Andreas Ebbert-Karroum

Codecentric is main sponsor for Java Forum Stuttgart, a one day conference with well over 1000 expected participants. Besides demoing live performance testing of a JBoss cluster with dynaTrace diagnostics, we are giving away a great Spring Source Toolkit (codecentric is spring source partner) and promote our trainings for Core Java Performance and Core JBoss Technologie. Carsten will give an in-depth session about Advanced JBoss Cache.

Have fun, everbody!

After enjoying the luxury of being exempted for two and a half months, time I spent well with my family, today was the day that appeared to be so far away in mid of April: today was my first day at my new employer: codecentric GmbH in Solingen (Germany).

Reasons to leave my previous Company (Nokia Siemens Networks) were numerous, the final trigger was that the R&D department of my business unit in Düsseldorf  was transferred to Tata Consultancy Services.

After some search I finally found codecentric (it was the other way round, to be exact). We offer consultancy, software development and training while specialising on Java performance, architecture and open technologies. But more than the technological orientation of the company, the personality of the founders, the athmosphere and company culture convinced me to switch to codecentric - and as far as I can tell after my first day, this was the right decision.

I’m looking forward towards my new challenges, which I even can adress together with one of my former collegues from Nokia (Fabian Lange) , who happened to do the same switch. Finally, working is fun again

IT projects have generally a high risk of failure if not managed and executed properly. The risk is even higher, if the project is to define a replacement for the current IT architecture, to overcome agility, scalability, security, extensibility, management or performance impediments. The Service Oriented Architecture is hyped to deliver that, what CORBA and EJB failed to provide, to be the new blanket for hiding the architectural chaos underneath.

When a new technology (or better a methodology) like SOA is not only difficult to master on it’s own, but also hyped (means, proposed by people, who have never written a line of code in their life), it is not a surprise, that “there are instances, where projects are going well, but most are over budget and under delivering, ” says David Linthicum from Real World SOA. He even proposes a money-back guarantee to share the risk, if the consultants take the training on the job on the client’s dime. As this is clearly a little over-provocative, he has a point that it is easy to fail in your job to implement a SOA sucessfully.

Rik de Groot, Viktor Grgic, Vincent Partington, and Gero Vermaas from Xebia posted the top 10 SOA Pitfalls for you to avoid the most common mistakes. They categorized these into implementation, architectural, and most importand also organizational pitfalls. These are the most dangerous, since with SOA also has to come a shift in the organizational culture, which reacts ponderous, sluggish and unwieldy to any attempt to change it. Thanks for the excellent and insightful postings.

With the knowledge about these Top 10 SOA pitfalls, maybe you can give your clients a money-back guarantee!

Machiel Groeneveld from Xebia blogs about how the Product Owner role should be filled out to get the most out of Scrum (Scrum The Mythical Product Owner role | Xebia Blog), and contrasts that with the reality that is met in most projects, where the product owner is a side job next to being a program or project manager, a release manager or information analyst. He concludes that it might be difficult to start off with Scrum, if the requirements to such an important role are so difficult to meet. While I agree with that conclusion, I’d like to add that in this case (and many others too), Scrum just reveals the problem and makes it visible, so that it can be adressed. The problem of no clear representative of the customer business case also exists if you don’t know Scrum, but it may just be hidden.

I’m using Google Reader for quite a while now and am still happy with it. A nice feature is that I can easily share items that I find interesting and valuable to read with others. So if you would like to know what I liked best, either visit the webpage with my recommandations, or subscribe to the feed.

I was just made aware, that Robot, the test automation framework that we used at Nokia Siemens Networks for huge projects, is now open source. Well done! It really has some very usefull features and because of it’s internal deployment scales up to huge projects where the whole test suite takes a good working day to execute and spans all layers of your enterprise application.

Testcases are described in HTML tables, as well as the test results:

Features include:

• Enables easy-to-use tabular syntax for creating test cases in a uniform way.
• Provides ability to create reusable higher-level keywords from the existing keywords.
• Provides easy-to-read result reports and logs in HTML format.
• Is platform and application independent.
• Can natively use both Python and Java test code.
• Provides a simple library API for creating customized test libraries.
• Provides a command line interface and XML based outputs for integration into existing build infrastructure (continuous integration systems).
• Provides support for Selenium for web testing, Java GUI testing, running processes, Telnet, SSH, and so on.
• Supports creating data-driven tests.
• Provides tagging to categorize and select test cases to be executed.

Now the only thing missing, to avoid that others have to go through the same pain as we had to, is the Robot Development Kit. That’s an Eclipse/EMF based editor for anything you need for Robot. It solves all the pains that you have with HTML editing, merging HTML in a version control system, etc. I hope that will be open sources at some point as well!

While I’m still trying to catch up with all the information gathered while JavaOne was happening (still 1000+ unread blog posts according to Google Reader), now also all presentations for technical sessions and hand-on labs are available for free (except for registering for the Sun Developers Network). I’d love to see the BOFs available somewhere too, some of them are scattered in some blogs if people cared to upload their slides, but this could be more organized. Presentations include a PDF and for some (and later more) sessions also audio and transcriptions. Some JavaOne ago it was suggested that the community could provide localised transcriptions - has that ever happened? Additionally, when will be able to see the live demos that were given during the session?

I’m in the unfortunate situation to actually have a full conference pass for JavaOne 2008 in San Francisco, complimentary of Sun for their Star Spec Leads - thanks Liz for all your support over the years, without actually being there (for various reasons). So how can you follow the conference as close as possible, without actually being there?

1. Reading Blogs - this is a recommendation of blogs to follow to keep up with what’s going on at JavaOne.
   • JavaOne 2008 Blog Central
   • Blogs from blogs.sun.com tagged with JaveOne2008
   • Blogs at java.net and the Editor’s Daily 
   • The Server Side
   • Use Google blog search for all JavaOne blog posts
   • Use Technorati to search for all JavaOne blog posts -
2. Twitter - apparently the way to follow the conference closest
   • To receive all tweets related to javaone, “track javaone” on twitter
   • Follow some people, who attend the conference, this is a good list to start from.
   • Follow @JavaOne2008 and @DukeJava
3. Webcasts - I wish they would also cast some selected sessions, maybe even only for those with a ticket. (I should propose that to Sun, to sell “virtual tickets” that only entitle you to follow the conference remotely)
   • Watch the General Sessions - live or recorded.
4. Magazines - even if this is the”old-fashion” way of doing it, it’s going to be more reflected and in-depth, I hope.
   • JDJ
   • JavaWorld
   • Artima Leading Edge Java
   • SD Times
5. Hands-On Labs - try out some labs yourself, they are fully downloadable
   • Hands-On Labs

If you have any other suggestions, please comment, I love to hear about them!

Update:

• @JavaOne2008 suggested to include the possibility to join JavaOne on Second Life. I’ve not included it in this blog previously because I’m not using Second Life, but if you are, this might be a good chance to take part.

If you’re developing software using Scrum or XP (or your individual collection of tools and processes) in multiple teams, make sure that you read this article “Version Control for Multiple Agile Teams” about how version control, trunks, branches and policies could be implemented. This is the depicted summary of the great article.

A key lesson that I took from it is that it’s a must to have feature teams. Spanning features (user stories) over multiple teams immediately breaks the pattern, because you not only need to sync each work branch with the trunk, but also sync all work branches that work on the same feature. Given you have n teams, you see that you multiply the need for synchronization from n to n*(n-1) in the worst case.

A couple of days ago, InfoQ spoke with Joseph Pelrine about Social Complexity Sience. With Josephs background in psychological theory on one hand and software development and management on the other hand, he now tries to lay the theoretical base for answering questions like: “Why does Scrum or XP work?”.

There are a number of interesting thoughts in the interview, plus there’s also a complete transcript available!