Jay Yan

Greenhorn
+ Follow
since Feb 04, 2004
Merit badge: grant badges
For More
Cows and Likes
Cows
Total received
0
In last 30 days
0
Total given
0
Likes
Total received
0
Received in last 30 days
0
Total given
0
Given in last 30 days
0
Forums and Threads
Scavenger Hunt
expand Ranch Hand Scavenger Hunt
expand Greenhorn Scavenger Hunt

Recent posts by Jay Yan

1. This piece of code is too trivial. It does not have to appear in your sequential diagram.
2. The creation part shall be in an AFactory class. And you can just show A and AFactory in your diagram, as no matter what kind of A it is, you use it in the same way anyway.
3. If you really want to show the branch logic, use UML2's "alt" frame.
Although only SCEA 5 is available now, many people are still on the track of the old SCEA, as they entered there last year. The upgrade is about to upgrade from the old SCEA to SCEA 5.

SCEA 5, the only currently available track for SCEA, is not the upgrade, but the whole set.

JPA persistence is not much covered in the exam. Design patterns are heavily tested in SCEA test, especially in part I.
I used MS Visio 2003, with a Stencil and Template for UML 2.0 http://softwarestencils.com/uml/index.html. UML 2.0 is still not supported in Visio 2005, but that Stencil and Template can do that job beautifully.

If you have Visio already, that's definitely an option. If you don't, I won't suggest you to purchase Visio just for this.

Originally posted by Jay Yan:

I emailed Sun Certification Support [certification@educationservicesgroup.com], ....



I emailed who2contact@sun.com, and got reply from that email address.
Thanks for replies.

I emailed Sun Certification Support [certification@educationservicesgroup.com], and got this reply:

"We apologize for the error. The results are entered manually and obviously there was a typographical error. We have corrected your score."

Entered manually? I guess they need a system to get the score summed up automatically. I can design the system for them.
I submitted part II and took part III early Feb.

I was told to go to certmanager.net/sun to check the result, but it's always "Pending". I happened to visit https://www.certmanager.net/sun_assignment/ today, and found my grade has been there since 3 days ago. Hmm...

Anyway, according to that site, my scores are:
Class Diagram (44 maximum) .......................... 39
Component Diagram (44 maximum) ...................... 40 Sequence/Colloboration Diagrams (12 maximum) ........ 10
So, the total should be 89. But on the same page, it says: Score: 82.

Anybody experienced this before? Is that because of part III? No matter how, PASS is what matters.

==================
The only book I used during preparation of Part II and III was: Sun Certified Enterprise Architecture for J2EE Technology Study Guide (Sun Microsystems Press)

I used MS Visio, with a Stencil and Template for UML 2.0 http://softwarestencils.com/uml/index.html (I paypal the developer a little money to appreciate that.) I could not draw a good loop or branch just using Visio itself.

To prepare part III, basically I answered the following questions by myself before the exam: "How does your system handle xxxx", while xxxx are those "--bility" concerns as the SCEA book listed, and key features like session/transaction/security/state/.... I found most how-to ideas on web.
Congratulation!

For part 3, how many sentences or paragraphs did you use to answer each question, averagely speaking? And I guess you did not use the whole 90 minutes, right? I just want to find out how much details I will need to provide.
Congratulations!

In part 3, how long did you answer each question? (like how many sentences or paragraphs, generally speaking.)
To answer questions like "How will your system scale?" (Just an example I imagined. I've not taken the exam yet.), I can write a couple of sentences to point out some aspects, or I can write a three page document explaining everything in details, which probably will annoy the grader.

So, a question to you who have passed it, how long should I answer each question? For example, is the length of the current question good enough?

Originally posted by Prabhu Venkatachalam:

After done with the exam, how do feel about Design Pattern Questions. Do think really worth in reading GOF for SCEA or Head First Design Pattern it self OK?

Thanks in advance.



I passed SCEA p1 a month ago, 91%. HF Design Pattern is good enough. But if you still have questions about some patterns, or you think some patterns were described too brief (like those two-page-patterns at the end), GoF is a good reference book.

Originally posted by rania ferrag:
Where is that Chapter at Cade book ?

So definitely I�ll rework on Mark Cade book, and reuse the Whizlabs (I had use them all before my exam)....So in your opinion Cade book and Whizlabs will be enough for my review ?

I have a question regarding the Survey we have just before the real exam start, does the survey have any relation with the exam ?



Cade book has two parts, Overview talks about IDL/JNI/etc., and Protocols talks about JRMP/IIOP/etc. Also did that book talk about Screen Scraper or I learned that from Wizlabs? Learn those parts by heart and you should be OK. You may spend an hour reading that blue book by Paul on this topic, but that book real covers unnecessarily too much.

Besides Cade book and Wizlabs tests, you will also need to read about EJBs more. EJB is not well covered by Cade book.

I believe that the survey is for Sun's own statistics. It is related to your test only in their own reports. It should have nothing to do your score.

Originally posted by rania ferrag:

The Question on Design Pattern ( had 80 % ), protocols (100 %) Internationalization (100 %)Security (54%) Applications of J2EE 66 %, Common Architectures 66 %
And thes bad score on EJB 44 %, EJB Container Model 25 %, Legacy Connectivity 20 %
Yesterday I had took the last exam from Whizlab and had scored 80 %


My own experience for your reference: I just used HF Design Pattern and Wizlabs and got 100% on Design Pattern. I used Cade book and Wizlabs for security, which got me also 100%. I used HF EJB to prepare SCBCD two years ago, so I did not spend much time on EJB topics this time, and got 100%. My Legacy Connectivity part was prepared with Cade book and Wizlabs, and 100%.

And I scored just 73% on the last Wizlabs test one day before my test, which turned out to be 91%.

So, I think that you do have enough books now. I would suggest you to spend more time on the Cade book, which is not too thick. For EJB, just state changes of those 4 kinds of beans, and pooling related topics. And after you take a Wizlabs test, read carefully the explanation of each question, and understand WHY each decision was made regarding what to use questions. And some time Wizlabs' questions are wrong, so read those with doubts please.

Originally posted by Ronnie Phelps:
Do you think that you could've answered the design patterns questions without the extensive studying of HF design patterns?



Well, I was not clear about the difference between Factory method and Abstract Factory, or the difference between Strategy and Template. So, my answer is negative. But I did not do it "extensively" though. Just a few hours each week. And it does not have to be HF. However, HF is much easier to understand. GoF book is nerdier.

BTW, I also used Pete Morgan's SCEA Part 1 Notes found on web. Thanks to Pete.
I just passed Part 1 today 91%. One error each in four areas.

I passed SCJP and SCBCD years ago with 9x%.

This time, I spent a few months on design patterns using Head First Design Patterns book. That's more a learning experience than exam preparation. Good book.

I spent about a month with Mark Cade and Simon Roberts book. Good book, but it does not cover 100% of the exam scope.

I spent several hours reading the Paul Allen and Joseph Bambara book. Waste of time. That book covers too many topics (good for your knowledge though), and many errors in those quizs.

I went through Whizlab's preparation kits in the last two weeks. Among those 9 tests, I scored 3 a little below passing line, 2 in 7x%, and 4 in 8x%. Generally, each Whizlab test contains more than 10% of ambiguous questions and wrong answers, and less than 10% of questions totally out of scope (like a question about a telecommunication concept virtual channel). That's why I scored finally higher than in all those mockups. I think the coverage of the those exams is about 90%.

My suggestions for people who are using those Whizlab's kit be aware of the existence of their bad questions, and read the explanations carefully, which are more accurate usually than the standard answers. My final exam, taken yesterday, was 73% only. So, don't panic about low scores with those Whizlab tests, as far as you know what's wrong and what's correct.