Jurgen De Jonghe

Coordinates:

Main activities:

CERN, never shy of being ahead of the curve, was standardising on SGML when I started working here in 1987. We were developing SGML parsers to load documents into a relational (Oracle) database and to translate SGML into LaTex for text processing on our IBM mainframe. I was the president of the Swiss Chapter of the SGML Users Group. It's funny that in this millennium, people are still tackling similar problems with XML, albeit not on a mainframe for most of them…

I then became responsible for the automation of business processes at CERN, and managed the development of the EDH workflow application, now the most widely used in-house developed software in CERN. The EDH software went through 3 different iterations: version 1 (1992) was a pure Oracle client-server application built in C for PC, Mac and Unix. It worked fine, but lacked scalability: each new business processes required a change in the client which was increasingly expensive to develop and re-deploy.

During 1994, we embarked on a more ambitious client-server model: the client became a pure rendering engine in C++, "unaware" of the business processes and all business logic remained on the server. This version was actually very similar to "the web" but unlike http, we used a stateful protocol.

The final version started somewhere in 1997. By then, we had about 1000 users for each developer (a team of 4 people) and it was getting hard to maintain our own client-server as well as supporting the client base with ever more requests for new business processes to support. During that time, I attended a course at IBM on their 'San Francisco' project which was really Enterprise JavaBeans avant la lettre. I was impressed with their component framework (a little less with their actual implementation at that time). We decided to port EDH to the web, using Java Servlets and modelling our components after the EJB component model, but without using the (then still unreliable and slow) "Java Enterprise" application servers.

After 6 years of managing EDH, seeing that the web version was well under way, I decided in 1998 to move on to a new and exciting project. My current responsibilities, from late 1998 onwards, are to lead a small team of developers to provide a Project Progress Tracking tool for the assembly of world's largest particle detectors, like the ATLAS or CMS detectors. These are pluri-annual, multi-million Euro projects, involving hundreds of partner organisations all over the world... an ideal Internet application! The current version of PPT is developed with Oracle Designer, using the web generator facilities (and our own extensions).

Around the end of 2000, I wrote a chapter of the book J2EE Technology in Practice: Building Business Applications with the Java 2 Platform, Enterprise Edition together with my collegue Derek (who is currently leading the EDH development). It describes how we intend to further migrate EDH to J2EE.

We're currently discussing internally whether to rewrite PPT on top of the J2EE platform, trying to reach at least the same level of efficiency as using Oracle Designer, but with UML modelling ... quite a challenge!

On a more personal note: most of my time goes to my family. But I do enjoy sports when time permits: squash, mountaineering and running especially. I also attempt to play the bass clarinet in the "Echo du Reculet" in Thoiry.