Eitan Suez's Blog

The Value of an Open Java

I just read on java.net an abbreviated news bulletin regarding a comment by Sun CEO Scott McNealy stating that Java will not be open-sourced any time soon. As I understand it, the main reason behind this decision was We're trying to understand what problem does it solve. That's a valid argument.

First of all, what is meant by an open Java? My interpretation is the open-sourcing of the J2SE libraries (collections, swing, etc..).

The value, then, of open-sourcing the J2SE libraries is that these libraries would quickly become much more robust than they have been up until now. I mean that the quality of these libraries both in terms of performance, number of bugs, and feature completeness would improve. If managed properly, J2SE could see significant improvements quickly.

Any features that developers claim are missing would quickly be contributed as patches to the J2SE codebase. If you subscribe to some of Sun's own forums on Java, you will quickly discover what features developers are screaming for. In many forums the typical case is too few Sun engineers with too many requests made upon them. This is a serious and absolutely artifical bottleneck. There are thousands of willing and able developers who probably would submit patches against J2SE.

A word of caution here is that the virtue of being open-source does not imply a successful project. The project leaders must know how to manage open-source projects. Eric Raymond's famous essay The Cathedral and the Bazaar explains this much better than I could do in this blog entry.

So, if Sun were to open-source the J2SE libraries, two things would happen:

• Developers would get exactly what they want (a better J2SE faster)
• Sun would get exactly what it needs to combat Microsoft's .NET

Personally, I don't see .NET competing in the first place since it is a closed technology running only on a closed platform. However, the only area where .NET does compete (quality) and where Sun is in certain ways failing its developer community, is what the open-source community is offering it, at no cost!

Why am I writing this? Reading Eric Raymond's open letter to Sun on the topic of open-sourcing Java, I thought that he'd described the problem eloquently. I discovered that lots of people didn't really get it. Worse yet, Sun does not appear to understand it. In my opinion, this decision could be the most significant business decision this company is facing at the moment, in terms of its fight against .NET or other competing language libraries that developers opt to use.

In closing, my hope is that this small blog entry, swimming in a sea of text and opinions on the matter, might help catalize the understanding that open-sourcing J2SE not only has value but tremendous value, if properly wielded.

Comments

• Better Libraries
  I think you hit the nail on the head. Right now there is zero incentive to contribute patches to the JDK. For one, your name won't be recognized, second, it will take years to see the outcome, and third, you are just giving your best ideas to Sun. Why bother?

  The success of the Apache Jakarta projects shows there are many open source developers who want to extend Java's functionality, and have the skills to do so. Sun should see that this is a net gain for them--their engineers could remain as commit-ers, or work on the implementation of the core language libraries, while gaining from the input of programmers around the world.

  I think Sun made a decision to position Java in the top-tools market. PowerBuilder did this as well: you build an ecology of companies who charge quite a bit for extensions (libraries, frameworks, tools) to the language. The problem is that if I want a good grid control based on a JTable, I have to pay hundreds of dollars for it--likewise with other extensions. It's aimed at a corporate developer's market. But it also means that open source projects in Java have to reinvent the wheel or suffer with the default controls in Swing, for example.

  MS approached VB different, by building an ecology around low-cost tools, so that after several years, there were thousands of ActiveX controls you could buy at very low cost.

  I think this is at the core of Sun's problem with Java. But it would take longer to discuss here, so I'll just leave it at that for now.

  Regards
  Patrick

  Posted by: pdoubleya on March 29, 2004 at 01:17 AM

• Compiler/JRE/J2SDK
  Java can benefit greatly from a Sun managed open-source project of the Compiler/JRE/J2SDK.
  The JRE is the primary source of concern, because it is almost an operational system (it doesn't actually manage the IO, scheduling, and other resources). But it is the access to the IO, scheduling and those other resources.
  OK so my opinion....
  I don't trust Microsoft code, of the operational system. I don't trust any proprietary code if it is operating (:-) my system. Why would I trust Sun?
  The compiler and librarys are just consequences.

  Posted by: levmatta on April 24, 2004 at 08:01 AM

This work is licensed under a Creative Commons License.