Oracle / IBM Pact: Does it really matter?

In a post on this blog earlier this week, my esteemed colleague Jesse Davis weighed in with his thoughts on the Java Pact singed by Oracle and IBM. Jesse talked about how his concerns about the fragmentation of the Java language were put to rest with what he described as a pact that had some meat behind it.

Well I have some good news for Jesse – but as so often is the case, I’ve got some bad news for him too. Let’s start with the bad news.

While peace may have broken out, I think some very large questions hang over the Java Pact - it’s my opinion that this is merely papering over the cracks. Despite what Oracle and IBM are saying now – and admittedly, they’re saying the right things – it won’t take much for one of these mega-entities to cause a splinter.

Let’s face it. The major Java backers have divergent needs. When a rift inevitably does happen, I think we’ll see the scenario that Jesse fears. The needs of the community are diverging, and we’ll ultimately end up with multiple versions of Java that will be unrelated.

But here is the good news, for the industry at large - this isn’t a bad thing! Sure, a unified Java platform certainly has its advantages – compatibility and portability of applications across multiple platforms have proven to be the backbone of Java’s success. But I suspect now the needs of the community may not be best served by boxing in the platform so tightly with the rules and regulations currently required. Java will evolve much more slowly with these restrictions, and developers are likely to scatter to build the applications they need. While the improved support for multiple languages on the JavaVM will help matters, the pressure of core feature demands may ultimately prove its undoing.

Moving forward, the needs Java users continue to diverge with different enterprises and different verticals needing different things, I think we’ll see the major Java backers splinter accordingly.  The end result, quite frankly, will be a more diverse Java platform, with end-users having more choice. To me, it sounds like the pros outweigh the cons.