Are applets making a comeback?

Good old applets -- do they deserve a place in your RIA toolkit today?

1 2 3 4 5 6 7 Page 4
Page 4 of 7

Testing the panacea theory

Q.What are your impressions of the Consumer JRE/Java SE 6u10? Is it enough to repair the existing problems with Java deployment and applet startup?

Danny Coward: I think we've nailed both of those issues in this release, but since Java SE 6 update 10 is currently in beta, I think developers will be the ones to answer that question for themselves by trying it out This release makes a massive leap forward in addressing the speed of applet startup -- especially from "cold" (when the JVM is not already running), by loading the JRE files into the disk cache we're on a par with warm startup. The Java Kernel project has brought the initial download down to just under 4MB which is way smaller than the current beta of the Adobe AIR download.

Romain Guy: Java SE 6u10 is a great step forward. The applets experience is a lot better than it used to be. I doubt this will be enough to make applets gain momentum again though. As I said earlier, it's more about the tools than the technology itself. It will still be great for companies that already use applets based solutions or even for the games market, still quite strong as far as applets are concerned. But again, I don't see these improvements help reverse the trend of applets going away.

Chet Haase: Update 10 is definitely making great strides toward fixing some of the pain points in Java deployment over the years. Between Quickstarter, Java Kernel, and the new plugin architecture, these changes should help address some of the biggest holdups to applet adoption.

There are still some fundamental constraints with the Java platform that cannot be solved with these changes, however. For example, Java Kernel can make the initial launch experience faster than it currently is because it lessens the amounts of bits being downloaded, installed, and loaded. But there's still a lot of data that has to come down simply because Java is a large platform with a lot of interdependencies in the Java APIs. I believe that Java Kernel is probably the best that can be done with the heavy and important constraint that Java continues to be a single platform that can be depended on for various capabilities in a backward-compatible manner; you can't simply start whacking stuff out willy-nilly and expect to keep the Java developer community happy when their favorite APIs disappear.

Java Kernel will certainly speed things up, but will it be enough to overcome some of the objections to slow install in general? Time and experience will tell. It's not clear whether these changes will be enough to hoist applets into the forefront of Web applications again, but it's certainly worth having these problems addressed for any developers and users that want to consider using applet technology (or any other desktop Java technology that also benefits from these changes).

Cay Horstmann: Decreasing start-up time is great. Doing this kind of work is Sun's strength.

The installation needs a lot more work. I installed the update on Windows and Linux and it failed on both. On Windows, re-running the installer solved the problem. On Linux, I had to fix the browser linkage by hand. Granted, it was a beta, but you'd think after ten years these guys could build an installer that works.

The Deployment Toolkit is a step in the right direction. Ask me in a year whether Sun let it wither on the vine, or whether they are actively maintaining it.

Jim Weaver: I really like Java SE 6 update 10 ... as it addresses the issues that made applets impractical in the past [... T]he incremental Java kernel installer loads just enough of the Java kernel to run a simple "Hello World" style program, and incrementally updates the kernel as needed. This avoids the long waiting time to install the JRE, the net result being that the applet starts up as soon as enough of the kernel has been downloaded to support it.

The next-generation Java plug-in addresses reliability, configurability (if that's a word), and communication with the environment.

JNLP support means that the user can click a link in a browser and the Java/JavaFX application is invoked in a separate window, external to the browser (but within the Java plug-in). By the way, I think that RIA (rich Internet applications) aren't just for browsers, and that developers and users will become increasingly willing to create and use RIAs that are external to the browser.

John Zukowski: I like what I've seen so far. It appears that Sun is listening and trying to address perceived problems with deploying client-side applications meant to be delivered through the browser. Will it cause more people to write applets? I doubt it. While there doesn't appear to be anything wrong with the new changes, certainly making things better, I can't see them as fixing perception on why people do not write applets.

1 2 3 4 5 6 7 Page 4
Page 4 of 7