Its almost a month now since the big fanfare and arrival of the Java 2 platform. For almost 18 months we had awaited JDK 1.2 and now its finally here. For UI Developers the world over, there was the promised Java Foundation Classes (JFC) which extend and replace the elementary AWT class library. The JFC (codename Swing) has been around for almost 15 months. So, is it finished in the Java 2 release?

Usenet newsgroup postings would tend to suggest not!

Broken APIs and Bugs

Swing 1.1 - the release for Java 2, seems to break or change a lot of the existing APIs. It has also introduced a number of often trivial but hugely annoying bugs. The truth is that the quality just isn't there.

I have been a great enthusiast and advocate of Swing since its announcement in 1997. I believe that the Swing team did a great job getting such a lot of the toolkit developed and into a respectable shape for Jan 1998. Those early developers were evidently great guys at getting the work started. They were real enthusiasts for what they were doing. It would appear that the team was lacking what Belbin psychologists call "Completer/Finishers". That's what Swing needs now.

Focus

In truth, Swing suffers from a general problem of focus which is endemic across the whole of the Java platform. Everywhere, there is too much width and not enough depth. Depth in functionality and depth in quality.

Java seems to be thrashing around like an headless monster, out-of-control. There is no clear direction. Sun appear to be trying to be all things to all men.

The development and direction that Swing has taken over the last 12 months, would clearly seem to indicate that it is intended to compete head on with MS Windows and Apple OS in the desktop applications market. Even the selection of the Metal Look and Feel would indicate and emphasise this conclusion. It is essentially a Sun corporate image - including the colour selections - for a desktop windowing environment.

Meanwhile, the continued insistence on adding more and more functionality to cover a wider and wider intended market is causing havoc with developers. Better to finish off what is there and start again with a new focus. The current toolkit is in any case fatally flawed. The restrictive threading model and the continued failure of the serialisation means that it will never meet initial expectations.

It seems that Java suffers from this all too much. A look at the jobs page at the Sun website will tell you that there are over 40 vacancies for Product Managers. Doubtless one of these people is needed for the Swing toolkit. Adding Product Management might help to get small scale customer focus and deliver finished product short on bugs and without broken APIs, it will not, however, solve the big picture.

The Product Management needs to be reporting to a small and highly focused team of strategists. A clear strategy for Java has to emerge. If Sun don't have the manpower to deliver a Java which is all things to all men, then they must focus.

A UI for the Masses

Over the last year of development without direction, little has been done to address the needs of Applet developers, or developers for other hardware platforms e.g. JNI enabled devices.

With the demise of the Network Computer concept, the notion that a different, simpler, more intuitive environment should be developed, seems to have disappeared. Neither Sun or IBM are pushing forward with the work which was done on HotJava Views or Lotus eSuite.

Just because NC hardware didn't sell in its intended market place, the need for a better User Interface for the greater population seems to have gone away. This is surely wrong!

We still need a new, better, simpler environment for the ubiquitious computing era which is coming soon.

Strategy Proposal

So what can I suggest? It seems to me that Sun is spreading the resources too thinly to achieve everything. Competing against Microsoft in the desktop market whilst trying to shift Java into new markets isn't working. They are failing in both areas. If they are to achieve the penetrating and defensible market position in new markets such as embedded JNI devices then they must focus the best people into that area.

If I were Scott McNealy I would achieve this by giving the market what it wants. Take the areas of Java which compete against Microsoft and release them as Open Source. Let the market fix the problems and move forward. This would certainly include the Swing toolkit in its current form.

Re-apply the freed resources to other areas, and develop a clear and focused strategy. Identify market areas for Java technology where it can be established as the no.1 player. This requires functionality, service, support, reliability and quality. Go after those market niches and knock them over 1 by 1. Then re-apply the talented developers into new areas as they complete work in a targeted area.

Sun is a clever and knowledgable high technology company. It appears that the size and speed of the Java market has swamped it and first principles of high technology marketing and development have been forgotten. Get back to first principles - focus, strategy, quality, timely delivery - then things will start to get better.