First of all, let me start by saying I am by no means a Java EE advocate. Not even by a long shot. I’ve been working with Spring for the last 10 years or so and aside from a couple of consultancy assignment where I was basically forced to work with the EE ecosystem (I’m looking at you, Websphere), Spring has been my weapon of choice when it comes to enterprise development.
Last week, the Java EE Guardians published an open letter addressed to Oracle regarding the rebranding of Java EE and the consequences of such a rebranding. Perhaps a little history: Oracle open sourced the Java EE codebase, which then moved to the Eclipse Foundation. With that move, a rename was pushed trough, Java EE was now called EE4J. You see, Oracle still retains ownership of the Java brand and therefor didn’t allow continued use of the Java EE brand in it’s current name. Additionally, most EE code extensively uses the javax.* packages, but Oracle also retains ownership of the use of those packages. This means any new API’s coming out of the EE4J project would not be allowed to use the javax packages.
In their plea, the Guardians asked to Oracle to retain the possibility to keep on using the Java EE name and to be able to use the javax and javax.enterprise package names for new APIs, under licence if it must. Oracle’s answer was clear: no. Java is a protected trademark for Oracle and they’re not willing to let it go.
Now, what’s all the fuss about, you may ask? A package rename is for most a single line sed script, and the name of a product can change, AngularJS changed to Angular after all.
Well, Java EE (or EE4J) is considered as being the ‘standard’ approach to building enterprise Java applications. It’s API are considered standardized through the input of various enterprise stakeholders by using the now defunkt JCP. On the other hand you had Spring, which was a product of Interface21/SpringSource/Pivotal, which is now a very popular, if not the most popular, approach to building Java applications. However, it’s not considered a standard, at least not by the Java EE guys, as they laid claim to that title.
Being considered as a industry standard has many consequences. Some companies only wish to work with standard because of the perceived safety using a standard means. Integrators tend to focus on integrating with a standard instead of a proprietary approach. So if Java EE loses the perception of being the official standard, it has a lot to lose.
Personally, I never considered Java EE the standard for enterprise development. It was a standard, sure, but on the same level as Spring. People either chose Java EE or Spring, not unlike one would choose Makita or Milwaukee when choosing an angle grinder. For me, the moment Oracle relinquished control of Java EE and turned it over to the Eclipse Foundation, it became a former standard and thus now proprietary approach, just like most other Eclipse projects. It just officially became ‘the framework formerly known as Java EE’ and just one of the ways to build an enterprise application.
So if EE4J is forced to use it’s own non-standard packages and no longer has a name that was linked in our collective mind to a standard, what makes it different from Spring?
Nothing.
But Spring cannot exist without Java EE, Spring implements a lot of Java EE standards, you might say. Sure, but that was out of necessity and because they could be integrated with relative ease. If they didn’t exist, Spring would have found a way to do it themselves. And they have done it themselves: there are a lot of things within Spring that do not have a Java EE equivalent.
So Java EE has a lot to lose because of Oracle’s standpoint. There’s a world of difference in perception here between a Java EE standard and an EE4J API. This is probably why the Java EE Guardians are fighting so hard to retain the name and the packages. They know it’s a war on perception, a war that in my eyes they have been losing for some time. And Oracle’s latest move could prove to be the nail in the coffin of Java EE. Without the protection of being considered ‘the standard’, they’re losing one of their main selling points: that they, in fact, ARE the standard.
Personally, I’d rather see the EE4J project just become a place where the old Java EE get a place outside of Oracle’s control. These are very stable APIs and aren’t that prone to change anyway. And I’d rather see the brains and knowhow of people like Reza Rahman or Adam Bien being put to use to help drive new features in the more fast-moving framework that Spring is.
Perhaps it’s time to bury the rivalry and rally under the same flag. Just imagine what could be achieved by joining forces. Spring would become the de facto standard (if it isn’t already), bringing everything and more to the table of current Java EE developers. Spring would need to provide a bit of support to accomodate Java EE developers, building some bridges to make sure Java EE developers need to change as little code as possible. Like I said, Spring supports just about every standard that Java EE has, with some small exceptions.
To me, it’s clear that Oracle doesn’t give a damn about the future of the EE4J project, they’re just protecting their copyright. And that’s their right to do so, they own the name after all. But to me, they’re killing Java EE in the process. Or at the very least the perception that Java EE or EE4J is the standard, which may be equally damaging. Personally I don’t care, I’m pretty sure Spring can handle it if they wouldn’t have to comply to any new EE4J standards.
Lastly, I don’t get why the Java EE Guardians keep on investing so much time and energy in a battle that seems less and less likely to have a good outcome. I respect their efforts, but I just don’t get it. Instead of antagonizing each other, perhaps it’s time that both sides joined forces and focussed on what really matters for enterprise Java developers: providing solutions for the problems we’re facing today. We’ve had two competing standards for over 10 years now. Perhaps the time has come to choose. Giving Oracle the bird on the way out is just an added bonus. God knows they wasted enough of our time with these shenanigans.
Edit:
So I received some input, which I’m happy to reply to. Some people see the EE4J movement as an open invitation for Spring. This would be their opportunity to standardize some of their work. But… what incentive would Spring have to do such a thing. Why would they want to standardize something they already have and risk having to do more work to comply to a new standard? And if there are standards that Spring would not have a solution for, it’s probably much more feasible for them to just build it instead of first finding consensus over an API… I don’t see any added value joining EE4J would have for the Spring people.
Someone else pointed out that I’m basically advocating a monopoly here. Even though I don’t like monopolies, the current duopoly is seriously out of balance. Java EE standardizes the work of frameworks, see CDI, JBatch, JPA, … I’ve yet to see innovation coming from a JCP spec, and granted, that’s not their goal. But if one part of a duopoly is just copying/standardizing the work of the other, is it really a duopoly? You don’t need a standard to introduce competition. The JSON-B standard didn’t exist when Gson and Jackson pushed influenced each other. The introduction of the JSON-B standard instead introduced a new player, Yasson, who brought nothing new to the table except for a standardized remake of the functionalities in Jackson and Gson. And in the meanwhile someone already made Jackson JSON-B compliant. So I fail to see what added value the standard here brings.
With regard to monopolies: monopolies are broken automatically when superior solutions are introduced on the market. J2EE has a monopoly on the enterprise Java market until Spring came along and broke that monopoly because of the technological advances it brought to the table. This made J2EE move into higher gear with Java EE 5, 6, 7 and 8, but they never caught up again to Spring, which just has a much higher momentum. And recent events have shown that even if (and that’s a huge if) EE4J would be able to build up similar momentum, it would have to do so at great expense (losing the name and packages). EE4J is running a marathon with a 40-pound backpack. And while I applaud the efforts of the Java EE Guardians to get it to the finish line, the Spring guys are already at the finish and training for the next marathon. So if Java EE becomes EE4J and they can’t use the packages, a de facto monopoly is born anyway. The only thing I’m advocating for is that it becomes a coorperative monopoly, a benevolent dictator if you will. And be assured: should the Spring people abuse their position, the community would be the first to hold them accountable and act accordingly. Their history speaks for them: in 10 years working with Spring (version 1 to 5), I’ve never had a ‘What the hell did they do???’ moment. On the other hand, I’ve lost count of the ‘Damn, that is cool’ moments. And apparently, I’m not alone and even avid EE supporters seem to agree. If there’s a technological superior alternative to Spring, I’ll be the first to look into it. But I’m also pretty sure EE4J won’t take that place for the next 5 to 10 years.
I’m still unwavered in my opinion that you need an (open) standard just for the fact that your need one to counter a proprietary market-leader, especially when that standard plays catch-up with the competition. This entire endeavor wastes our time, and more importantly, our client’s time. They want solutions, not rhethoric on need for a standard because of ‘big bad Spring’. If you look at the definition of an open standard, it’s hard to deny that in its current incarnation and through its community Spring doesn’t have an open environment into deciding the direction they’re going to take. Sure, in the end there’s one company behind Spring, but is it that much different with Java EE, where it’s mainly dominated by Oracle, Red Hat and IBM?
I wholeheartedly support initiatives like the MicroProfile, but these should exist in a open market system where people get to decide based on technological merits and not based on something arbitrary as ‘being the standard’ in order to pick the better choice. And as it stands today, Spring (with things like Spring Boot, Spring Data and Spring Cloud) are miles ahead. And I don’t see EE4J closing the gap any time soon.
Tweet this article