Alexander Petrov

Hello, I have a Usecase that includes another usecase. Basicly I have a case where I have to Pay a Bill but in order to Pay the Bill I need to first review the Bill. If I have understood correctly the Sequence diagrams are on per "use case" basis. I have couple of questions:

1.Is it acceptable if I have one Use Case for Review Bill and then my next Use Case - Pay Bill start where the first has ended.
2. I saw that there is this REF element. I don@t understand If I should stretch the REF element over all my life lines or I should only visualize it between the first the User lifeline and the second lifeline which is actually represented by the end lifeline of the included UseCase.

Is there any other mechanism to represent such a sequence diagram. And how this REF thing is defined when it comes to the overlapping with the lifelines ???

Thanks

Hei I am trying to create a sequence diagram. And would like to depict the communication between JSF and JAAS. There is very little online in this direction. I came across one of the posting on the javaranch forum, there are posted some sequence diagrams of JAAS, but never the relationship between JSF and JAAS is demonstrated. I read a bit online and check several stack traces.

If I follow the stack traces I have the feeling that the LoginContext is created as Part of the JSF lifecycle. Can we assume that both login on LoginCntext and create on LoginContext first pass through the FacesServlet and are happening as part of the JSF lifecycle?

The alternative is that this is happening somewhere in the Web container ? I dont see JAAS having own servlet registered in the web.xml so I kind of assume that this is happening possibly on container level.

I dont get it 100% how JAAS play together with JSF. I understand the whole JAAS flow of events, but dont get how it is executed together with the JSF part.

Yep the CDI beans are lets say a little bit more flexible Managed beans and the CDI provides better dependency injection and the use of interceptors and named factories. If I remember correctly CDI a lot of the features that the SEAM framework presented http://www.andygibson.net/blog/article/comparing-jsf-beans-cdi-beans-and-ejbs/

Thanks I was thinking the same that it should go under the presentation tier, but a reality check is always a good thing.

Alexander

Hei If I have a setup with CDI beans and EJBs that are holding the business logic. And the CDI beans are more or less handling the presentation state and also serving the data model. Are the CDI beans part of the Presentation Tier , or are they part of the Business Tier.

According to oracle I am quoting:Managed components that may provide the business logic of an application, but do not require the transactional or security features of enterprise beans are part of the Business Tier.

But on the other hand I am planning that most of my logic will go to the EJBs, and the CDI beans will work only as thin wrappers between the JSF pages and the Business logic and of course provide some state management for the presentation layer.

Any hints on that? If I follow the Oracle statement I should place them in the Bussiness Tier, but on the other hand if they are handling mostly the presentation state isn't it logical for them to be part of the Presentation tier ?

I think JPA comes first as it requires least dependencies from the JEE. I have never took the OCEWCD but if I think that it comes second place, or you take OCEEJB I would place both of them on second place. Does@t really matter which one you take first. I personally found OCEWSD most demanding, also had one of my worst certification exams fails ever on it. Probably if I had previously taken the OCEWCD this could have been avoided. But either way.

OCEJPA->OCEEJB->OCEWCD->OCEJSF or OCEWSD thats how i see it.

Alexander

Hei Tsang, thanks for the replay. The thing is that in the assignment it is explicitly specified that the backend protocol for communication with the payment system JMS. But if I assume that the payment system provides a form for entering the credit card details and so on.. I am basically going away from the protocol they have specified. It is still true that I can capture notifications for received payments or initiate the recurring payments from the backend. But I am a bit afraid that I am violating the requirements if I dont hold my own form for setting up the credit card details.

What do you think ?

Alexander

Hei, lets suppose that we have AJAXified JSF pages and we use facelets as well. If we have an application that is with separate pages we will put them in the CLASS diegram as XHTML pages. But if we have a situation where we have everything in one page, can we again put several XHTML classes in the class and component diagram After all these controls that are placed on the single page application are XHTML fragments. Basicly my question is in terms of UML how do you declare the XHTML stuff. It should be the same as multi page web application?

Hei, in my assignment there is a specific backend protocol defined for communication between the application and the payment platform. The thing is that when it comes to payments the first thing I can think about is something like PAYPAL where the customer is redirected directly to a form provided by Paypal itself.

But in order to conform to the requirements where the protocol is a backend protocol I need to create a form within my application where the customer will enter the credit/debit card information and then send it over this protocol to the payment platform.

I have a very strongt desire to make an assumption that the payment platform provides access to a web based form where the customer may feel the payment details. But I am afraid that this will go agains the requirements defind in the assignment. On the other hand I can still use this backend protocol to capture payment notifications or initiate monthly payments on abonament basis.
Do you think that such assumption about the Payment portal is a bit too much of an assumption.

Probably I should stick to the protocol and just implement myself the payment information form.

Looks like the direct debit is a very specific thing and I have already used it in Norway as "Avtalegiro". In Germany it is known as Giro. In US as Automated Clearing house. There is also a common service for all EU countries which is called SEPA Direct Debit.According to wikipedia and some other sources I read. There are two general ways to give consent that you agree the service provider to draw dircetly from your account. One method is to write a concent letter to the service provider which the service provider makes available to the Bank at later point starts collecting the payments. The other method requires the payer to instruct his or her bank to honour direct debit notes from the payee. The payee is then notified that he or she is now authorised to initiate direct debit transfers from the payer.

So when I read a bit I believe that the payment system should be able to send notifications to our system that a client has given allowance for our system to draw directly all the payments. Or alternatively we can hold a concent form with some digital signature or whatever to make sure that people/companies who use our service are real. Send the concent form to the bank and at later point get the payment. This is my view on the direct debit. According to the assignent the payment system is already present so I will just assume that the ability to receive notifications that a certain client has given access for direct debit.

Hei can someone clarify about the direct debit. Is this something that can be managed within the application, or is it something where you get redirected to external payment portal where you sign it ?

Am I correct that if the concept form is provided by my web site then I would need an electronic signature to sign it ? After all we need to be 100% sure that it is exactly the person/company that is legal signing the contract.

If I dont want to deal with electronic signatures and contracts and stuff is it possible to outsource it to some external portal ?

Thanks

I am sharing the same view. In general I try to minimize the amounts of DTOs in a project to only the places where they are absolutely needed.But this is a classroom assignment and what the reviewer thinks is a factor. And since I have been challenged on several occasions about my viewpoint of DTOs I will choose to go with the Mapper. Sounds to me like none will challenge this opposed to your(our) view which I completely share. I make it as a safeguard measure.

Hi, I am a bit confused about the level of detail I should place in the component diagram. I have created the following scenario. I have a timer service which is communicating with EJB that holds the business logic this EJB communicates with a External web service which returns a DTO that through Mapper is converted to Entity that later is stored by a repository class.

Is it too much if I put everything in the component diagram this would say the DTO to be part of the diagram plus the Entity that will be generated plus the repository class.

Is it better if I remove the mapper and the repository. And leave just the timer service the EJB with the business logic and the Webservice consumer. Is the repository and the mapper too much for the component diagram ? I think that by including them the whole algorithm is much more visible.

I am kind of not completely sure how much details is ok for the component diagram

Hei Tim, did you present this kind of model on the assignment or you are yet to finish the assignment ? Just wondering if you think it got accepted well. So far I have placed a Mapper in my assignment as a way to insurance myself if the reviewer see it in a negative way. Seems to me like the safest way to build it. I mean a reviewer may potentially find arguments agains co-existance of JaxB and JPA annotations but hardly will find any arguments agains the Mapper right ? What do you think ?

Yeah, it is. I am planning to use both JaxB and JPA , and wonder if I annotate the Pojoes with both of them or just keep two sets of porous plus a mapper. The thing is that while I believe it is probably better to have separation in real life very often I annotate the same pojo both with JPA and JaxB. Is the question inappropriate ? I think that it is not inappropriate question because it is very very general. Is it ok to combine JaxB with JPA annotations or not.

Hei, I have the case where my data is comming from external system web services. I am thinking store local the external data. I am thinking to use JaxB to parse the web services. I am thinking to use the same classes that are annotated with JaxB annotations and used as a model for the web services and annotate them with JPA annotations and use them to persist directly the data to the database. Theoretically I can have a second set of classes and mapper that is moving the data from the DTOs to the Entities,but I am questioning the needs since I think both JPA and JAxB are quite flexible and can co-exist.

What do you think ? Do you think this is a bad practice to mix JaxB and JPA(why I have the feeling that it is almost rhetorical question )?