Where to put logic with Transfer Objects?

I'm playing with Transfer Objects for the first time, and have got this one class I'm not sure how to design the Transfer Object for. The problem is that I have some logic that is needed by both the domain model and by the client.
The class is named Moment. It's basically a wrapper around the java.util.Calendar class because I don't like passing objects of that type around in my code. That Calendar class is just to complicated to expose everywhere.
So, an easy example is that internally, Moment stores time in millitary time. I need a method that will take an hour, a minute, and AM or PM and convert it to millitary time for storing internally in the Moment.
Before Transfer Objects, this was easy enough of a decision. Moment had several factory methods called makeMoment. One of these took three arguments: hour, minute, and AM or PM. And, it internally converted to millitary time and stored the time in the Moment. However, now that the client is operating on Transfer Objects and not Domain Objects, if I put that method on Moment in the domain model, it won't be available to the client. If the client takes user input in AM/PM format from the user, it would have to duplicate logic already implemented in the model.
Options I think I have:
1.) Put the conversion method in MomentTO, maybe make it static. Then, my Moment class calls this method on MomentTO to do the coversion. This would work, but it seems funny to have the domain model calling business logic inside the Transfer objects.
A variation of this solution is just whenever my domain model wants to create a Moment, have it first create a MomentTO via a factory method that could do the coversion, and pass the MomentTO to a Moment constructor. This seems even funnier though to have the domain model creating Transfer Objects for internal use. It just seems to me that the point of Transfer Objects is to export information about the domain model to the client, leaving a lot of the specifics behind. Not something that the domain model uses for processing internally.
I could put the conversion method on the MomentTO and just say anybody who wants to create a Moment using anything but millitary time is going to come from the client that will be using Transfer Objects. Nobody else will be creating moments using anything but straight millitary time. But, this approach seems inflexible.
2.) Make a separate library / package that is available to both the domain model and the client. The inital reason I don't like this solution include the fact that I don't need much more other logic to be shared like this, so this package would be really, really small.
Any suggestions? Just can't seem to find something that makes good sense. I do have several other things like this for this Moment class in creating a Transfer Object for it. But, it is just this one class that I'm having this problem with.

>>
If this is for the pursuit of knowledge, then thatis
all well and good. However, if you actually wantto
do something with this application, I wouldrecommend
the following link:
http://xp.c2.com/YouArentGonnaNeedIt.html
- SaishI think you missed the part in the thread where I'm
not seeing the evil of Transfer Objects. Please
explain. They really don't seem that much work to
develop to me.
Fair enough. However, my point is that any time you develop something that 'might' be needed is time you are taking away from time developing things that 'are' needed.
One of the listed advantages of the Transfer Object
pattern in the Core J2EE Patterns is hiding
complexities of the domain model from the client.
This seems to be very handy for my application.The session facade pattern would accomplish the same effect.
. Many of the domain model objects are versioned,
but the first three iterations of development that I
want to complete before deployment need to know
nothing about the versions. With transfer objects,
this can all be safely hidden in the Application
Service objects.
I fail to see how transfer objects themselves would not require versioning or modification. Given a large enough transfer in the domain model, it woud be difficult to insulate an associated transfer object. Granted, these changes may occur less frequently.
Plus, I'm trying to figure out how the hell Commons
Validator works with Spring. Really wish I could
find a getting started guide. What I've found so far
makes it look like it's going to be easier to
validate Transfer objects than my domain model
objects, because all the validated fields will be on
one object. Not mostly on one object with some
fields scattered about like they would be with domain
model objects...I've found good FAQ and HOW-TO documentation at Jakarta while working on several Commons projects, including Log4J and FileUpload. Validator has no such online documentation?
For Spring, I recommend the new Johnson "J2EE Development with Spring" and (I forget the author) "Spring in Action".
- Saish

Similar Messages

Where to put Logic Plug-ins?

When using a product like Miroslav Philharmonik, I can get the program to run, but I can not figure out how to make it work with logic. The plug-ins do not show up! Someone please help me know what folders to put what in. Thanks

luke79 wrote:Why is it, that when I run philharmonik in logic, I have to load the intruments into strange channels, I can't just use channel one, I have to switch it around. Does that make sense?
No, not really... what channel(s) do you mean? I have never seen "strange" channels in Logic (or Philharmonik). A bit more detail, or even screenshots, might help.

I am trying to organize a zip file with a bunch of links and files. But when I drag an image file to a folder, it always "Snaps to grid" instead of just landing where I put it.

I am trying to organize a zip file with a bunch of links and files. But when I drag an image file to a folder, it always "Snaps to grid" instead of just landing where I put it. All of the files are going into the same zip file but I want them visually in the space that I put them and not at the bottom of the snap to grid folder location they are looking like. I am trying to drag and drop files in groups so that I can group them for now, but image files always do their own "sort by" when I drag them from my web pages. Links don't do it though. I tried resetting the folder, and deleting the DS files. No luck.

You keep asking variants on this same question. You've had replies in all your other threads. If you can't find them, go here and click where it says Activity:
Thomas Cannon Jr.

I bought a new mac and my music in itunes didn't transfer over from my old cpu. Most of it was put on with cd's I have. Can I transfer the music from my I-phone to itunes on my mac?

I bought a new mac and my music in itunes didn't transfer over from my old cpu. Most of it was put on with cd's I have. Can I transfer the music from my I-phone to itunes on my mac?

It sounds like there's something amiss with either your ID or the new system you're using. I've been using iTM on a Win7 and OSX SL system and have been able to upload and download to the Cloud on both. I've been using them interchangeably for the last week or so.
I would also recommennd giving Apple support a call, to see what they say.

My iphone is locked how i can know with which id i can unlock it .tell me any web site where i put some information and that tell me with which id i can unlock it?

my iphone is locked how i can know with which id i can unlock it .tell me any web site where i put some information and that tell me with which id i can unlock it?

http://support.apple.com/kb/PH13695

Where to put .txt file with SimpleInput

Hello everyone,
I'm rather new to Java and have a slight problem. The thing is that I need to write an app that can read a text file in this format:
2
Orc
name = Gnorck
experience = 20
maxHealthPoints = 50
healthPoints = 30
maxBerserk = 9
berserk = 7
Human
name = Aragorn
experience = 45
maxHealthPoints = 40
healthPoints = 30
maxGreed = 20
greed = 10
The first number is de number of characters in the file and after that every character is described in 7 lines.
Now the application has to be able to read the file. I've choosen to first just make a list of characters. A friend helped me with this code but I'm not completely sure if I know how it works. The code:
package RPGApplicatie;
import java.util.*;
public class Karakters {
     //Constructor
     public Karakters(int size){
          karakters = new ArrayList(size);
     //Methode voor het toevoegen van een karakter
     public void voegToe(Karakter k){
          karakters.add(k);
     public static Karakters readFile(String filename) throws ClassNotFoundException, RuntimeException, InstantiationException, IllegalAccessException{
          SimpleInput file = new SimpleInput(filename);
          int size = file.nextInt();
          Karakters result = new Karakters(size);
          for(int i = 0; i < size ; i++) {
               Class c = Class.forName(file.nextLine());
               result.voegToe((Karakter)c.newInstance());
          return result;
     //Variabelen aanmaken
     private ArrayList<Karakter> karakters;
}First question: I have the classes Karakter (which is the superclass for class Human and Orc), Human and Orc. So does the method forName scan for strings that are also classes? Like so it scans the file and when it comes across a line that says either 'Orc' or 'Human' it adds it to the list?
Second question: Where to put the text file that holds the data? I now have it in the same folder called test.txt. I use this code to run it:
package RPGApplicatie;
public class RPGApplicatie {
     public static void main(String[] args) throws ClassNotFoundException, RuntimeException, InstantiationException, IllegalAccessException {
          Karakters.readFile("test.txt");
}Thanks in advance! If more info is needed I can post that.
Regards,
Sander

I assume you have defined your Human and Orc classes to include fields name, experience, etc and a constructor that takes these values as parameters?
You can remove the Class.forName() method altogether. Example:
int count = file.readInt();
// define variables for Karakter values (name, experience etc.)
for (int i = 0; i < count; i++)
String characterClass = file.readLine();
// read common character components
if (characterClass.equals("Human")
    // read human specific components
    karakters.add(new Human(...));
if (characterClass.equals("Orc")
    // create Orc
//...Cheers

How to combine Session Facade and Transfer object?

Hello All!
I'm working on an enterprise application. Presentation layer is a stand alone client, business logic is build on the Glassfish v2.1 and MySQL is used as a database. The client is connection to the GlassFishj server remotely using EJBs.
I have problems with business logic architecture.
Here is the brief description of backend application architecture design:
1. Session Facade pattern is used to simplify the client and application server interface and to provide application layers between backend (http://java.sun.com/blueprints/corej2eepatterns/Patterns/SessionFacade.html).
2.Transfer Object pattern to define update transfer objects strategy in order to decrease network overhead during client and application server interactions and to provide version control for objects. Transfer objects are designed as simple java business serializable objects. (http://java.sun.com/blueprints/corej2eepatterns/Patterns/TransferObject.html)
3. Originally the backend application consisted of three modules: users, storage and orders, but at the end I have decided to divide my application into the following parts - assortments, map, menu, orders, transactions, users.
4. All MySQL database transactions are via JDBC using procedures. No use of entity beans.
Questions:
1. I have some doubts about using Session Facade and Transfer object patterns at the same time. At first I'd mike to cite the definitions of the patters from the SUN official web site.
* Use a session bean as a facade to encapsulate the complexity of interactions between the business objects participating in a workflow. The Session Facade manages the business objects, and provides a uniform coarse-grained service access layer to clients.
* Use a Transfer Object to encapsulate the business data. A single method call is used to send and retrieve the Transfer Object. When the client requests the enterprise bean for the business data, the enterprise bean can construct the Transfer Object, populate it with its attribute values, and pass it by value to the client.
* So, if I use Transfer Object along with Session Facade, it makes some difficulties with object version control, because I 2 or
3 transfer objects controls with the 1 bean class. The best option for Transfer object Pattern is that each transfer object should have its own bean class to provide ability of object's version control. In the case it can bring the network overhead because of frequent remote calls caused by the large number of the bean classes.
* So, should I use the both patterns? If yes, how to manage the interaction the patterns. If no, which one to use.
2. E.g. I have a huge list of the Order objects and each Order object consists of other complicated objects. So, would I have trouble to transfer that list over network? If yes, how to manage it.
Thank you!
Astghik

Astghik wrote:
Hello All!
I'm working on an enterprise application. Presentation layer is a stand alone client, business logic is build on the Glassfish v2.1 and MySQL is used as a database. The client is connection to the GlassFishj server remotely using EJBs.
I have problems with business logic architecture.
Here is the brief description of backend application architecture design:
1. Session Facade pattern is used to simplify the client and application server interface and to provide application layers between backend (http://java.sun.com/blueprints/corej2eepatterns/Patterns/SessionFacade.html).
I would simply recommend establishing a service tier. Your services should be stateless. You can go the extra mile and have a session facade, but in the majority of cases, coding to an interface for your service accomplishes the same goals.
2.Transfer Object pattern to define update transfer objects strategy in order to decrease network overhead during client and application server interactions and to provide version control for objects. Transfer objects are designed as simple java business serializable objects. (http://java.sun.com/blueprints/corej2eepatterns/Patterns/TransferObject.html)
The idea of the transfer object is very similar to the Command pattern. I think if you investigate that pattern, it will be more obvious. The transfer object reduces network latency by consolidating all the parameters into an object, ideally, this also consolidates multiple method calls. If you combine a transfer object (or command object) with a service tier, you get the best of both worlds. The service can delegate calls to helper objects (or other services or components) using the data in the transfer / command object.
3. Originally the backend application consisted of three modules: users, storage and orders, but at the end I have decided to divide my application into the following parts - assortments, map, menu, orders, transactions, users.
The is your domain. It will vary from application to application. The principles above are more general (e.g., patterns and architectural tiers) and should apply to most domains. However, your actual use case may require something different.
4. All MySQL database transactions are via JDBC using procedures. No use of entity beans.
Consider using something like iBatis or Spring's JDBC templating to make your life easier with JDBC.
Questions:
1. I have some doubts about using Session Facade and Transfer object patterns at the same time. At first I'd mike to cite the definitions of the patters from the SUN official web site.
* Use a session bean as a facade to encapsulate the complexity of interactions between the business objects participating in a workflow. The Session Facade manages the business objects, and provides a uniform coarse-grained service access layer to clients.
* Use a Transfer Object to encapsulate the business data. A single method call is used to send and retrieve the Transfer Object. When the client requests the enterprise bean for the business data, the enterprise bean can construct the Transfer Object, populate it with its attribute values, and pass it by value to the client.
* So, if I use Transfer Object along with Session Facade, it makes some difficulties with object version control, because I 2 or
3 transfer objects controls with the 1 bean class. The best option for Transfer object Pattern is that each transfer object should have its own bean class to provide ability of object's version control. In the case it can bring the network overhead because of frequent remote calls caused by the large number of the bean classes.
* So, should I use the both patterns? If yes, how to manage the interaction the patterns. If no, which one to use.
Versioning is a separate issue. Generally, the more coarsely grained your transfer / command object is, the more changes are likely to impact dependent objects.
Your command or transfer object does not have to be a vanilla JavaBean, where you are basically creating a bean that has data from other objects. You can simply use your command / transfer object to encapsulate already existing domain objects. I see no need to map to a JavaBean with what you have described.
Generally, a method signature should be understandable. This means that many times it is better to pass the method, say, two coarsely grained objects than a signature with a dozen primitives. There are no hard and fast rules here. If you find a method signature getting large, consider a transfer / command object. If you want one service to delegate calls to a number of other services, you can also create a transfer / command object to furnish the controlling service with the data it needs to invoke the dependent services.
2. E.g. I have a huge list of the Order objects and each Order object consists of other complicated objects. So, would I have trouble to transfer that list over network? If yes, how to manage it.
This is a large, open-ended question. If you are going to display it to a user on a screen, I do not see how you avoid a network transfer with the data. The general answer is to not pass the data itself but rather a token (such as a primary key, or a primary key and a start and stop range such as you see on a Google search result). You do want to limit the data over the network, but this comes at a cost. Usually, the database will receive additional load. Once that becomes unacceptable, you might start putting things into session. Then you worry about memory concerns, etc. There is no silver bullet to the problem. It depends on what issues you are trying to address and what trade-offs are acceptable in your environment.
Thank you!
AstghikBest of luck.
- Saish

Transfer Object Pattern and Concurrency Transaction Mgmt

I am developing an application that implements a remote rich client. For
performance reasons I have chosen to use the Transfer Object pattern. I
have been very successful with this from a performance standpoint, it
really paid off. I am using it in combination with an assembler pattern
to construct Transfer Objects for the view and to also reassemble domain
objects based on changes made to transfer objects on my client side.
Anyways the only problem I am having with it is I can't seem to figure
out how I should implement optimistic locking with it. Is there a best
practices to handle transaction control here?
I generally try to keep visibility of stale data down, but there are still
situations where concurrent clients could have requested to edit the same
the Transfer Object at the same time. I would ideally like to handle this
using a flavor of optimistic locking. I in fact have implemented
optimistic locking on the domain side, but now I am not sure how or if I
should integrate this into the Transfer Object or View side. The version
field used in optimistic locking is generally a hidden field. The only
way I can see to handle this with the view side is to expose this field on
the Domain Objects and actually store it into assembeled transfer objects.
This seems like it may be a bad idea from a design standpoint.
The problem is the client requests data and it assembled and delivered to the client. Then at some later time the client may send a request to the server to update this data. At this time a new transaction must be started where the server reloads the domain data and copies over the requested changes. Since the domain data was just loaded you wouldn't ever trigger a versioning problem unless the version was maintained on the transfer object and copied over as well. I am developing this project with hibernate on the back end and unfortunately hibernate doesn't acknowledge manual changes to the version field.
I just feel like there must be a better way to do this. I believe that using the Transfer Object or DTO pattern for remote client/server architectures is pretty common. So there must be a best practice to deal with concurrency? Suggestions, insight, lay it on me please.
thx

I personal respect both concepts and am using both in an application I am currently work.
Firstly, I have my Transfer Object which I call xxxData.
Next, the entities are called xxx
I have my DAO classes that handles all CRUD operation. but within the DAO class I have two methods
private <EntityClass> retrieveEntityFromObject(<EntityClassData> data) {}
private <EntityClassData> retrieveEntityFromObject(<EntityClass> entity) {}so with these methods, I separate my business logic from my data layer. My codes will alway use data objects instead of entities. For example, a create method will be
public ProductData createProduct(ProductData data)
entity.persist(retrieveEntityFromObject(data));
}I hope someone understands
Regards

New LabHSM Toolkit - Agile development of complex event-driven maintainable LabVIEW applications with active objects / actors based on a universal Hierarchical State Machine / statechart template.

Dear Fellow LabVIEW programmers:
Most of the systems you deal with are reactive. It means that their
primary function is constant interaction with their environment by
sending and receiving events. But most likely, they can have something
happening inside them too, even when they are not processing messages
received from outside. So, such systems have to continuosly react to
external and internal stimuli. Right? Moreover, most likely, they
consist of subsystems that are reactive too and, in turn, can have
their own "life", to an extent independent from other parts (with
which they still communicate, of course). Reactive (event-driven)
systems are more naturally modeled with active objects. So, why then
should we try to model and code them with GOOP and its passive
("dead"!) objects?
"Flat" State Machines have been known for decades to have severe
limitations. It's been more than 20 years since Dr. Harel invented
Hierarchical State Machines (statecharts) to fight those limitations.
Then why does NI still tout the same old good Moore FSM as the
ultimate tool for event-driven programming in LabVIEW in its $995
State Diagram KIt?
The LabHSM toolkit we are happy to present, makes it possible to
easily create and then maintain complex event-driven applications in
LabVIEW as a collection of HSM-driven active object VIs using a higher
level of abstraction and agile software development methodologies.
These active object VIs are created based on a universal Hierarchical
State Machine ( HSM or statechart ) template. So. all your code looks
similar regardless of its functionality!
We all love just jump to code, right? However, to be good boys, we
need to do design first. Then implement it in code. If the logic is
modified we need to redo the design first and then redo the code. When
using LabHSM where behavior information is abstracted into a separate
HSM data file editable with a supplied editor, there is no need for
coding separate from design any more. The modified behavior becomes
code automatically as soon as the HSM file is saved. Design is code!
The implementation basically follows Dr. Samek's Quantum Programming
paradigm. (see http://www.quantum-leaps.com). However, as already
mentioned, LabHSM stores the behavior information in a file separate
from the code itself. It also adds state dependent priorities to
events, a separate queue for public events/messages, and, of course,
some LabVIEW specific code like capturing front panel user events and
putting them into the private Events queue. Communication and
instantiation functions are also rather specific for LabVIEW.
It is available for UNLIMITED PERIOD trial. Please visit
http://www.labhsm.com for details and download. The site also contains
references which you may want to check to learn more about
hierarchical state machines and active object computing.
Since this is our debut we will appreciate any comments and
suggestions. Our contact information is available on our site, of
course.
Have a G'day!

Symtx is currently hiring the following position. Please contact me if interested.
Amy Cable
Symtx, HR
[email protected]
Symtx, the leading supplier of functional test equipment, hires the brightest & most talented engineering professionals to design & manufacture complex custom electronic systems for advanced technology leaders in the defense, aerospace, communications, medical, transportation & semiconductor industries. Symtx’ challenging & dynamic work environment seeks to fill openings with highly qualified electronic engineering design professionals.The ideal candidate will be responsible for defining the requirements, software design and code development, and integration of test control software for custom functional test systems. Candidate should be familiar with data acquisition concepts, instrument control, complex test, measurement and calibration algorithm development and definition and implementation of control interfaces to hardware. Prefer familiarity with instrument control via GPIB, VXI, MXI, RS-232 desirable. Requires BS/MSEE and 3 -7+ yrs of experience in one or several of the following test applications in a Windows NT/2000/XP environment using Labwindows CVI, TestStand, Labview, Visual Basic, C++ and knowledge of RF systems is a plus. Job responsibilities will include software design, development, integration, team leadership, and interfacing with customers( includes PDR’s & CDR’s).

Where to put EJB design patterns?

Im considering where to put the EJB design pattern classes and files.
i.e. Facade, Data Access Objects, Transfer Objects, Delegate...
I was thinking since the EJB business logics and implementation should be hidden from the user, then it is more logical to place these classes at the webapp.war inside of the ejb.jar.
What are your thoughts guys?

Photoshop patterns and normally stored in sets the set files have an extension of .pat. You may also add individual patterns using menu Edit>Define Pattern. In some Photoshop Pattern dialog you will see a little gear icon to open a Fly-Out menu in that menu there is a entry for Photoshop Preset Manager. The are also other ways to get into the Preset manager. In the preset manager you can manage many types of presets. Using its pull-down menu select patterns. You can Load an save patterns sets. To save a set tou heed to hilifht the paterns yoy want to include in the set. You can also delete and rename patterns.

Where to put java code - Best Practice

Hello. I am working with the Jdeveloper 11.2.2. I am trying to figure out the best practice for where to put code. After reviewing http://docs.oracle.com/cd/E26098_01/web.1112/e16182.pdf it seemed like the application module was the preferred spot (although many of the examples in the pdf are in main methods). After coding a while though, I noticed that there were quite a few libraries imported, and wondered whether this would impact performance.
I reviewed postings on the forum, especially Re: Access service method (client interface) programmatically . This link mentions accessing code from a backing bean -- and the gist of the recommendations seems to be to use the data control to drag it to the JSF, or use the bindings to access code.
My interest lies in where to put java code in the first place; In the View Object, Entity Object, and Am object, backing bean.....other?
I can outline several best guesses about where to put code and the pros and cons:
1. In the application module
Pros: Centralized location for code makes development and support more simple as there are not multiple access points. Much like a data control centralizes services, the application module can act as a conduit for different pieces of code you have in objects in your model.
Cons: Everything in one place means the application module becomes bloated. I am not sure how memory works in java -- if the app module has tons of different libraries are they all called when even a simple query re-execute method is called? Memory hog?
2. Write code in the objects it affects. If you are writing code that accesses a view object, write it in a view object. Then make it visible to the client.
pros: The code is accessed via fewer conduits (for example, I would expect that if you call the application module from a JSF backing bean, then the application module calls the view object, you have three different pieces of code --
conts: The code gets spread out, harder to locate etc.
I would greatly appreciate your thoughts on the matter.
Regards,
Stuart
Edited by: Stuart Fleming on May 20, 2012 5:25 AM
Edited by: Stuart Fleming on May 20, 2012 5:27 AM

First point here is when you say "where to put the java code" and you're referring to ADF BC, the point is you put "business logic java code" in the ADF Business Components. It's fine of course to have Java code in the ViewController layer that deals with the UI layer. Just don't put business logic in the UI layer, and don't put UI logic in the model layer. In your 2 examples you seem to be considering the ADF BC layer only, so I'll assume you mean business logic java code only.
Meanwhile I'm not keen on the term best practice as people follow best practices without thinking, typically best practices come with conditions and people forget to apply them. Luckily you're not doing that here as you've thought through the pros and cons of each (nice work).
Anyway, back on topic and off my soap box, as for where to put your code, my thoughts:
1) If you only have 1 or 2 methods put it in the AppModuleImpl
2) If you have hundreds of methods, or there's a chance #1 above will morph into #2, split the code up between the AppModuleImpl, ViewImpl and ViewRowImpls. Why? Because your AM will become overloaded with hundreds of methods making it unreadable. Instead put the code where it should logically go. Methods that work on a specific VO row go into the associated ViewRowImpl, methods that work across rows in a VO go into the ViewImpl, and methods that work across VOs in the associated AppModuleImpl.
To be honest which you ever option you choose, one thing I do recommend as a best practice is be consistent and document the standard so your other programmers know.
Btw there isn't an issue about loading lots of libraries/imports into a class, it has no runtime cost. However if your methods require lots of class variables, then yes this will have a memory cost.
On a side note if you're interested in more ideas around how to build ADF apps correctly think about joining the "ADF EMG", a free online forum which discusses ADF architecture, best practices (cough), deployment architectures and more.
Regards,
CM.

Where to put Validation Code?

Up until now, Im still having second-thoughts of where to put validation code when setting attributes of an entity.
Right now I have created lots of custom validators --(implement JbovalidatorInterface) that calls stored procedures to validate the value entered. But sometimes, i just use a ViewObject and query it on the setterAttribute method of the Entity and just manually throw a JboException of the value is invalid based on some business rule.
Question is, what are the best practices where to put validation codes? do we have to be strict that we always put all validations on Validators or are we free to just throw JboExceptions anywhere on the BC classes' code.
regards,
Anton

1. The reason I have a custom validator and I don't normally use the built in declarative validators is that the error message generated when the validation fails is fixed, only one message. I decided to have create a custom validator is that I need to test a one attribute for many cases in each case should produce a distinct error message. So if I use the built in validators, I would have to create lots of built in validators for that single attribute only. (and i have lots of entities and lots of attributes that needs business rule validation). So, I decided to create a custom validator, that calls the stored procedure, the stored procedure takes care of all test cases, for that attribute only, and I can return a dynamic error message depending on the test case that failed. What do you think about the approach?
It's a little extra work to create a reusable validator class that will only be used once, but whether you do it that way or encapsulate the call in a helper class that your one-off method validator code delegates too, it seems similar to me. So it's more of a stylistic choice for you which you like better. Now, if your reusable validator were enable to encapsulate
2. When I said anywhere; I meant inside the setterAttribute methods on the Entity and on the ViewRowImpl, orThe ViewImpl class or inside a method on an ApplicationModule?
Rather than writing code in the setAttribute, I recommend using attribute-level method validators. This makes it more clear where all your validation code lives.
I don't recommend performing validation in the view object level since entity objects are the component in the system that are tasked with handling validation. It would be easy to circumvent view level validation checks unless you make a lot of assumptions about exactly how your application is working.
3. One other issue is that Validator methods are for validation purposes only. So its not a good idea to put in attribute setters to other attributes inside there. So you put the attribute setter logic outside of the validator usually inside the setAttribute() just after validator returns. But there are cases that is very straightfoward to put validation logic inside the setAttribute; meaning, inside the setAttribute() method, I test for a condition, if it fails, just throw a JboException, if its true, continue with the otherAttributes setter logic.
Whether attribute setting of other attributes is performed in a setter method or in an attribute-level method validator, either way you will need conditional logic to avoid going into a validation "loop" (which eventually will throw an exception if after 10 attempts the object is still not valid at posting time.

Transfer object, FindByPrimaryKey, session facade

I would like to request help on coding a 'facade session' bean that invokes the FindByPrimaryKey method on a BMP entity bean, and how to put the returned values into a transfer object. I've been trying to modify the code provided in Sun's 'Duke's Bank Application' to accomplish this. Below is the code copied and pasted from Sun's session facade. 'accountIds' is a Collection that contains the values (customerId, balance, etc) of bank accounts which were retrieved with a FindBy method. AccountDetails produces a transfer object.
try {
Iterator i = accountIds.iterator();
while (i.hasNext()) {
Account account = (Account)i.next();
AccountDetails accountDetails = account.getDetails();
accountList.add(accountDetails);
I was wondering how to accomplish the task of transfering values into a transfer object when invoking FindByPrimaryKey.
I've taken a few stabs at it. My FindByPrimaryKey method doesn't throw any exceptions (as far as I can tell):
try {
Account account = accountHome.findByPrimaryKey(id);
} catch (Exception ex) {
throw new AccountNotFoundException("darn it - " + ex.getMessage());
The error (which appears in the catalina log) occurs in the following section of code, which immediately follows the above:
ArrayList myAccountDetails = new ArrayList();
try {
AccountDetails accountDetails = account.getDetails();
myAccountDetails.add(accountDetails);
} catch (RemoteException ex) {
throw new EJBException("getDetails isn't working in AccountControllerBean: " + ex.getMessage());
And here is the business method from Account:
public AccountDetails getDetails() {
return new AccountDetails(id, surname, firstname, balance);
If any of the experts out there have spotted a mistake, I sure would be grateful for your help.
James

Thank you for your help. I think the ejbCreate method in the 'session facade' takes into account your advice:
public void ejbCreate() {
try {
accountHome = EJBGetter.getAccountHome();
} catch (Exception ex) {
throw new EJBException("ejbCreate: " +
ex.getMessage());
account = null;
accountId = null;
I included your piece of code just the same to see if it would help anyway. Unfortunately the same error appeared in the catalina file.
Also, there were error messages I hadn't noticed until now in the Verifier.
One read:
"No findByPrimaryKey method was found in home interface class [database.AccountHome]"
The other:
"No single argument findByPrimaryKey was found in home interface class [database.AccountHome]"
Here is the method in the implementation class of the entity bean:
public String ejbFindByPrimaryKey(String primaryKey)
throws FinderException {
boolean result;
try {
result = selectById(primaryKey);
} catch (Exception ex) {
throw new EJBException("ejbFindByPrimaryKey: " +
ex.getMessage());
if (result) {
return primaryKey;
else {
throw new ObjectNotFoundException
("Row for id " + id + " not found.");
Below is the 'selectById' method that it invokes:
private boolean selectById(String primaryKey)
throws SQLException {
makeConnection();
String selectStatement =
"select id " +
"from accountTable where id = ? ";
PreparedStatement prepStmt =
con.prepareStatement(selectStatement);
prepStmt.setString(1, primaryKey);
ResultSet rs = prepStmt.executeQuery();
boolean result = rs.next();
prepStmt.close();
releaseConnection();
return result;
Here is what I have in the home interface of the entity bean for this find method:
public Account findByPrimaryKey(String id)
throws FinderException, RemoteException;
In addition here is the entire method in the 'session facade':
public ArrayList getAccount(String id)
throws AccountNotFoundException, InvalidParameterException {
if (id == null)
throw new InvalidParameterException("null id");
try {
account = accountHome.findByPrimaryKey(id);
} catch (Exception ex) {
throw new AccountNotFoundException("dag it" + ex.getMessage());
ArrayList accountList = new ArrayList();
try {
AccountDetails accountDetails = account.getDetails();
accountList.add(accountDetails);
} catch (RemoteException ex) {
throw new EJBException("getDetails isn't working in Account ControllerBean: " + ex.getMessage());
return accountList;
Thanks again for helping me out.
James

Flex design pattern for BlazeDS transfer objects

Dear Community,
I have a question regarding the nature of objects transferd from BlazeDS backend to the Flex client.
According to some adobe docs all the presentation logic should be in flex and BlazeDS should provide only services.
The question is what kind of transfer objects should be used? with what level of abstraction?
To make it more clear let's talk about a simple forum application like adobe forums.
I have created a Forum service, which returns a Forum transfer object, which has a List of Threads, with each thread having a List of Messages.
Flex gets the Forum transfer object and does all the presentaiton work.
This is a very nice seperatin of concerns architecture BUT there are two major performance issues:
- when the forum becomes bigger the transfer object will be huge
- the flex will use huge client resources to create the presentation of the forum transfer object.
So my question is what is your strategy/ design pattern for such a problem.
An obvious answer would be create services that include the presentation logic and use transfor objects like ThreadListPage which will include only the first 10 threads.
Thanks in advance

I think you are going to get a few varied types of answers to a general question like this, but I will comment on a few things. First, you should only be transmitting the data that you need to use or display to the user at the time. If this forum was implemented in Flex (I wish) you wouldn't transfer the text for all the threads. You might not transfer the thread text at all (just the subjects), you'd make calls back to the server to fetch that data when and if you need it.
Also, I think you are overestimating the resources required for the Flex "presentation layer". If you think about it, these workstations we have here are pretty idle most of the time when you're at a website. Nowadays there's plenty of processing power just sitting there waiting, so as long as your design isn't flawed the CPU or memory are rarely going to be a bottleneck with Flex.
The way I design, BlazeDS (or GraniteDS etc) should mostly just be "serving data". There is some flexibility available, for example I don't see a problem with putting certain types of business rules (such as a limit on the number of results that can be recieved in a search) within server side code, and validation rules on the Flex side. So there's not too many "industry standards" in Flex yet beyond maybe what's done in the various standard architectures, at least not that I'm aware of.
You might find it beneficial to look into some of the various frameworks/architectures available like Cairngorm, Mate, Swiz, and so on. Learning how they work (looking at examples) would probably answer some of your questions too.

Transfer Object in JSF, Spring, and Hibernate

The application on which I am working deploys the JavaServer Faces, Spring and Hibernate frameworks.
We are going to use the "transfer object" to move data between J2EE tiers from the front end to the back end and vice versa. In addition, data "types" may need to be converted --
At the front end, the "type" of the data that are entered by users may not match that in the database table. For example, checkboxes may be marked or left blank by users. The "type" of the checkboxes is "boolean" in the JSF backing bean. However, we have either "Y" or "N" for the corresponding field in the database table.
The JSF backing bean holds data as well as controller functionality. Of course, the transfer object holds data only.
With regard to class properties, Is the transfer object an exact copy of the JSF backing bean? And at which tier the data are converted to match their correspoding "types" in the database tables?
We also have the "data objects" which are exact representation of database tables.
Thanks for your advices.

Small example: you can decide in the business layer
whether to convert it to boolean or Boolean (which
has a 3rd state: null).But it all goes back to your data model. If your data model supports null for that property, then of course boolean is a poor choice. If that case one could use Boolean. None of this has anything to do with re-usability, which was your objection.
I think the re-usability question goes back to the data model as well. Suppose your data model says this property is boolean valued. However if your database does not support booleans natively, then you need to map the value for storage in the database. But that is the data persistence layers problem.
OTOH, suppose your data model says that this property may be 'Y' or 'N'. Then the Java class represent the business object should not have a boolean property, we should make it match the data model.
However, I think that most cases where people are storing 'Y' or 'N' in a field in the database fall into the first case, where the data model is a boolean and they are working around limitations in the database. There is no reason in that case to perpetuate the same limitations in the Java code.

Where to put logic with Transfer Objects?

Similar Messages

Maybe you are looking for