I am designing a system in for Java EE based web application wherein there are lots of Master Data tables. There can be new master data added or a new field may be added in the master tables.
The master tables can fetch the data from database or from webservice.
I am thinking instead of creating logic for creating each master table separately, I create some sort of code generator which does the code generation for master data management.
Essentially what does all master data have.
A view to add the master data to table
A view to update the master data
A view to list all the master data
with user actions being
Search and Filter
Has anyone created something like this design for Java EE based application
We are creating a product. User doesnt need anything. We hat lots of master data which will dynamically change based on once we go to market. Master data like city, country, district. Basically admin screen has actions like add screen, update, listing, search.
Normal approach is to design table, create dao, have controller and service layer and then frontend screen. If a new table comes in again same thing needs to be created.
I was thinking for a way to automate creation of such code or create something which has common set of code to accomodate any master table introduced in future or update existing table.
What you do is based more on what you posted here.
I can say for sure that attempts to "make my work easier" which result in dynamic/meta data solutions are always wrong. They do nothing but make the final solution MUCH harder to maintain.
That said there are numerous solutions that already exist for creating multiple layers based on actual data models. Those work when the actual concern is not a dynamic solution but rather the work involved in creating code for many data entities which is basically the same.
Depending on the solution it can create any or all of the following
1. The DDL
2. The DML - stored procs that act as an database API
3. DTOs and DAO in your language of choice
4. DAO API layers.
Myself I have been rolling such solutions myself for decades.
Only suggestion to the above I would make is that you must not allow the ease of use of the DTOs to allow you to extend their use into other layers of the application UNLESS they are free from all database hierarchy abstractions. And perhaps even then.
We have designed a 3-tier web app for a finance application.
The business tier is divided further into layers like manager, helper, util layers to modularize code and isolate the different functions i.e. core business vs non-business code from each other
Util layer has non-business functions which are required during a particular process e.g. DateUtils.java, EncryptionUtil.java etc
Helper layer has business logic which is specific to particular business process and not required in other business processes e.g. SomeThirdPartyInterestCalculationHelper.java, SpecificRequestBuilder.java
Manager layer has business process which controls the flow as well as implements some parts of business logic e.g. CustomerAccountManager.java has different methods for CRUD operations for customer account. It calls different helpers, utils, DTOs etc and gets the work done. It also implements some pieces of business logic. So,it performs mix of BPM role as well as parts of core business process logic.
As the processes become complex and lengthier, my manager layer is growing and does not look like well organized code.
I want seperate layers doing specific roles i.e. business process controller, business process execution (core business logic), CRUD operations which are DB specific, helpers (specific to processes), non-business logic
What can be a better design pattern to achieve this?
I am trying out Business Objects patterns to isolate different parts of business logic and coupling it with Application Service pattern.
So, for executing a business process, I would have:
1. ApplicationService - Would be a pure business process controller calling different business objects and controlling execution based on results of BO methods
2. BusinessObject1 - Core business logic in different methods - Called by ApplicationService
3. BusinessObject2 - Core business logic in different methods - Called by ApplicationService (if BusinessObject1 grows bigger or BusinessObject 1 and 2 can perform specific business functions)
4. IntegrationBusinessObject - To call other third party services required in business process
5. DomainEntityBusinessObject - CRUD operation for a particular domain entity required in process...will also have some business level checks required before or after CRUD operations
6. Adaptors - To convert formats for third party services - May be called by IntegrationBusinessObject
Idea is to make classes more compact and doing specific business functions. Also, control the process from a single class (Application Service) so that changing the process can be easier.
I have 3 DNN webservers behind a Citrix Load Balancer, the load balancer is configured for SSL Offloading.
I discovered that the login link doesn't work anymore. It just refreshes whenever it is clicked. The URL of the login link is: https://test.abc.net/User-Login?returnurl=%2f. The link when clicked supposed to take users to the page where they will login.
When I changed d Citrix load balancer to HTTP, everything works normal. I.e http://test.abc.net/User-Login?returnurl=%2f takes the users to the Login Page.
Any suggestion on how to resolve this issue will be appreciated.
In an imagined situation we have a file without extension. We open the file with a hex editor. Is there any chance for us to interpret what we see in hex editor and ascertain whether the file has only instruction or only data (maybe as a text, maybe in some other format).