afongen

Java SE 6 was released last week. How many of us are now running 2 major versions behind? :)

There are a few goodies in this release, including performance gains and better debugging and monitoring, but the one I’ve been waiting for is JSR 223, explicit support for “scripting” languages. I first got excited about this JSR when it seemed that the reference implementation would be PHP. At the time I did most of my work in PHP, and I was excited about bridging the languages. Instead Java 6 ships with Rhino, a JavaScript interpreter written in Java, but that’s just fine if not better. There’s a longer list at scripting.dev.java.net anyway, including PHP on the JVM.

Why does this matter? I believe that the future of Java is not so much Java-the-language as Java-the-platform. I have felt this from the day I first encountered Jython (four years ago already?). Encounters since then with JRuby, Rhino, and Scala have only made my convictions firmer. Recent actions from Sun (and the JCP) lead me to believe that more than a few people there there recognize that if the JVM is one of Java’s core strengths, then the Java platform has a future somewhat distinct from the language. Sun hired two lead JRuby developers (locals Charles Nutter and Thomas Enebo). With JSR 223 and now JSR 292 (“Supporting Dynamically Typed Languages on the Java Platform”), which I believe we can expect in Java 7, it will be taken a significant step further. As Danny Coward points out in a recent interview, JSR 292 introduces the first bytecode in the JVM that is not used by Java. To me, this is a fairly clear endorsement of non-Java languages as part of a broader Java platform.

Microsoft is doing the same thing, by the way. Sure, much was made of multiple languages running on the .NET Common Language Runtime when it was first announced, but I have the distinct impression that C# is very much the canonical language for .NET development. But a few things have happened in the past year that tell me this is changing: Microsoft hired Jim Hugunin, lead developer of IronPython, a Python implementation for .NET (Hugunin is also the creator of Jython), then released IronPython for ASP.NET They’ve also been working with Zend to make the PHP experience better on IIS, including writing FastCGI for IIS 7.

I’m kind of surprised that I haven’t written about this more here. I think I’ve avoided it because it seems so damn obvious. But it isn’t, not really. It may be obvious within the areas of the blogosphere where I spend time and in the local Ruby community. But in the real world of day-to-day Java development to which I subject myself, in which most are unaware even that Sun has open sourced Java (it’s a culture thing), much less that Java 6 has been released and what it includes, talk of languages other than Java is very strange and uncomfortable news. I still get polite nods and bemused or uncomprehending looks. Daily.

It also occurs to me that there are readers of this blog who do not live immersed in the world of Java and who have valid reasons for being unaware of recent events. :-)

A quick note. I put “scripting” in quotes above because labelling languages like Ruby and Python as “scripting languages” is unfair and indicative of the historically dismissive attitude that some programmers have held toward them. To sound au courant, you should know that the currently favored term is “dynamic” or “dynamically typed” to distinguish them from statically typed, early binding languages like Java. The wrinkle is that in JSR 223, those languages are used for scripting, playing second fiddle to Java. JSR 292 shifts this balance.

So there. Nothing earth-shattering, but now at least maybe you understand why I talk about JRuby a lot.

I’ve spent my life immersed in language. By the time I graduated from high school, I’d studied at least a dozen: French, Spanish, Italian, Esperanto, Russian, Mandarin, Irish Gaelic, German, Latin… French is the only one I learned very well, but still I developed a profound appreciation for how study of one language can enrich my understanding of others, including and especially my native tongue, English. In college, as a French major with a linguistics emphasis, I of course studied even more languages. Frequently I found myself better able to express or understand certain ideas in languages other than English.

My father was a programmer and often expressed his conviction that there is a connection between languages used for programming and those for human communication, asserting that they occupy the same area of the brain. I have no idea whether that’s true, but it sounds reasonable and I’d believe it. He, too, had occasion to learn new languages for his work, although the only ones I ever knew about were C, C++, Ada, and possibly COBOL. I think it no accident that the first programming language in which I felt truly at home was Perl, which is strongly influenced by linguistics and ideas about human language and expressiveness (see, for example, Allison Randal’s “On Topic“).

It should come as no surprise that now, as a working programmer, I believe the study of multiple programming languages to be not only extremely valuable, but essential. And maybe it’s close-minded of me, but it surprises and disappoints me when other programmers do not feel the same. I work on a team of Java programmers, and with one exception I am the only one who is even remotely interested in things outside the Java world. This is crippling. That sounds like an exaggeration, but I believe it to be true. A monoculture is limiting and dangerous. It doesn’t matter whether you’re working with .NET, Java, Ruby, or whatever. Working under the assumption that a single tool can be an effective one-size-fits-all tool is a serious mistake. It limits your thinking, it limits your creativity, it limits your ability to solve problems.

Here’s some advice I recently gave someone who was looking to hire a .NET web developer:

I’m not suggesting that you hire someone who isn’t strong in .NET. If you don’t currently have strong .NET skills, then obviously that’s a very important consideration. Neither am I suggesting that you look only at developers with experience in .NET and something else. I’m suggesting that a candidate who has experience with .NET and another platform may — may — turn out to be a stronger candidate because s/he can critically assess what .NET has to offer rather than blindly follow the One True Microsoft Way.

I work with Java programmers. That’s what they are: Java programmers who happen to be working on web apps. Except for maybe something like Scheme covered in an intro computer science course, they don’t have experience with anything but Java. And it hurts us. Because when they look at a problem, all they see is Java. When I suggest a way of doing things that isn’t a canonical approach in the Java world, I get strange, uncomprehending looks. And we end up with overarchitected, overcomplicated application designs that solve problems we don’t have — or that fail to solve the ones we do have.

Example. An complicated call stack for a form submission that retrieves search results. POST form, push criteria to the call stack, redirect to a GET to the results display page, pull criteria off the stack. Okay, the call stack isn’t complicated, but the code behind it was. The programmer who designed this spent a long time working this out and unsuccessfully trying to address problems like dealing with multiple windows. Solution? Submit the form with GET in the first place. Dead obvious to a web developer, not so much to a Java programmer who’s not used to the web. But maybe that’s not fair since it’s a domain problem.

Example. A former coworker insisted that the text output of a certain COBOL program could never be parsed and made into something more meaningful. Why? Because he was exclusively a Java programmer, and Java’s a pain in the ass for text manipulation. Well, I work with languages like Perl and Ruby, steeped in a history of text manipulation. I munge text for breakfast. I looked at the problem and figured it would take about half an hour with a few regular expressions. Java didn’t even have regular expressions in the core API until four years ago, so this way of approaching a problem still doesn’t occur to most Java programmers.

Example. I find myself cleaning up a lot of clunky JavaScript code, written by people who treat it as “Java Lite” and who don’t understand JavaScript’s object model, which is quite different from Java’s. Different, not worse — and certainly not the One True Java Way.

I’m not writing this to rag on Java programmers, it’s just that they happen to be the source of my daily frustrations.

All this said, it’s quite possible to have an excellent web developer who knows only Java, or .NET, or ColdFusion, or PHP, or whatever. Regardless of whether they’ve worked in a diversity of languages or have been steeped in a monoculture, I think it’s worth asking about the strengths and weaknesses are of their chosen platform, because it’s a strong indication of how carefully and how creatively they’re thinking about the problems before them.

This is part of why I worry about what to suggest instead of Java EE, or rather what we can introduce to Java EE to be more productive. Part of my growth as a developer has been learning that not everyone has an easy time learning another language. I have always had an easy time learning languages, whatever the sort. I love to learn a new language, to stretch my mind and see what it has to teach me. This is far from universal. But I have to keep reminding myself of that, especially when I have trouble understanding why so few people are interested in extending the Java platform with non-Java languages.

More on that another day.

A prediction: if a couple years from now I am still mired in a Java monoculture, I will strangle someone. Probably myself.

As I have explained before, and with some apology for the double negative to which I am about to subject you, I do not believe that Java can never be viable for web application development, or that it is a bad language. I simply assert that it is an exceedingly poor choice for the web applications that I work on. Because it’s past 1 a.m. and my son will be waking me up in less than five hours, and maybe because despite my curmudgeonly nature I am reasonably polite after all, I will spare my employer the (mild) embarrassment of an all-out rant. Suffice it to say that working among the Convinced as I do, I am very much in the minority in my belief that Java EE — really, a Java monoculture — is the number one culprit for the project I’m working on being so very, very late. It isn’t the only problem we have, but it’s a big one.

The question I’m left with is this: what do I propose as an alternative?

.NET is out. We’d need lots of new hardware and be locked with a single, closed-source vendor. Please do not bring up Mono.

Cold Fusion? Same single-vendor problem, and I remain unconvinced.

There are Python frameworks like Django and TurboGears. I like them. They’re just not compelling enough for me to suggest using them.

Tcl? Sorry, private joke.

To my mind, it comes down to PHP (using any of a number of frameworks) or Ruby on Rails.

We would stand to benefit more from both Ruby and Rails, but I am concerned about deployment scenarios. Large-scale Rails deployments are possible, but it’s still a new enough platform that people are still working out the kinks for how to do it well. To be honest, we’d just be talking about a medium-size deployment, but the same concerns apply. Could we do it? Yes. I have faith in our system administrators and our developers. I would just feel really guilty about going to them every 3-6 months with a new way of setting up the servers to deploy Rails apps. I admit that I haven’t been following that scene for a while, but it does seem every time I poke my head in that there’s something significantly different.

On the other hand, I’m satisfied that we know how to work well with PHP, at least from a sysadmin perspective. On the development side we would certainly move a lot faster than we do with Java. As PHP becomes more Java-like, you’d think that a transition from Java would be easy. That’s the whole point, right? PHP has a low barrier to entry. On the other hand, its similarities to Java (at least in the object model) may make the transition harder, as I’ve found that Java programmers are somehow blinded to or misunderstand significant platform differences — e.g. PHP’s share-nothing architecture. We would also struggle with maintaining a large PHP codebase, much as we struggle maintaining Java code in reasonable order.

But honestly? In my gut, I hesitate to propose PHP because of the language itself. It doesn’t feel that much better to write PHP code than Java code — sometimes it’s worse, especially some of PHP 5’s Java-inspired syntax. Mostly, though, I think it’s the lack of closures and blocks, language features that I’ve come to expect and rely on. Peter Williams brings up the same point about PHP.

I first learned blocks and closures about two years ago and now find programming without them mildly painful. I think that Mark Jason Dominus got it right when he said

in another thirty years people will laugh at anyone who tries to invent a language without closures, just as they’ll laugh now at anyone who tries to invent a language without recursion.

There are just so many common classes of problem that are simply and cleanly solved by closures that not having them seems like a crime.

I think I’m holding out for JRuby on Rails.

On a tip from Todd Ditchendorf, I’ve had JavaScript and Ruby interactive shells running in Dashboard widgets for some time now. It’s fan-frickin-tasic. If you do anything with Ruby you probably already spend a fair amount of time in irb, and I imagine that Python programmers also understand. But it’s a fair guess that few JavaScript programmers are firing up a JavaScript shell to, say, test a regular expression. Maybe that will change with Java 6 adoption, since it has a native, defined API for interaction with dynamic languages.

Having the shell just a keystroke away is a huge help, all through the magic of WidgetTerm.

In “Taming the Rhino: Making Mozilla’s Javascript Command Line a Little Less Brutish”, Charles Lowell suggests using JLine to bring Rhino’s shell up to snuff. I have to admit, not having history is a pain in the butt. So I’ve dropped the JLine jar into ~/Library/Java/Extensions so it’s in my classpath along with Rhino’s js.jar, changed my js alias to java jline.ConsoleRunner org.mozilla.javascript.tools.shell.Main, and life is a tiny bit sweeter. Thanks, Charles!

I snapped this from the bus on the way into work yesterday:

What a damning title for this post.

Peter Williams has started working with PHP. He comments mostly about the syntax and with the understanding that he’s writing about PHP 4. Some of that has been improved upon in PHP 5: exceptions, for instance, to which Matt Zandstra has written a good introduction. I agree with a lot of what Peter says. DHH has a point: PHP is not pretty to look at, and sometimes it’s ugly to use. Using -> as a method invocation operator is unpleasant (Perl does the same thing, but in Perl 6 it’s a .). It’s a small thing, but small things add up. I don’t like to use PHP because I like the syntax of the language or because it’s a joy to write PHP code. I like to use PHP because it gets the job done, sometimes quite powerfully. And it’s a helluva lot better in PHP 5.

It is a joy to write Ruby code. Just want to say that. Coming from Ruby to PHP, that’s gotta be hard.

At first I thought Peter was being a bit over the top complaining about PHP’s requiring explicit statement terminators. Then I thought back to how much time I’ve spent tracking down bugs that turned out to be a misplaced or missing semicolon. He’s got a point.