I passed the exam part just last week. I got a decent score but far from my score in SCJP and SCWCD. I intend to get better score in Part 2 and 3.
Just my thoughts on Part 1.
- First, the coverage of the exam is so vast, unlike other certifications where you focus on one or two technologies. This exam test you on almost all aspects of computing technology. Although not as thorough as the programming level, even going for the API level is not that easy.
- The questions are really not clear-cut on finding the correct answer but more on the best and most appropriate answer. There are only few questions which ask for specific term or technology. Most questions are hypothetical and scenario-based. And I believe this is where I found the exam not very convincing. As a technical person, I am geared towards objective/true/clear/concise concepts. I find it difficult to work with hypotheticals or if/but/maybe.
- I've studied (wide and deep) on all the technologies that were mentioned in the exam objectives down to the code-level. So I know how to implement them, not just in theory but on the programming level. But it seems to me that the exam is geared towards the philosophy behind the technologies. And again, I am not a big fun of philosophy. I guess this is why architects are being labeled as not in-sync with reality, or having their feet off the ground.
- Beware of the mock exams. I bought whizlabs mock exam. There are so many questions that are outdated. I've tried some free mock exams but I find the questions also outdated. You have to be very aware of the API version. And this is where whizlabs failed. They are asking questions which is not in JEE 5. They have lots of questions which are remnants of previous exam. And there are questions which are grammatically and semantically confusing. I hope they fix this soon. It's just not worth it to pay $100 for outdated questions.
- It's easy to pass the exam, but it's hard to get a higher score. It's because of the nature of the questions, the way they are posted is not exactly what you might expect for a clear answer. As an analogy, if you're taking a math exam, if the answer is 6.54, that's it. But here, you have to choose 3 answers from 6 choices which most of them also satisfy the same question. And this is where it becomes subjective especially if you have varied experiences in applications, technologies, and industries.
But all in all, I benefited a lot from the review. It cleared a lot of my gray areas in technologies and considered aspects of development which I never paid too much attention before (e.g., security). I consider myself a better Java EE technologist now than before.
I am gearing up for Part 2 and 3. I hope I could do better on these sections because this is where the heart of the architecture is.
Now I realized that you don't get a score (%) for Part 2/3.
So you just know you pass or fail. But you won't know how well (or bad) you did.
This again adds up to the subjectivity of this certification. There is a huge gap from the passing rate of 71% to 100%. So if you pass, you're safest assumption is you got at least 71% - that is not a very pleasing assumption.
That leads me to another question. What's the use of the ff criteria in the assignment?
Component Diagram = 40
Class Diagram = 40
Deployment Diagram = 24
Interaction Diagrams = 16
Risk & Mitigation List = 16
Part 3 Short Answer = 24
Total = 160
Hmm... I was hoping I could use the certification as a proof to my current (or prospective) employer that I am really good (or better than non-certified) at architecting and designing an application. Instead I can only prove that I can do it but barely.
I just wish Oracle will change this policy and make this certification quantifiable (i.e., scored) rather than qualifiable (pass/fail).
Congrats on passing part I and thanks for the great tips in your first post . What books did you study from? I don't think studying from just one book would be enough, or would it? what do you recommend?
I share your toughts about part 1. There is subjectivity involved in the exam's proposed answers, its much different than 1+1= ?. But it reflects the kind of decisions that an architect has to take.
I'm surprised that you think that a high score would convince your boss that you are a better architect. If I was an employer and I had to choose an architect, I would base my decision on that person's experience, reasoning, technical knowlegde, "soft skills",... certainly not on an exam score.
Same is true for all certifications IMO: they validate a minimal level of techincal knowledge on a subject (a very broad and deep subject in our case), but they don't tell that you are "good" in any way. In my opinion only successful projects validate your skills.