A good Devoxx day

Wow, it has been quite a day! I’ve attended quite a few interesting sessions, but most of all I met a lot of interesting people. I have the habit of starting up a conversation with the people I meet at a conference, but Devoxx makes it quite easy.

The downfall of java EE
I went to post my blog and charge my laptop and found myself in the company of the guys who presented Spring Roo yesterday. We had a nice conversation about web development and Ben Alex from SpringSource opened my eyes to something. I’m still evaluating, but at least it triggered me. His statement is that all the development that is being done at the moment for java EE is doomed. His argumentation starts with the statement that the future is in the cloud. Of course Ben is a VMWare employee, so you might say he has an interest, but global sentiment is proving him right in his cloud statement so far. He goes on to state that the big application servers introduces an enormous overhead concerning startup and memory usage of applications. Cloud vendors want applications to be lightweight so they can start them on usage and decommission them when there is no demand. These 2 statements directly conflict each other, hence the likelihood of the demise of java EE.

As I’m writing this I’m giving this some more thought and although it sounds plausible, I’m not sure I agree. I have to agree that the number of companies that I know of that split up the runtime of their EJBs from their webapps is limited to say the least, so there is definitely a lot of waste in this spec. I also agree that there are things that need definite and immediate improvement, such as the further introduction of convention over configuration.

Where I hope he will be wrong, is that I hope the java EE spec will evolve into something that adopts all best practices that the people from SpringSource, wicket, restlet etc introduce. The spec should not lead like it tried to with EJB and JSF, but follow and complete good ideas like the W3C HTML4 and CSS2 standards.

Furthermore, there really isn’t that big a difference from having the SpringSource folks trying to optimize our runtime than there is in having Oracle or IBM trying to do the same. It’s all a matter of motivation and that usually stems from money lost or gained. As the cloud field progresses, the JavaEE app servers are going to have to step up and do their jobs. As they foresee their revenues dwindling, they will gain the motivation to do just what is needed.

I’m hopeful we’re not on a dead end by implementing Java EE 6, but only time will tell.

What good is open source?
Another interesting talk was at the Quick on a stop for a hamburger. I bumped into the guy I just bashed on twitter for presenting his web framework. I still stand to my statement that there are already way too many of these, but he, Marc Portier, had an interesting point on this. At his company they open sourced the web application framework they developed in-house. Of course the wisdom of writing your own web framework can be disputed, even in 2007 there were plenty of web frameworks available. Still, they took their labour and open sourced it and then continued working with and on it. They are applying “eat your own dog food” to the extreme while at the same time they are giving their customers the guarantee that the apps they write will always be maintainable. I doubt that they will ever find their web framework, Kauri, the talk of the town. Nonetheless, they do get customers through the sales point of being open source on occasion, so Open Sourcing is bringing them something. I guess most Open Source projects start from a particular itch a developer wants to scratch, but this goes to show that even if your project does not skyrocket, you can still benefit from being Open Source

The Java Posse
Finally, I got the chance today to meet and thank Dick Wall of the Java Posse. They have been bringing me the news I need for my development work, so it was great to shake his hand and thank him.

Advertisements
This entry was posted in development, devoxx. Bookmark the permalink.

2 Responses to A good Devoxx day

  1. Ben Alex says:

    Thanks for your thoughts, Martin.

    I don’t believe I used the term “doomed” to describe Java EE 6, but was simply noting that tomorrow’s challenge is how to build apps that deploy successfully to low-cost public and private PaaS clouds. I noted that complexity increases costs for cloud providers and thus, in the drive toward lowest cost, the lightweight, low constraint environments will likely win. Spring is well positioned to be at an advantage here, while Java EE 6 is not. I’m happy to discuss the reasoning with you again if you want to meet up at Devoxx.

    Cheers
    Ben

    • sunsear says:

      Hey Ben,

      I completely agree with your statement about the lightest weight environment eventually winning out in an environment where low overhead is key. I would definitely agree that a web container is usually low overhead, while a full EE container starts at a loss.

      I also think that the JEE vendors have not really had the necessity to tune their app servers for startup time and low overhead. You can see the Glassfish team is now drawing the same conclusion you do, albeit for different reasons. They are working on tuning their app server for low startup, so that developers can work with it better. Developers are a good share of their market, so they have a vested interest in helping them out.

      I must say I appreciated our discussion, it got me thinking and that was fun, although it hurt a little 🙂

      Hope to see you guys at a next conference again.

      Grtz,

      Martin

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s