Screaming Penguin - Source Does Matter

Chris has a big commentary up about the decline -- perhaps another word would be better here -- of Java on the Mac.

Just some notes here...

Dr. Gosling writes, “as much as I love the Mac’s eye candy, it really hasn’t been keeping up as a developer’s machine.” Two points to make here. First, he didn’t used to just see the “eye candy”; years ago he wrote that he thought of OS X as “Linux with QA and Taste.” The very term “eye candy” has become something of a pejorative for dismissing the Mac, to suggest its strengths are purely visual. This despite the fact that the Mac can, by Gosling’s own admission, handle laptop power management better than Solaris.

Now, I am someone who moved from Linux TO the Mac, because I got tired of dicking around with Centrino Wifi Drivers vs Firmware releases, power management and all that crap. I still have Linux on my desktop at home, and love it, but certainly in the laptop world, Linux doesn't cut the mustard on just working. And frankly, my attempts to even get Xorg working right on Solaris leads me to say it is WELL BEHIND Linux in terms of day to day usability.

Secondly, it’s been keeping up just fine as a developer’s machine… for those developing Mac software (or Flash, but that’s another story). Those who are playing with Core Image and QTKit and all the other neat stuff in Apple’s API’s are pretty happy with it. The missing word here is “as a Java developer’s machine”. JDK 6, which Apple develops for OS X because Sun doesn’t do a Mac version of their Java runtime, has yet to be released despite some 2006 preview releases tracking fairly closely with Sun’s previews. In absence of official information, most resigned themselves to the idea that Apple had made JDK 6 a Leopard-only feature, and then Leopard slipped by six months. There’s also the fact that Apple deprecated Java as a development language for Cocoa, which brought the predictable screeds of Apple Doesn’t Like Java, Never Has, Probably Never Will (despite developing and maintaining their own VM, because Sun wouldn’t, since the late 90’s, and including it in every copy of OS X, something you don’t see on Windows or Linux).

Yeah, Mac is the dominant Ruby developers box right now, with TextMate being the dominant Ruby editor.

Here is the thing, though. Over the last couple years I have spent more time at developer conferences/events than I have for years previous, and what I have noticed is that at general (Non-Windows only) events -- like the recent BarCamp Atlanta -- Apple now has near parity with Windows in terms of users. At Java-specific events, Mac is the DOMINANT platform. The thing is, most places with Java developers don't give them the choice. I still have 3 Dell boxes on my desk at work because that is just what they give you. I would love to have a Mac Pro, but that isn't an option. However, I personally own 3 Macs, two of which are in daily use.

n a followup to Gosling, Keith Weinberg writes, “Apple needs to get the picture on this. Not only will they miss out on the developers, they’ll also miss out on the applications they’ve developed.” Really, applications they’ve developed? Where? They’re all on the server, which is probably running Windows or Linux, and so whether a client is a Mac or Windows or an iPhone or a Wii is irrelevant. End-user Java applications? Name a few, not counting NetBeans or LimeWire, which have been the default answers for years.

Azureus, JXplorer, SquirrelSQL. OK, the last few there aren't "Top Tier" applications by any stretch, but they are certainly widely used.

As a bit of an aside, but also remember that the Flex server and Flex Builder environment, as well as ColdFusion are Java-based. Dragging on Java is eventually going start putting a kink in Adobe's product offering. Dragging on 1.6 might not be huge, since there weren't any significant language changes since 1.5, but once 1.7 drops, it is gonna be a whole new deal. Wait till Adobe is the one bitching about Apple not supporting Java and see how fast that changes.

Anyway, name for me a big Python or Ruby client side app. There's uh, Odeo (Python), which Apple killed when iTMS started indexing podcasts, and... nothing. Hell, I can't even say I have ever seen a Ruby client of any description.

The thing here is, Python and Ruby are maintained FOR Darwin outside of Apple. Apple adding them to the distro is like saying apple is "supporting" bash. They don't have to, they just have to put it there.

Flash is wiping out the last of the applets, and installation/maintenance hassles made double-clickable Java wretchedly impractical. I’ll grant you the point on missing out on the developers, but is it that big a deal? A few million Java developers, some fraction of which own Macs, makes them just another special interest group, like educators, media professionals, scientists, etc: they’ll be a factor in what’s included in the OS or default apps, but since Java developers use Macs to develop apps that are rarely if ever run on Macs, they don’t contribute to the well-being of the platform like other kinds of developers do, and are thus more like the other special interest groups.

Now, here is the thing... Java developers, which anecdotally I will say 50% of which own Macs -- about the same ratio as even graphic designers in my experience, are certainly a larger market than the Ruby folk. But I want to turn Chris's last point around:

If you really think Apple should support Java better — that it should get the runtime and JDK out faster, keep Intel and PPC support up to snuff, talk it up to developers, etc. — then answer this question: how will doing so make more money for Apple? In what way is supporting Java going to move more Macs, iPods, and iPhones, or demand a higher price, than if corporate resources were directed elsewhere. Say, to the hardware lab where they make sure that the laptop goes to sleep when you close it.

How about in stead of making more money, Apple goes for "costs less." If Apple would just merge their fork into the OpenJDK project, I suspect there would be plenty of people who would work to make sure Apple sees first-class support. Apple could even keep one or two staffers on OpenJDK to oversee it and check it, THEN distribute OpenJDK the same way they will be with Ruby or Python.