Sun's open sourcing of Java: avoid the red herring

Comment The other day I attended a briefing with Simon Phipps, Sun's chief open source officer which was ostensibly aimed at updating the assembled on Sun's progress in its open sourcing of Java.

The "big news" was that Sun now plans to bring its reference implementations of the mobile "editions" of Java (Java ME CDC and CLDC) to market under open source licenses.

It also has a bit more clarity on the timescale for its open sourcing of Java Standard Edition: code will start being released under the new license regime before the end of the calendar year. And in addition, it's created a portal to serve the interests of potential Java committers.

Sun hasn't yet settled on the type of license that it will release its Java SE and Java ME implementations under (although Java EE is already "done", using Sun's own CDDL license): the situation here is a bit more tricky than with Java EE, because one of the things Sun wants to do is to make it easy for desktop Linux distributions to ship with Sun's Java implementation - something that's been tough to do so far because of licensing incompatibilities and ambiguities.

Having laid all this out, for me, the importance of these details in the broad scheme of things is a bit questionable. What's much more interesting is what effects these moves might have on the ultimate beneficiaries of Java technologies. I have my own position on this, and Simon agreed with it when he responded to my question. To paraphrase him a little:

Most people don't know or care about how Java is made...so for most people it [Sun's involvement in open-sourcing Java] will have a negligible effect.

Exactly. Sun's active involvement in bringing Java and the open source community-of-communities closer together might matter a great deal to a small but highly influential group of industry insiders and black-belt developers, but to mainstream enterprise audiences Sun's involvement is a red herring. At least, in the way that most people seem to be reporting it.

Why? Because the truth is that (1) most enterprises already use Java from an established vendor which has licensed the technology from Sun, or which has already managed to work around licensing issues quite nicely; (2) there are already a number of free and open-source alternatives to Sun's JDK out there (see Apache's Harmony, Kaffe and the Classpath effort for three examples), for those who want them - and with the caveat that these alternatives aren't all completely 100 per cent compatible implementations, they are already shipped with desktop Linux distributions today; (3) the Java Community Process (JCP) now allows for open source implementations of any JSR (as long as it was created under version 2.5 of the process or higher).

There are only two things that Sun's active involvement does do, as far as I can tell: (1) allow Sun to say "look, other people have been asking us to open source Java for ages, and we're listening, honest"; and (2) enable small vendors to attempt to build interesting businesses on top of the pukka Sun-originated Java reference implementation - something that many find prohibitive today due to licensing fees.

In theory re: (2), over time this will enable small vendors to push more innovation into the mainstream Java codebase. Note, however, that this does not mean that said small vendors will be able to build viable commercial businesses. Why? Because most enterprises, whether we like it or not, prefer to buy services and support from large, established vendors. Hmm.

To me, as someone who thinks about the business value of IT, what's really at the heart of all this, came almost as a side comment from Simon (although he's also said it before). And that's that to Sun, the open source model is a means to an end: and in large part the "end" is about aligning Sun's remuneration with the business value that customers get from using its software - rather than making them pay for software up front, before they've had a chance to get the value from it.

In other words, for Sun, open source software distribution is to a great extent interesting because it's an established vehicle for offering software for free and then charging for support and services around the software.

For a company like Sun, which has always struggled to make lots of money from software, there's an obvious appeal here which also nets it a lot of "industry shaper" kudos and brownie points.

I'm in violent agreement with the view that the enterprise value of open source software is mostly about only having to pay for something if and when you use it in anger, rather than when you first acquire it, and I think this model will grow and grow - for infrastructure software categories primarily, in the short term.

The challenge with this view is that it's politically incorrect, and companies like Sun that see open source as a tool to turn over their business models end up spending inordinate amounts of time trying to justify their intentions in the name of "building communities" or similar.

Copyright © 2006 Macehiter Ward-Dutton

This article was originally published at IT-Analysis.com.