Antonio Fornie

Hello again!

You there're many Use Cases where user first asks for a list of something (let's say a list of cars) and in the next 'click' it asks for the details of a car from the previously given list. So, if we used JSF/EJB we could have a CarBackingBean get the list using the CarEJB (SLSB as a DAO). Here's my question: should the BackingBean keep the list in order not to ask the EJB in future requests (calling DB again) or should my BBean forget about the list and ask for a selected car details in the next request.

I think it's better to use the BBean as a cache, I mean, let it be a Session Scoped Backing Bean to avoid going to database once again. Anyway, it should not be much data as I used pagination so the BB only would use page-length cars at the same time FOR EACH USER... That's what scares me, maybe there are many users and it's better to have a Request Scoped BB and query to DB everything everytime. What do you think?

By the way, if you finally decide to cache this data, would you choose any other place than the BackingBean?

Thank you very much!

antonio

Hello.

I'm finishing my part 2 assigment and I think I can answer correctly most questions about my design. Anyway, I'm starting to forget some of the contents I studied for Part 1 exam. Will I have problems in the Part 3 essay because of this? Will they ask me about complicated subjects from part1 other than why I did my design the way I did it?

Thank you very much once again

antonio

Thank you both. I think I'll finally give some text right behind the image. In fact, I think I'll have a Class Diagram for the whole model with little detail (no methods) and them I will have a couple of Class Diagrams with more detail, one for each part. Otherwise I'd end up giving a single HUGE Class Diagram.

antonio

Hello.

First of all, thanks to all of you.

Dinesh Kumaar wrote:Hi Antonio
I would suggest you to go through the examples in Chapter 9 in Marc Cade's Book.

You can go through these links.
https://coderanch.com/t/426825/java-Architect-SCEA/certification/modeling-JSF-managed-beans-component
https://coderanch.com/t/517797/java-Architect-SCEA/certification/JSF-compliant-diagram-component-diragam

1. You are right
2. From what I read from old post Sun wants to keep class diagram framework agnostic, so just use controller in class diagram.
In Component diagram, you can use jsp/jsf, managed bean ..etc.

I have 4 Backing/Managed Beans, but each one comes for different matter. For example: CustomerMB, SaleMB, ProductMB, InvoiceMB. In this case, would you place them four in the same component or would you create a component for each one?

Let's say I've got this 4 BackingBeans and their EJBs: CustomerBean, SaleBean, ProductBean and InvoiceBean. My first idea was to have two components, one for each tier: the four BackingBeans in a "web" component and the four EJBs in a "business" component. Do you think that's right or I should spread them among different components.

By the way, do I have to say the package name of each class?

Thank you very much.

antonio

Hello.

In Epractize Labs mojos I've seen a couple of weird things, I'd like to share with you:

1º In the Component Diagram they have 1 component for each class or jsp page... Would you do the same? I think this is wrong: We should have one or more classes or one or more jsp pages in each one, otherwise, Component Diagram would be like Class Diagram. For example, if I have three related Backing Beans, I would create ONLY ONE component with the three of them, wouldn't you?

2º Both in the Component Diagram and in the Class Diagram they include the Faces Servlet... I think if we use JEE5 framework, we don't have to include it completely in our diagrams, but just OUR components. So I would include my jsp's and Backing Beansm but neves the Faces Servlet, would you?

Thank you once again.

antonio

Hello again,

how can I specify the scope of my Backing Beans? I thought about using stereotypes in my UML diagrams.

And what about specifying if an EJB y Stateless or Stateful? Do I use a stereotype, do I use the name of the EJB(*SLBean or SFBean)?

Thank you very much!

antonio

Hello again,

Do you know what conventions are there for specefic kinds of classes? For example, EJBs are called *Bean, and they're interfaces *Local or *Remote. So how should I call JSF Backing Beans? And what about a Web Service client?

I'm using this naming convention (let's suppose it's about Sales):
SaleLocal
SaleBean
SaleBB (for JSF Backing Bean)
SaleWSClient (for Web Service consumer)
SaleManager (a sales general purpose pojo)

Is it right or is there another nameing convention?

Thank you very much!

antonio

Hello,

I'm about to finish my part 2 deliverables and I would like to explain the reasons that made the design the way it is, I mean, why I used an EJB for one case and not for another, why I need to use paging when returning a long list, and so on. Can I include a text with this information in the deliverables or am I supposed to deliver ONLY the diagram images, the 3 top risks and the assumptions? I just feel it would be easier to understand my decisions if I could explain them.

By the way, should I include in my assumptions, technical matters or only my assumptions about the essay?

Thank you very much!

antonio

Hello!

My question is: for JEE5 (SCEA5), do we still need to use Service Locator as we used to (prior to EJB 3.0)? When accessing a Statelss EJB, why should we use a Service Locator if we can just inject it through annotations?

On the other hand, when it comes to accessing Stateful EJB (from a Stateless EJB or any non-stateful object) we can't just inject it since we would keep injected the same Stateful Bean for the other users, so what to do there? Do we use a Service Locator there?

Thank you very much. Bye

antonio

Yes that's always one of my first references, but I didn't find what I want, I think it's not possible.

I can do something like the following, just in case anybody wants to know it. Maven installation brings a jar (uber) which defines the lifecycles in the file uber.jar/META-INF/plexus/components.xml. There you should define a knew lifecycle with its own id and own phases. Then you'll have to define default goals (plugin:goal) for each phase.

You can define goals already defined for other lifecycles, but you cannot repeat phase names. The problem is that even when I declare here that I will use MyPlugin:MyGoal in the phase MyPhase of a new custom lifecycle, I'll have to modify dozens of projects's pom.xml files if I want them to have the behavior they used to have for common phases in my phases. I mean, I can add this to components.xml

<id>my_lifecycle</id>
<phases>
<phase>my_deploy</phase>
</phases>
</lifecycle>

But not this

<id>my_lifecycle</id>
<phases>
<phase>deploy</phase>
</phases>
</lifecycle>

Because I'm using a phase name already used for default lifecycle: deploy.

My scenario is there are hundreds of projects that include a pom with lot of info about compilation, testing... AND DEPLOYMENT. But as it takes so long any of these phases, I want to create a custom lifecycle that just go directly to deploy phase (the same as default lifecycle but without phases prior to pdeployment), that way I can invoke the default lifecycle until install phase and leave it all prepared sometime before deployment, in order to be fast when it comes to actually deploy the app because it wouldn't go through previous phases.

Whith my solution I would have to mention the phase my_deploy in hundreds of pom.xml files as all of them have a lot of info related to deploy phase now... I think my best choice would be to invoke mvn deploy skipping all the previous phases.

I hope anybody found this topic worth reading. Best regards and thank you very much as always.

antonio

Hi,

I need to have a new lifecycle for Maven 2, in which I define my own phases. Does anybody know how to do this?

Do I need to introduce changes in the poms? Because I would like not to touch anything in my already working poms, just create a new custom lifecycle, with new phase for which I can create plugins, but avoiding changes in every pom IF POSSIBLE.

Thank you very much!

antonio

Just in case anybody else is facing the same problem:

CR786ML: WAS Network Deployment V7.0 for Linux on x86-32 bit, Multilingual eAssembly

CR787ML: WAS Network Deployment V7.0 for Linux on x86-64 bit, Multilingual eAssembly

CR77WML: WAS Network Deployment V7.0 for Windows on x86-32 bit, Multilingual eAssembly

CR77XML: WAS Network Deployment V7.0 for Windows on x86-64-bit, Multilingual eAssembly

Bye!

antonio

Hello all!

I'm trying to download WAS 7 from IBM web page. I have a partner user so I can download it, but I'm diving into the web and I can't find it. Can anybody tell me its part number?

Well, in fact this web changed its structure in the last days, so if anybody can tell me the navigation to get to the form where I can introduce the part number it would be a great help!

Thank you very much!

antonio

Thank you very much for your reply.

And about point 1, is not my personal opinion and I really don't think so, neither think otherwise. I just don't know. It's just that anytime I ask about this topic people tell me "WAS is far better than WASCE" but they give me no reason. So to avoid this kind of answer I meant that I GUESS WAS is better than WASCE but what I wanted to know were the reasons why one is better than the other and the inner differences. So thank you for your replies: I was mistaken for so long thinking that WAS was just a WASCE with a few improvements: in fact, as I don't usually like IBM products I supossed they would be improvements of little value. (That's what I think of IBM IDEs: they're Eclipses with plugins of little value in most of the cases : that's my personal opinion)

Good to know. Bye

antonio

Hello.

I don't understand something about all of this. I know that WAS is far better than WASCE, but is there a WASCE inside a WAS? In other words, is a WAS some kind of WASCE with a lot of plugins, modules and so on or they don't share anything at all?

And by the way, if I develop an app for WAS and I test it on WASCE and it all works fine, should I trust it will work on WAS?

One more question, about the wsadmin tool for WAS, I know there's a similar one for WASCE. Do they share commands, syntaxis and so on?

Thank you very much for your replies! I can say this topic will be very interesting for many people around me, so we're looking forward to read your replies.

antonio