by Andrew Oliver

The long death of fat clients

news
Jun 28, 20125 mins
Core JavaJakarta EEJava ME

Why JavaFX is a last gasp

shutterstock 2344090325 fat cat in pink glasses and a chain around his neck on a pink background
Credit: Anna List / Shutterstock

With Adobeโ€™s divestment of Flex and mobile Flash and Microsoftโ€™s move from Silverlight to Metro, Oracle now seems all alone in believing that a fat client framework โ€” in the form of JavaFX โ€” is a worthwhile investment.

Fewer and fewer options exist for developing purely fat client desktop applications and fewer still for RAD applications with Web-based delivery (aka, โ€œthick clientsโ€). We are on the verge of a purely HTML/JavaScript client world.

Or we would be, if it werenโ€™t for mobile pushing us back to client-side development. If youโ€™d asked me a decade ago whether Objective-C would come back from the grave, I wouldโ€™ve waited for the punch line to the joke. Meanwhile, Google decided to follow suit with its own slimmed fork of Java and an XML UI definition language.

Tools like PhoneGap didnโ€™t make quite the splash they might have because users came to expect that native look and feel. A new niche was created with tools like Appcelerator Titanium that gave a look and feel without making you code in Objective-C, but they were too slow to develop to keep the faux-nativeness going.

Yet HTML5 is starting to come on strong in the mobile world, at least in terms of video and other functionality. Increasingly, developers trying to target both Android and iPhone are going this direction, a choice made easier by the dinosaurs RIM and Nokia continuing their long march into extinction.

Death to fatties

In fact, the new mobile world is driving us to the standards that were always resisted. Microsoft has started supporting jQuery โ€” at least in spirit. Even Apple, the reigning king of all things proprietary, has become an HTML5 advocate.

Oracle is the last holdout with its JavaFX framework. While visiting my companyโ€™s Chicago office, I attended a Chicago Java User Group presentation by Oracleโ€™s Roger Brinkley. With compliments to Mr. Brinkley, who had a tough job, there was a bit of โ€œparty like itโ€™s 1999โ€ feel to the presentation.

The JavaFX demos donโ€™t work well on the Mac, let alone on Linux. They smelled of SwingSet demos from back in the day. Every developer in the room was probably thinking, โ€œWhy would I want to do that fade effect with a big fat Java install when I can do that in jQuery without one?โ€

The XML language looked like Androidโ€™s XML UI langauge. When I asked Mr. Brinkley any question that required a comparison between Android and JavaFX, the response was essentially, โ€œThis is the direction of Java on the client.โ€ Talking points are the last bastion of an intellectually dead argument.

Soothsaying

Oracle seems to be developing JavaFX as a way to hold onto legacy clients that have been leaking to .Net. The strategy will fail because it requires the adoption of โ€œsomething else.โ€ While youโ€™re at it, you might as well adopt the something else everyone else is adopting that doesnโ€™t require a client-side install, especially since all the stuff JavaFX can do HTML/JavaScript/jQuery can do better.

Developers will continue down the HTML/JavaScript/HTML5/jQuery path. The costs are dipping as WebKit becomes the browser monopoly that Microsoft tried for but couldnโ€™t maintain. Capabilities that were traditionally locked down in Flash or native code are being opened by standards and open source tools like jQuery. The standard has created a market, and that market is the new monopoly. Ironically, Apple may rue the day it promoted HTML5, as the cost of multiplatform support is lowered.

Whether or not you think this is the direction we should go, you have to admit the options are shrinking. JQuery is the killer framework for HTML, and itโ€™s a monster swallowing the world of fat client options. In the cloud era, why would a company want to install something native and pay the costs of traditional PC maintenance? Full-scale adoption of HTML5 will end this argument for good.

The fat client been a great ride, but Iโ€™m glad itโ€™s finally over. Client-side installs for corporate applications have always been fraught with problems and have required draconian IT policies to maintain securely. The new browser-based world is the joint dream of both IT managers and developers alike. Those holding out for more fat, rich, thick client development had better face the fact that the world is slimming down.

This article, โ€œThe long death of fat clients,โ€ was originally published at InfoWorld.com. Follow the latest developments in business technology news and get a digest of the key stories each day in the InfoWorld Daily newsletter.