Marta De rossi

+ Follow
since Sep 22, 2006
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 Marta De rossi

sorry, I hit "Post new topic" instead of "Reply"
Sorry :roll:
Thanks for your reply.
Yes I agree, it's not necessary to persist unpaid itinerary.
Maybe I could offer seat-booking for the duration of the session.
I could save the seat booking as a flag.
Then when the customer buys the ticket OR when the session expires
the flag is removed.
No, wait a moment, it's not so easy. Because travel agents don't have session expiration......they have a Desktop application.

I think it's very complex to offer a seat "lease". So I'll state in the assumption that for the sake of keeping the architecture simple and consistent seats are only reserver after customer pays.
The drawback is that is customer waits too much in the Pay Use Case an expection could be raised because seats have all been booked.....but I think it's acceptable....what do you say ?

Before submitting the work I read again my work and I have a terrible doubt.

There's a sentence in the Prepare itinerary "Customer not logged" case which says that itineraries are saved before confirmation.

I will not say more because I don't want to discover too much of the assignment.

My question is: is it possible to argue that an Itinerary is saved WITHOUT actually assignment of the Seat ?

This has a big impact on the System because if seats are assigned ALSO for unpaid itineraries, it's compulsory to use Entity Beans because there are cuncurrency problems which can hardly be solved just with Session Beans.

What do you say to it ?
Dear all,
I have a doubt if inserting or not the Business Delegate in the Component Diagram. (Between the Servlet Controller and the EJB layer).

I can see on Mark's Cade book the Service Locator in inserted into Component's Diagrams.

I think that inserting it should give a clearer idea about how the two layers communicate......
what do you think ?
Hi all !
I have read a topic about Unpaid itineraries a few days ago.
I have decided to give a leased lock to the seat allowing the user
to reserve a seat within 30 minutes. After that time the seat lock will be released if it's unpaid.
My question is: with this design in mind, there can be multi-user access of Seats EJB PK (SeatIdPK) ?
If so, then I have to implement Seats EJB as Entity Bean.....

at the moment I'm only using Session Beans in my project.

Should I modify my design if favour of Entity Beans ?

Originally posted by Brian Smith:
Also, Is it scalable? I don't know - please how do you think it would be a scalable approach.

well in my opinion it should be a scalable solution because I don't need to create a SFSB for every client but I just use some SLSB to serve all the request.

On the other hand, as reported on this thread, I agree that passing session information across tiers could be a poor design choice.
I'll take a look at this forum is somebody passed using just SLSB.....
Yes I agree. Thinking over, I don't need a distributed transaction since I have a return code from the Web Service. I'll use that as discriminant for continuing or throwing Exception.
Hi all,
I'm modelling the last part of "Pay Itinerary".
Forgive the silly question: since CreditCardPayment and SeatReservation should be in the same transaction, it's not important which one is called first in the sequence diagram ? (So I can debit CC first and reserve Seat after- and viceversa)

My doubt arised because TransMaster communicates with the J2ee application using XML-RPC. I'm not sure if the transmaster Tx (XML-RPC) can join the Tx used for reserving a seat (written in Java), so if one tx fails also the other rollsback.

Hope somebody can shed some light on it....
Hi all !
I would like to join the discussion since I have taken a different approach.
I started with a SFSB facede for keeping Session information, then I decided to change it.

In fact I don't like very much to use both a Stateful SB and HTTPSession to store the handle(of SFSB). It seems more complicated and error-prone to keep up with both this session information.

Now I re-engineered it with just SLSB. I have inserted the ShoppingCart object in the HTTPSession for web clients and just kept it in memory for plain Java Clients.

I think it could make also the application more scalable.
I hope that my approach will not be a failure when I deliver my assignment

What do you say ?
Hi all,
we have a problem with the number of cursors opened in our application.
The application is deployed on Bea Weblogic 8.1
Obviously in some pieces of code (called in a loop) Statements aren't closed.
Unfortuntely the number of methods which open/close connection to the DB are countless.

Does anybody know if there is a way (maybe using MBeans ?), or a monitoring tool which detects all the pieces of code where Statements aren't closed ?

Thanks a lot
14 years ago
Hi all ranchers,
in the assignment (part II) it's not very much clear one point (really more then one !):
when the User receives a list of priced flights and alternative flights,
the flow continues with "Customer select flights" and
"System responds with seat availability for selected flights".

My question is: the Client can request seat availability for different flights at the same time ? or only for the departure and arrival segment ?

In other words: the Client can ask seat availability for: New York->Miami ,
Miami->Buffalo, New York->Washington...etc.

or only: New York->Miami , Miami->New York

Understanding this is crucial for my design, because my FlightVO holds both departureSegmentInfo and arrivalSegmentInfo. So If the first scenario is true I should pass a Collection of VO to the System. With the latter approach I simply pass the FlightVO, which has all the information needed.

Hope somebody can shed some light on this.
Hi all !
I've started the Scea part II examination.
I thought as the first thing to draw a component diagram.
I have some doubts on the assignment background. I have the following elements :

1. The FBN travel agents use a Cobol legacy system that accesses an IMS database

2. FBN uses TransMaster (via SSL) to process credit cards for flight purchases.

3. The IMS database will be replaced with a relational database

4. TransMaster Api won't be replaced at the moment and it's getting web enabled using XML-RPC over HTTPS.

So how does this all fit together ? this statements seems a bit contradditory.

Doubt 1:
Where does the relational database fits in this picture ? if IMS database is replaced with Oracle f.e. what happens to Cobol legacy system ? should it be replaced with EJB accessing Oracle ?

The Transmaster Api will receive input from the web, but what is its output ? it will connect the web to the legacy system ? or it will persist data on the J2EE tier ?

I will be very thankful if somebody gives me an help, maybe a minimalist description on how they drew the component diagram....