Matt Dole

i passed in early december of 2003, still no certificate and the same canned response from sun mentioned earlier. very unprofessional. after all the hard work i put in and the hundreds of dollars spent for the exam!!! We need to organize and draw attention to this. slashdot anyone?

i passed p2/p3 early december and *still* did not get my certificate! just wondering if anyone who passed since then got it yet. this is very unprofessional of them.

give this a shot in your web.xml
<error-page>
<error-code>403</error-code>
<location>/mycustom403.jsp</location>
</error-page>

certmanager is not letting me upload my part2 solution and the autorespond out of office email says they will respond in 48hrs but I need to finish the certification this week. can i take part3 before part2 is uploaded?

good job! i should have taken your approach of "just get it done" long ago. i am taking that approach now after suffering through analysis paralysis and trying to make this assignment a bigger learning opportunity than it should be.

I've looked through a few books and have not seen how to model this.
what is the proper way to develop a sequence diagram for a use case that << include >> another?
for example, [distributor search] --- << include >> ---> [login]
I am using analysis classes in my seq. diagram classified as either boundry, controller, or entity.
for [distributor search], do i have a class called [login] with sterotype << usecase >> with a message going from my controller to [login] labeled << include >> ?
for [login], do i have a class labeled [distributor search] << usecase >> with message going to the entity? should it instead go to the controller?
[ November 03, 2003: Message edited by: Matt Dole ]

did you supply sequence diagrams for the use cases which were not given a description - for example log in?

I have the same question. the class diagram is technology neutral and the component diagram is not. Should the sequence diagrams be technology neutral or use what is shown in the component diagram?

ttt. i'm also interested in the proper way to depict this in uml.

I'd like to try suse and maybe mandrake but i mainly do java development with weblogic 7, jboss, tomcat, & eclipse. would there be any problems developing on a distro other than redhat?

my job description use to be design, develop, test, java applications. i would implement business requirements using java. my employer changed direction and has gone towards using 3rd party tools instead of custom software. now my job is now to analyze a business process and re-engineer it to use the 3rd party tool. of course, i have to learn the 3rd party tool and implement the new process using it.
career wise, do you guys think i am better off now? My knowledge of this 3rd party tool will be useless as soon as the next version comes out. java knowledge would be useful longer and for more companies but it is starting to be seen as a commodity. maybe the business analysis skills are more valuable than java coding or knowledge of this 3rd party tool?

to those who passed part2, did you draw your component diagram following the example in mark cade's book? I have read many people put .jar files on their diagram.

i understand that a PreparedStatement is precompiled but where is it stored and what is its lifetime?
Is it stored by the database and the database manages its lifetime? Thus the db maintains a pool of PreparedStatement objects that it refers to?
Is it a regular java object thus obeys the standard java object lifetime rules? If so, would there be any benefit to using a PreparedStatement object instead of a Statement object in a method that created a method local PreparedStatement object and used it once?

we use a data access layer. they contain all our db access code. These are commonly used to centralize your db code. other objects call methods in the data access object.

i have a method that needs to return a collection of objects, say Customers. each row of a table represents the fields of the Customer object. performancewise, is it better to have the method return an array of Customers (ie Customer[]) or a java.util.List/java.util.Vector containing Customer objects?
I personally like the method returning Customer[] because it makes a cleaner method signature.
since i don't know the number of objects that will be returned beforehand (size of array to allocate), i could do 2 queries: select count(*) and the data gather query. Alternatively, i could use 1 query, make a List, then call List's toArray(Object[] a) method. Is there major overhead with that method?
How do you guys deal with methods that return objects created from db queries?