruby, PHP, and Lua are all fundamentally quite similar. That doesn’t make them the same language, and more importantly it doesn’t make them the same community.

So, we stepped into Perl 6 expecting the full power of the mighty Perl community pushing it forward. What we actually got is a tiny band of free-thinkers, re-imagining what “programming” should be like, and what “programmers” should be like. That’s not a bad thing. As new languages go, Rakudo is among the most exciting. But, it’s in that thinly-stretched startup mode where you only get to pick one of “quick, cheap, or good” and it’s optimizing for “good”. In the long-run, that focus will be crucial to Rakudo’s success.

Back to the impact on Perl. Ultimately, the wonderful distraction of Perl 6 has proven… well.. distracting. What was once a very good thing for Perl, is paradoxically now bad for Perl. I recently explained this to a friend as a story of two brothers, Perl and Rakudo Wall:


Perl Wall has finished his advanced graduate degree, and is out building
his career. He was hugely successful for a while, but lately something
strange has been happening. When he goes on interviews, for some reason
people keep pulling up his younger brother’s resume by mistake, and then
tell him “Sorry kid, you don’t have the experience for this job”. But
really, he’s perfect for the job, if only they’d look at *his* resume,
instead of looking at his brother’s.

Rakudo Wall is still a teenager, and walks to the tune of a different
drummer. He’s smart, but he does things his own way. Sometimes
he takes a little longer than the other kids, and sometimes he
leap-frogs past them with a brilliant insight even the teachers don’t
understand. People keep telling him that he should be just like his
older brother. But he’s not, and he doesn’t *want* to be exactly like
his brother. He wants to be himself. Someday, he’ll be awesome, even
outshine his brother. But he’ll get there in his own time, and his own way.

Right now, Perl and Rakudo are getting in each other’s way. They’re like conjoined twins, trying to live separate lives, but always anchored to their brother. That doesn’t mean I love Rakudo any less than I love Perl. I love them both, and want them both to succeed. But their paths are very, very different, and they each need the freedom to walk their own path. The way to grant that freedom is stunningly simple: accept that it is what it is, and let each go its own way, with its own chosen identity. Let Perl be Perl and let Rakudo be Rakudo.

I sincerely hope to see Perl 7 released quite soon. No fuss, no bother, no long list of “blocking features”. Just BAM! ship the next version of Perl (5) as Perl 7. And I sincerely hope the greatest success for Rakudo. I don’t even care if it takes another 13 years to release, it’ll be worth the wait.

The King is dead. Long live the King!

8 thoughts on “Mythbusters – Why I (still) Love Perl

  1. As a 13-year Perl programmer I started with the language around the time the Perl 6 hype was beginning. As with Python, I wasn’t sure what to pursue: the way of the future or the tried and true. Fortunately I stuck with Perl 5, and after 13 years of everyday use, it has been the right language for the problem domains I have found myself in countless times. Yes, I have dabbled in .NET and Python, and they have their places, but Perl has been my go-to language for a long time and will be for the foreseeable future.

    I find the situation Python is in to be very similar to Perl’s, but I note well your observation that the 3.x path really is intended as a continuation of the 2.x path, where Perl 6 turned out to be something other than a continuation of Perl 5. I struggle similarly with trying to decide which way to go with Python, and I have to admit that my experience with Perl has biased me towards the Python 2.x path. I may have to re-examine this someday before long.

    Above all: your suggestion of Perl 7 being a release along the Perl 5 path is fantastic! That needs to happen, and it needs to happen soon. Yesterday! I bow to whoever came up with that, they’re much more perceptive than I; it absolutely would help solve the “Perl is dead” conundrum. I hope it happens.

  2. Marvellous. I never understood the perl bashing. I script in it all the time – I still use my faithful Perl5 CD burning script (calling cdparanoia and cdrecord) in favour over the hoggy and buggy brasero that has infested modern Linux.

  3. “So, I’ve got a news flash for you: no language is perfect, no syntax is perfect, no implementation is perfect, no community is perfect, no foundation is perfect, no license is perfect, nothing is perfect.”

    – Mental Note: “Throw together a new language and name it ‘Perfect’. Then I can use play on words to make bold statements like ‘This language is perfect!’, ‘This language uses perfect syntax’, and ‘Get the software with a perfect language implementation’”
    =)

  4. I wrote a blog post for Perl’s 25th anniversary (link below) about how Perl is Free Software’s COBOL, and that it’s not a bad thing to be. I think the Perl community too often keeps chasing the dragon of Perl’s former popularity. This is where Allison’s King of the Hill point is really apt.

    What if Perl had not ever been huge in the way it was in the late 1990s, but had slowly risen from in the last 25 years, to a peak level that we had today? We’d call it a pretty successful minor programming language community and no one would say it’s dead.

    Perl advocates should let go of the pain of not having the “good old days” anymore, and be happy that the language and a user base survives.

    I once talked with Drizzle founder Brian Aker who told me that Drizzle was still small because it only had about ten thousand users. He was comparing it to MySQL of course, and in his mind MySQL’s ultimate success will always be the yardstick of success for a database project. But, I know plenty of Free Software projects that would be delighted at the idea that ten thousand people used their code! I’d easily bet that there are more than ten thousand people writing new code in Perl every day. Shouldn’t the Perl community be happy with that success?

    http://ebb.org/bkuhn/blog/2012/12/18/perl-cobol.html

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s