Win a copy of Microservices Testing (Live Project) this week in the Spring forum!

Richard Carr

+ Follow
since Apr 15, 2003
Cows and Likes
Total received
In last 30 days
Total given
Total received
Received in last 30 days
Total given
Given in last 30 days
Forums and Threads
Scavenger Hunt
expand Ranch Hand Scavenger Hunt
expand Greenhorn Scavenger Hunt

Recent posts by Richard Carr

For what it's worth, I submitted two class diagrams.
My facade session beans appeared in both diagrams.
The main diagram centred around the domain model classes.
On the second diagram I showed the relationships between various ancillary classes.
I wanted to keep the domain model to implementation mapping clear, but also show significant other classes I thought necessary for the implementation.
BTW, it's four weeks tomorrow, and no result yet.
I think most of the patterns remain.
But Local Interfaces remove much of the motivation behind Composite Entity.
See a Sun Description
The question could be better worded, but yeah,
I'm pretty sure it's Mediator.
All the messages I've seen seem to indicate that grading part II takes all of the four week period allowed, or even a bit more.
But maybe only those with something to complain about mention it?
Has anyone had a grade in much less than 4 weeks??
I just want to know if it's worth checking before then.
Having done Part III and uploaded Part II more than a week ago, should I see some evidence of this in certmanager alongside my Part I result etc.?
Or will there be no mention of it until the whole thing is marked?
Poor performance from CMP does seem to be an old idea.
In the second edition of Mastering EJB, Ed Roman writes "CMP entity beans, if tuned properly, are much higher performing than BMP entity beans."
Unless things have changed, I think you'll have to buy it. Only available as far as I know in paper-and-ink format.
I seem to remember I bought it on Amazon.
I have 12 diagrams.
I have 7 sequence diagrams. I have a diagram with my own version of the domain
model, a main class diagram, a supplementary class diagram, and a component
diagram. I also have a separate navigation diagram for the JSP's.
The largest are the sequence diagrams. These I drew at the smallest scale
that was readable. Some of the others I could have made smaller - but I was
satisfied that they fitted on a page OK.
The largest sequence diagram in dimention is 1630x1620 pixels, but with just
a little scrolling it seems pretty readable. Maybe I'm too influenced by my
quite hi-res 1600x1200 screen.
The largest file size (another sequence diagram is 91k)
I have about 90K of HTML commentary and assumptions organized into a few
documents. This might well be over-the-top .
P.S. The part III exam wasn't too bad. We'll see when they sort out letting me upload the assignment.
If that question is for me it will have to wait till I get back.
I'm just walking out the door for the part III exam.
I had no problems with PNGs in any of the browsers I tried - Netscape 7 & IE 6.0 on Win XP Pro,
and Mozilla (whatever version comes with Red Hat 7.3) on Linux. But none of these is very old.
However these diagrams don't depend on fancy PNG features, and according to a PNG resources site there has been support in the main browsers since 1997.
Anyway, unless you want to squeeze that last few percent out of your images GIF should be fine.
I emailed to complain about lack of permission at 01:33 this morning (GMT +2:00).
I had already seen your message and Chris' response, so I also sent the message to 'Correspondence'.
In addition I sent it to the Sun contact I had in the UK.
I was worried, because my Part III exam is booked for tomorrow. I assumed that I had to upload the assignment before the exam.
But my Sun UK contact tells me that I can do the exam before uploading the assignment.
I've had no response yet from the US about permission.
But when I had an earlier problem, Sun UK's prompting seemed to speed things up in the US.
If you're not in the states maybe you should contact Sun where you are.
To finalize my diagrams I was using Rational XDE which also
outputs JPGs.
I believe that JPGs are meant for photos etc. They give poor compression
for diagrams. It seems that for diagrams you do much better with PNGs or
Initially, as I was using Windows (no choice with rational XDE), I just
loaded my diagrams with Paint and saved them as GIFs. This makes for a
dramatic improvement. I'm not sure if Paint does PNGs.
After I finalized the diagrams, and since my jar was still a bit big (<800K)
I switched back to Linux and used the GIMP to convert the GIFs to PNGs. This
saved another 10 to 20 percent.
I'm now planning to upload (whenever they sort this permission problem) my
630K jar.
Sun's comments on the file size are a bit strange. They say that > 1M has
caused problems for people uploading, so try to keep to 200-400k (by
optimizing graphics).
Hope this helps.

I'm close to submitting Part II.
To keep the component diagram managable I plan to have just a couple of token 'Any JSP's.
I have a separate navigation diagram showing all the main JSPs and the route through them for most of the use cases.
On the component diagram I also show only the most important entity beans - those accessed directly from the session beans and the one BMP bean. All the beans and their relationships are shown on the class diagram.
Is this a bad idea? I've read quite a few comments that absolutely everything has to be on the component diagram.
Can I declare a unidirectional CM Relationship from
a CMP bean to a BMP bean?
I will have a complex of CMP beans persisted in one database. I have to make use of some existing data in a second DB and plan to encapsulate that as a simple BMP bean (I can't modify the schema). I'll want to access that BMP bean from one of the CMP beans - a CM relationship seems the simplest solution.
Or must the targets as well as the sources of CM relationships be CMP beans?
You found a couple more mocks than I did.
I took the P1 exam last Wed (9th April 03). The
mocks were very useful for finding gaps in my
knowledge, but not a good indicator of my final
My mock scores in the week before I took the exam varied from 67% to 88%. But by that time most of my 'errors' were actually errors in the mocks, or due to ambiguities/lack of clarity.
Not everything is crystal clear in the real exam,
but it's much better. I ended up with 93%.
I seem to remember a similar situation a couple
of years ago when I took the Programmer exam.
Still I'm very grateful the mocks are out there. We can't expect the authors of mocks to spend as much time proofing questions as I hope Sun does.
Note also that some of the mocks are much better than others.
Good luck,