Hi all:
I relied on many of your posts as I was studying for the OCAJP 7 exam, and I took it (and passed it!) today, so I wanted to create an account and share some thoughts. This version of the test had 70 questions in 120 minutes with a passing percentage of 63%.
Preparation:
I spent about 10-12 weeks on and off studying for the exam. Prior to that, I had absolutely no Java experience and about two years' worth of developing in C++. I wanted to teach myself Java since it seems to be the language of choice in industry, so I jumped in.
I used Mala Gupta's text as my primary reference, and it turned out to be valuable for both learning Java as well as studying for the exam. I also liked her fun analogies (buying diamonds in a try block, coffee in a finally block... classic). Be warned, however, that her end-of-chapter sample exam questions and full mock exam were easier than the actual exam. That said, I actually pre-ordered her upcoming book for the next level of Java certifications because of how informative and easy to read it was.
After studying my tooshie off, I also purchased the Enthuware (by Hanumant Deshmukh) set of exams, and I worked through Exams 1-6, skipped 7 (since it was marked as "Advanced"), and then took the Last Day Test. It was a good purchase. Exams 1-6 were HARD, but I learned a lot. I am very impressed by the quality of Enthuware's answers, and I highly recommend you purchase it regardless of your primary test reference, especially since it's only $10.
In terms of difficulty, I would rank Gupta's as the easiest, followed by the Last Day Test, and then Exams 1-6 as the hardest (in some order).
Here were all of my scores:
Gupta sample exam questions after each chapter: 81%
Gupta full mock exam: 80%
Enthuware Exam 1: 81%
Enthuware Exam 2: 78%
Enthuware Exam 3: 78%
Enthuware Exam 4: 78%
Enthuware Exam 5: 78% (notice a pattern?)
Enthuware Exam 6: 75%
Enthuware Last Day Test: 88% <- this gave me hope
I also was caught up in reviewing my incorrect answers relentlessly. In some cases, my mistakes were silly, but usually, I learned something new. By yesterday, I was able to go through Enthuware's first four exams, of which I hadn't taken for more than a month, and nearly effortlessly could correctly answer the vast majority of questions.
The exam itself:
First of all, I had a fly bugging me for the first 30 minutes. That was annoying. I wish you a fly-free test environment.
Honestly, my second biggest fear (first to come later) was time. After having a very consistent ~80% on the various mock exams, I figured that I could answer ~80% correct on the real one, too, and other posts that I read suggested that time was an issue. As I started, though, the first 10ish questions were done in under 5 minutes, the last 10ish questions took about the same amount of time, and I only had to guess on one of those ~20. That gave me a lot of extra time to go back and review the middle 50ish questions. The ones in the middle were largely more detailed, and some required extensive scrolling through code on the screen. I actually finished the exam with about 35 minutes to spare, and I used all but a few seconds of those 35 minutes going over the middle 50ish questions.
A hint: Mark every 10-15 questions as for review. Since I needed to jump back to the middle 50ish questions at the end, it was a lot faster to jump to a question number near my target than to click Previous/Next a bunch of times. A second saved is a second used for thinking.
Make sure you pay attention. Both Gupta and Enthuware taught me to look out for silly compilation errors, but the exam was more sophisticated. You're not going to pass by assuming that you can pick out obvious compilation errors. You'll need to fully understand the exam objectives.
Speaking of exam objectives, I was pleasantly surprised to see that the real exam followed the objectives exactly and thoroughly. My single biggest fear about the exam correlated with my sole issue with the Enthuware product: in Enthuware, many questions contained enums, transient/volatile/synchronized/native, wrapper classes, etc. with Enthuware caveating them as saying that some candidates received those types of questions on the exam. I did not; the real exam went step by step through the objectives, covered them all in-depth, and did not stray away. Of course, the Candidate Agreement prohibits disclosing the test contents, so I can't answer any questions about the contents of questions they asked. If you know the fundamentals specified in the objectives, though, you'll be set. The writers of the exam did a remarkable job making sure that the candidate knows the inside and outside of every objective. The test was extremely well-written.
I walked out of the exam thinking that, if I split 50/50 the questions I narrowed to two answers, I would have received a score in the low 80% range. In reality, I received a 91%. During the exam, I thought that it seemed a comparable difficulty to, or slightly more difficult than, Enthuware's Last Day Test, and my score suggests that as well. It was nowhere near as difficult as Enthuware's Exams 1-6, thank goodness. If you can pass those and understand where you went wrong on your incorrect answers, you'll do great on the real exam.
Best of luck to all of you.