Category Archives: Javapolis

My Discoveries of the Year

logoEvery year, the main reason why I go to Devoxx is to discover new stuff. For me it’s all about technology watch. The internet and RSS feeds are my main tech watch instrument but there is one thing that is harder to get through RSS: feelings. Conferences like Devoxx are a unique opportunity, not only to see what’s happening but also to sense how the community is feeling about it, which is at least as important to anticipate on what’s going to be important.

Now you’ve certainly read here and there that there have been a lot of talks about Java EE6 and Closures in JDK 7. There sure were, and there was quite a lot of reactions to those. But frankly, I’m not interested in any of those. I’m not interested in Java EE6 because even though it finally leverages the concepts that have been pushed by the community for so long, the very fact that Sun has been so late in implementing them shows one thing to me: it’s not about them, it’s about us. Sun has been trying to partner with big companies around the JCP, to create a lot of business around those standards. And why were those standards so complex? Because there were so many companies involved in their elaboration? Or was there any interest from those companies to create technologies complicated enough to require a lot of consultants, and books, and trainings, and tools to make things easier? If the first option is true, then how did it happen that a gigantic community of individual software developers made it so that Spring, Hibernate and other Open Source technologies became de facto standards?

Which brings me to the second point I don’t really care about: closures in JDK 7. People have taken matters into their own hands. Other languages have appeared implementing some of the missing features of Java: Scala, Groovy, etc. Some other languages have been imported into the Java landscape, like Ruby and Python. I’ve been using Groovy myself for some time now, and I couldn’t be happier with it. Now Sun is coming after the war, but does it matter? What I see here is that it’s good to have a base language, a base platform. But as soon as you start extending it for purely mercantile reasons, or as soon as you start avoiding certain innovations because you’re afraid it might harm your business or the one of your partners, then things go messy. But once again, the good news is that software gives us power. We developers have the ability not only to decide which technology is better, but to build and promote our own technologies. And I think this is why our world is so dynamic and why things change so fast. What I see here is that languages matter less than our ability to create new ones, to solve more specific problems, to provide some more advanced features. And that strengthens my belief that Language-Oriented Programming is the next big step in the evolution of our technologies.

Now about the things I do care about. My discoveries.

kanbanMy first one is definitely Kanban. I’ve heard about it for quite some time. But I didn’t understand why people were already trying to push it forward, even though we are still fighting to push companies away from waterfall messes. It’s even more radical than Scrum and for that it’s very interesting because it gets closer to what software development really is. The key phrase that kept popping in my head during the 3-hour session about Kanban was “It’s going to take the time it’s going to take.” And that’s why it’s brilliant. Now that I understand Kanban a little better, I see Scrum as a compromise. We’ve taken some of the principles of Lean Manufacturing, we have dissolved them into trusted concepts like RUP’s iterative cycles. And for years, we have been trying to pour that cocktail into their mouth. And in doing that we forgot something important: if they don’t trust us, they will never drink the whole thing, all the more so as it looks weird with all those colors not really mixed together. The way I see it, Lean and Kanban are all about getting back to the basics, and relying on trust. We have to build trust first, we have to make them understand that at least some of us are not interested in building artificial business on top of poor practices. That at least some of us desperately want the software we build to have a real impact on their business. We have to show them our good will so that they let us do our job. And they have to understand that the more they trust us, the faster we will be, the more we will be able to solve other problems. Big software vendors and resource providers will not like that, because every new project comes with its own overhead, because their business thrives on poor practices, stupid methodologies and complex technologies. But once again, power is in our hands, it lies in collaboration, not in corporation.

My second discovery was Spring-Actionscript. Those guys really have a thing to do things in a clever way. Cairngorm, PureMVC, Swiz, they all impose some sort of a structure to your Flex applications, forcing you to surrender some great powers of Flex itself in the process. And here comes Spring-Actionscript, more like a toolbox than a real framework. It doesn’t impose anything. It just gives you all the tools you need, all the glue you miss, to make things fit together perfectly. Their asynchronous abstraction is just brilliant, their configuration options are complete, your code just looks better with it, simpler, more natural. I think that’s what I love most with Spring: not only does it create great technology, but it also instigates a whole pragmatic and elegant way of thinking into the minds of a whole generation of developers, thus encouraging the community to come up with their own technologies: Spring Actionscript used to be called the Prana Framework, developed independently by a Belgian guy who took inspiration in Spring for Java. That’s just awesome. I will definitely integrate Spring Actionscript in a couple of Spring/Flex projects. I think I will even update my todolist sample application with it. Stay tuned.

My third discovery is a couple of technologies to detect and prevent coding errors BEFORE they actually happen. I insist on “before” because of my previous post about TDD: unit tests are all about writing code to check that some other code you have already written (or not written yet) does its job. But to me, this is equivalent to the old prevention versus repression debate. TDD is just repression, and it’s tempting to go there only, because it’s always easier to catch the bad guys than trying to understand why they became bad in the first place. JSR-308 and its pluggable type checkers is all about strengthening the compiler so that it prevents more of the most frequent bugs like NullPointerExceptions. It allows us to make our code more expressive, to give the compiler more information about our intents so that we can prevent bugs from happening. Brilliant! Project Lombok also goes in that direction: it adds a couple of annotations and customizes the Java compiler in order to minimize the amount of boilerplate code we have to type. Once again, by doing so, it improves the expressiveness of the language, allowing us to say more things with less words, thus reducing the likeliness of ambiguities and errors. Awesome! Lombok and type checkers will definitely be part of a couple of projects too. The only thing that really made me uncomfortable with both of these presentations was this question: “Why the hell weren’t those techniques in Java 5 already?”

pomodoroMy fourth and last discovery was Pomodoro technique. Once again, heard of it before, but never dug into it. And then we had this guy with a strange Swedish accent in front of us, playing with dolls, showing us handwritten slides with simplistic drawings. And you could hear the disappointed reaction of a lot of people in the room: “sounds nice, but not applicable to me”. That was my first reaction too. Software development requires long slots of concentration because we need time to load the whole conceptual model of what we’re working on into our mind before being effective, and this implies some overhead. But then when someone asked this very question to the speaker, he answered something like “what if you are loading too much? what if limiting the amount of time you are going to spend on a given task forces you to load just the minimum you need to solve the matter at hand? what if it made you more productive?”. And it made me think: “hmmm… It’s worth a try.” So I will probably try that as well soon.

Overall, this edition of Devoxx was great! The first 2 days, I was somewhat afraid that it would be disappointing, because you could feel that everything was “cheaper”, that there were less sponsors, less schwag, less tempting hostesses. But then the most important part was preserved: amazing independent content and a great community spirit. Finally there was an interesting special guest this year: Twitter. Twitter was everywhere. People were tweeting live from the sessions, there was a projection board with all devoxx-related tweets in the hallway. I and a bunch of my colleagues were even using twitter to cover Devoxx live for our fellow Axenian java developers on our intranet. Twitter was really everywhere this year.

So thanks a lot to Axen for allowing me to go there. Thanks to Stephan and the BeJUG for putting it all together. Thanks to all the great speakers and to my colleagues. This was really a great edition and I can’t wait for the next one.

PS: All the talks will be available in the weeks to com on Parleys.com. So stay tuned.

Conference Guide Available

Conference Guide Icon

Woohooo! Exciting times! It’s confirmed, I will be at Devoxx the whole week, and my Conference Guide iPhone application has just been approved. For your information, it only covers Devoxx 2009 for now, but it’s meant to be generic and I will add more information for other conferences soon: TEDxBrussels, JFokus, etc. The application is free for a limited period of time, so go ahead and install it!

More features for the iPhone application are coming, maybe not for Devoxx (the Apple review process makes things really slow), as well as other mobile clients. An Android client is already in the pipe, and if you are a Palm Pre or Symbian developer, feel free to contact me. I can give you access to the same public API I’m using for the iPhone version.

Finally, if you are a conference organizer and you want your event information to be available on mobile platforms without having to develop your own applications, feel free to contact me as well.

And by the way, if you are a company, and you are thinking about developing your own iPhone application, whether it is for marketing purposes, to sell content or to improve the productivity of your salesforce, I’m your guy. And I can do much more than just iPhone development ;o)

My Devoxx Planning

This year is a little exceptional for me: it’s the first time I attend 2 big technical conferences in a row like that. Last week, Adobe MAX, this week, Devoxx. And all I can tell you is that I love it! I love it as an attendee because those conferences are incredible gatherings of talent and smartness, and because I get to discover exciting technologies that can make my life easier and my software more beautiful. And I love it as a restrained presenter/showman because I envy those who get to introduce all those amazing technologies, and travel all around the world to do so. I’ll do that one day!

Back to Devoxx, I finally built my agenda for this week and it’s very loaded as you can see here or here. My main focuses this year should be:

  • JavaFX: because from what I’ve heard, it seems doomed to failure, but I want to make my own opinion and see what it really looks like
  • Groovy/Grails: because I’ve gotten hooked to those and I am currently investing a lot of time into learning those technologies, possibly for You And The World and for TagOSpot.
  • Random technology watch: topics I don’t have a clue about, but that could enlighten me somehow.

Hopefully, I’ll find some energy left at the end of the day to write a few posts about what I see. Woohooo…

Towards a mainstream Open Source OSGi application server?

That’s official: SpringSource has just acquired Covalent. Or as I explained it to one of my colleagues, “the company behind our IoC framework has bought the one behind our application server”.

My first reaction was satisfaction, because it’s another step forward in the direction of corporate Open Source adoption. It’s always amazed me to see how big companies can be afraid of Open Source. And the fact that there is now one bigger support service offer behind two of the most popular Open Source technologies in the enterprise will certainly reassure some skepticals.

The second “kiss cool effect” was undoubtedly about something I’ve really been playing with lately, since Javapolis: OSGi. The fact that componentization-related JSR’s are so fragmented and so alpha, plus the recent works of Spring Dynamic Modules for OSGi, added to Peter Kriens’ presentation at Javapolis, all of that really got my attention. And now I’m dreaming of building a collaboration platform using Flex for the front-end and OSGi, Spring and Hibernate JPA for the back-end. The only component that’s missing in my big picture is a deployment target platform, i.e. an OSGi application server. Of course, JOnAS is working on that but there’s no documentation on their 5.0 server whatsoever. And I’ve heard Websphere and Weblogic are using OSGi too, but hey, I’m talking about Open Source here! Now have a look at the documentation of both Tomcat and Spring. They’re probably amongst the best Open Source documentations on the planet. Now imagine that quality of documentation for a brand new Tomcat 7 server using OSGi as a core deployment mechanism, and integrating Spring DM libraries to ease the development of web application bundles.

Yummy!

Being creative again

Amongst the things that really amazed me this afternoon in Gent was Lovely Charts, an online charting application that’s currently under semi-private beta. My first impression was something like “if it’s just a Flex alternative to Gliffy, I don’t see the point”. By the way, Gliffy is developed with what was the most advanced RIA technology before Flex 2 came in, namely OpenLaszlo.

Then Jérôme Cordiez proceeded with a small demo which showed something very interesting and reminded of Javapolis 2007. There was a lot of animosity there between JavaFX and Flex advocates, the most prominent JavaFX advocate being certainly James Gosling himself. And it was funny because the only answer Sun seemed to have against Flex was like “everything you can do with Flex, you can do it with Java, and even more”. But of course the problem is not what is possible, but what is easy to achieve. And Flex is far far ahead of Java on that topic. And even though Sun perfectly understood that and probably launched the JavaFX initiative to solve that very issue, I think they’re still too late.

But why is it so important to be able to achieve advanced rich features as easily as possible? Well, that’s precisely what Jérôme demonstrated today: because you lose less time on putting things together, you have much more time to be creative. And as a matter of fact, Lovely Charts is not just a Flex Gliffy counterpart, as it proposes a whole new way to create charts with a very interesting “create and connect” tool. This allows you to basically create your chart very fluently without having to change your tool continuously. That’s the kind of feature you wonder why it was not included sooner in other charting tools, including Visio, Omnigraffle and others.

Hopefully I’ll get my beta login soon so that I can test it further.

AndroMDA, an Open Source MDA toolkit

I gave a quickie at Javapolis today about AndroMDA. It was a first for me, and boy it was hard! My goal with this presentation was to act as a counterbalance to last year’s scripting language hype, to show that yes, you CAN be productive with traditional Java frameworks. It’s just a matter of tools and methodology.

And because I wanted to demonstrate that, I actually challenged myself into developing a full-blown JSF/Spring/Hibernate application in front of the audience in just 15 minutes. And I must admit that I had my eyes bigger than my stomach on this one. I’ve repeated my demo something like ten times, and I only managed to do it once in less than 15 minutes. So that one was risky. And as a matter of fact, I made just a small mistake in my modeling that actually led me to fail showing my running application.

But just for the sake of precision, and in case you need to see it working, I’ve actually attached the final application to this post: TodoList sample application. And here is what you need to do to build it:

  • You have to install Java 5, Maven 2.0.7 (I’ve not tested it with Maven 2.0.8), MySQL 5, JBoss 4.0.5 GA
  • Download mysql-connector-java-bin-5.1.5.jar and copy it to server/default/lib subdirectory of your JBoss installation
  • Set JAVA_HOME, JBOSS_HOME and M2_HOME environment variables accordingly
  • Create a database called ‘todolist’, and allow a user named ‘todolist’ with a password as ‘todolist’ to do everything on this database
  • Run the following commands in the unzipped directory:
mvn -f core/pom.xml andromdapp:schema -Dtasks=create
mvn install -Ddeploy

Please let me know if you have any issue or question related to this sample project. Comments are here for that.

And I’ll probably use this sample application as a starting point for a more complete learning thread so stay tuned.

And by the way, if you need more information about AndroMDA itself, please refer to their official website. The AndroMDA forum is very active as well, and if you can’t wait to get started, you will definitely love the official Getting Started Guide, which exists in Java and .Net versions.

First you’ve got Agile…

…and then you’ve got Management. That’s the main idea I remember from the excellent university presentation I attended this morning at Javapolis: the Zen of Agile Management, by David J. Anderson.

When the guy introduced himself and told us that he was working for Corbis, privately held by Bill Gates, a great deal of worry went through my mind because since the management fiasco with Vista, I don’t really see Microsoft as a reference in terms of agile project management. But then I remembered that Ken Schwaber has worked for Microsoft too. And it took Mr David only a few minutes to remind me that you can’t judge such a big company by only one of its mistakes, even such a big one.

First of all, it’s the first interactive presentation I’ve attended at Javapolis. Sometimes people can get a little disturbed when you ask them to bring their own content into the presentation. But at Axen, that’s something we are really promoting in our trainings, so we find it far worse when we do nothing. Here, there were 4 or 5 exercises that the audience were asked to do collaboratively. We gathered in groups, and discussed given topics and after 5 minutes, he came back to us with a microphone and we talked about what we had come up with. Interesting idea, especially for such a 3-hour presentation with such a big audience.

That was for the “layout”, now for the content.

There were really two parts in this presentation: first you’ve got Agile… and that was the main point of interest for me. First of all, he insisted on the importance of trust with knowledge workers, defining knowledge workers as people who know how to do their job better than their bosses do. And I love that definition. The thing is that sometimes it’s difficult to forget about the traditional role of a project manager, to remember that there are better ways to have visibility and control over the project than micro-management. And it’s even more difficult to limit yourself to just being a facilitator rather than a leader, to get out of the way when the team doesn’t need you and be there only when they do. That’s what a Scrum master is all about. And trust is obviously the basis for it to work.

I also saw a couple of very interesting usages of cumulative flow diagrams, reporting artifacts that I have never used before, but which give some interesting information. The main point there was that the best way to optimize your interventions as a Agile project facilitator is to consider a few metrics, but only the good ones, the best one according to him being the Lead Time, in other words the time it takes for a feature/task/use-case/whatever to go through your workflow from start to end. The thing is that if you implement things faster, you see bugs earlier so there are less of them, so you lose less time fixing bugs and you do things faster, and the virtuous circle goes. And how do you implement things faster? Simple! Do less in parallel. That’s what of the main things that I loved about this presentation: David destroyed many common beliefs, and at that moment, the way he explains this, it seems obvious. And it’s only when you think about it for a minute that you realize how most project managers do things the wrong way.

The second part of the presentation was more about making decisions at a higher level, analyzing metrics, identifying bottlenecks and reacting properly with surgical interventions. Sometimes I was a bit lost with all those numbers but it made it obvious to me that there should be more project managers at conferences like this. One of the exercises was about identifying the bottleneck in the project we’re currently working on, and what we would do about it now that we have some leads. I know exactly where the bottleneck is on my project, but what am I gonna do about it? That’s another question.