I have found Dan Haywood's article on TSS titled "MDA: Nice idea, shame about the..." very interesting. It's kind of a high-level analysis of where MDA is today and how it will "probably" evolve in the future. Dan also tries to answer very relevant questions as to why an organization would adopt MDA (or not) in the future.

Although I was already aware of the fact that there are two different actors on the MDA scene (elaborationists and translationists), the following statement hit me right in the face:

"At a minimum what this means to MDA adopters is that they should not expect PIMs developed with a tool from one camp to be usable in any form in a tool of the other camp. And the fact that two camps exist at all is disconcerting, to say the least. There ain't ever gonna be such a thing as a standard MDA tool."

If you come to think of it, this is really scary. MDA is still at an embryonic stage and there are already two very well demarcated camps who think they are on the right path. This is understandable in some way since usually when a new technology comes to life, everything has to be created and it is sometimes not easy to get people to look in the same direction and share the same opinions. Still, so-called *MDA-compliant* (whatever that means) tools already exist on the market. They comply to MDA but not to each other. Anything wrong here?! As an analogy, think about the chaos if a Java program written with JBuilder could not be opened in NetBeans. This is often the case when you use vendor-specific libraries (X-Y layout, etc) for building up your applications.

I would just like to continue on the same subject I started yesterday and give some more insights on the ongoing discussion.

Basically, Francis says: "Stop wasting your time with the SCBCD exam since EJB 2.0 will soon be overhauled by EJB 3.0. Better learn Spring and Hibernate instead."

While his statement his comprehensible up to a certain point, it does not make much sense given the way things work and the pace at which technologies evolve. JSR 220 mentions that EJB 3.0 will be released with J2EE 1.5 which has just been filed under JSR 244 and is not due until late summer 2005. As pointed out by Kathy Sierra in the same thread, "it [EJB 3.0] just won't be mainstream for a long time, and at Sun we want people to certify in something that is practical and useful for at least the next 18 months". She further adds that "...the main point is that EJB 1.1 and EJB 2.0 are what customers are using *today*, and will be continuing to use [...], for potentially another 18 months to two years or MORE".

So the bottom line is that the perceived value of a certification should principally lie in its practical applicability given the current industrial settings and community acceptance. Nobody in the thread argues that Spring and Hibernate are bad (I'm a fervent supporter of both), just that learning EJB 2.0 is neither a waste or time nor money given the fact that EJB 3.0 will be backward compatible.