The initial part of the project kick-off is spend on Project Configurations, which is kinda tedious to handle during the initial phases due to component versioning. Maven does come to rescue but misses on closed project component templates. This bring us to another exciting yet to mature project from SpringSource called Spring Roo. Spring Roo is RAD development tool, which makes life easy for J2EE developers to work with Spring based project configurations. It helps you build model/domain driven Spring projects and generates the code based on the model/domain specified. Its just not a code generetion tool, but does all the code plumbing to integrate various components like, security, jms, logging, mvc, testing etc. which means the developer concentrates more on the entities rather than component internals.

Spring Roo heavily relies on AspectJ and Maven for most of its behind the curtain scenes. My initial reaction when I got started with Spring Roo was “Holy Grails, it does make life easy.”, but after closing looking at the code generated, its meant for prototyping and not for production deployments (just like Adobe Flash Catalyst design-to-code conversion, ugly and fat). So it does what it promises, and then I had to roll up my blue sleeves to make the code production ready. When working with Spring Roo here are my realizations :

1. The usage scope of Spring Roo is limited, doesn’t makes sense in collaborative enterprise project development.

2. Good for simple data models, not meant for complex ones. Also the domain modeling needs be visually enabled as it is in case of SkyWay Builder.

3. Code back-tracking is a mess, developers just don’t write code in incremental fashion.

4. Hard to sync up code when modified with further Spring Roo generation, generated code overrides your customized code.

5. Been working with Flex/J2EE projects for quite sometime, I am used to DAO design pattern, sadly this is missing in the code generated by Spring Roo, for a good reason, but still I am a spoiled one. Will take some time to accept the change.

6. Generates unit test cases and integration test cases, makes easy for QA and relies on Selenium for web application testing. (GOOD)

7. Generates the web tier for you to perform object CRUD operations, and relies on Tiles framework. (GOOD)

Further I must say Spring Roo does impress me from the point that it makes project configuration easy and the ability to bring in your own add-ons, no more getting paranoid with WEB-INF XMLs and dependency management. Considering the project timeline of Roo, I am sure the best is yet to come, for now limiting its application to prototyping itself.

NOTE: I have still got to put Spring Roo to further testing scenarios. Will write on my experience evolution soon.

With this we have got the Queue created and pretty much configured the ActiveMQ, now we would need to configure the Primavera P6 eventing system, so that it will post all the events triggered on to the message queue we created in ActiveMQ. For this login into your Primavera Administrator Console and navigate to your P6 Database Instance and configure the Eventing and Directory Services. Below is the attached screenshot with needed field-values that you will need to enter in order to post the events on to the ActiveMQ message queues.

Primavera P6 V7 Eventing Configuration

Once you have entered values as defined in the image for Eventing and Directory Services section, you can test the configuration by right clicking on Eventing title and click “Test Connection” form the menu. This will dispatch a test message to the P6WS7 queue which you can view it in the ActiveMQ web admin console. Hope this info helps you out, as none of the documentation that comes along with the Primavera P6 v7 platform has this defined, even not in the documentation thats on the support CD pack.

As in my earlier post I mentioned, Fiber is a suite of technologies which helps you do model driven development using Flash Builder 4 (with modeler plugin) and LiveCycle Data Services 3. And the major part of the Fiber is the Adobe Data Model, which lets you build data models, create services and bind services to data controls. Below are the slides attached which I used at the TAC 2009 and also find attached source code which was used for demos at the conference.

In order to work with attached codes, you will first need to setup and get started with Livecycle DS 3, about which you can find it here.

Demo II : Getting started with FB4 & LCDS 3
Demo III : Working with Annotations & Styles
Demo IV a : Working with Filters : Criteria based Filters
Demo IV b : Working with Filters : Pass-through Filters

Model Driven Development using LCDS 3 – Annotations & Styles from Mohammed Khan on Vimeo.

You can download the the source code which was used in the above screen cast here, and the sample MySQL database script here. When importing the project into your Flash Builder 4, you will need to change the LCDS 3 turnkey path in the Flex Compiler settings, also the RDS username and password would change accordingly. If you have any issue in getting started, feel free to share.

Getting started with Model Driven Development using LCDS 3 from Mohammed Khan on Vimeo.

UPDATE: The next part of this screen cast titled “Model Driven Development Using LiveCycle DS 3 Beta – Annotations and Styles” is available here.

1. Execution runtime support for Adobe Data Model

With the introduction of Fiber, a suite of technologies that enables Model-Driven-Devleopmet using Flash Builder 4 and LCDS 3, also comes Adobe Data Model. Adobe Data Model, is an XML-based persistence modeling language, and under the hood it implements the Java Persistence Query Language (JPQL) which is part of the Java Persistence API (JPA). In order to work with Adobe Data Model, you will need Flash Builder 4, Modeler Plugin, and LCDS 3 Beta. Without LCDS 3 Beta, you will not be able to execute or deploy your data model definitions, this also means Adobe Data Model wouldn’t work with BlazeDS.

2. Data throttling

For developers who worked with Flash Media Server, this should not be new. Throttling means the server increases or decreases the amount of data streamed based on how much the Flash Player can consume and process. LCDS 3 beta has various throttling mechanisms, and the one that automatically controls the server’s data outflow rate based on the client consumption rate is known as Adaptive Throttling.

3. Reliable Message Transactions

In LCDS 3, the Quality of Service (QoS), has been largely improved, for instance when working with data transactions applications like a Banking App, its necessary to have an acknowledgement made at all ends of the the data pipeline. This guarantees that whatever is send/receives is reliable.

4. Recover over Network Failures

Another feature that has been added to improvise the quality of service, is the ability for the Flash Player clients to recover data over network failures. A particular business case where this feature can be handy is a Trading application. Data Transmission is crucial in any data driven application, and also it becomes necessary to have high network redundancy. With multi-tier features it makes it possible for the Flash Player client to switch sessions with different LCDS 3 server definitions, if an network failure occurs with a particular destination.

5. Pause/Resume Capabilities

Another interesting feature, which can change the way the Flash Player client handles large chunk of binary data transmission is the Pause/Resume capability. Exchanging documents, large set of data can be controlled in more orderly fashion, the user gets the ability to control the exchange of data.

If you feel the above features excite you and want to test drive the MDD feature of LCDS 3, watch for this space. Till then I hope you have a great time.

1. Interface
2. Concrete Command
3. Receiver
4. Client
5. Invoker

Command Pattern Structure

In this post I will touch upon Interface element of the command pattern. The command interface defines the execute() method, which is responsible for executing the requested operation. Also multiple command types can be implemented using the same interface. The command interface provides a programmatic approach to manage the objects which have disparate modes of operation.

//General interface implementation
package com.thepixelcode.sampleCPExample.commands {

  public interface ICommand {
     function execute():void;
  }

}

The above snippet can be further extended to support the redo/undo functionality, and in order to implement, we would need to create two command interfaces, one for redo and another for undo which shall extend the ICommand.

//Interface for undo
package com.thepixelcode.sampleCPExample.commands {

  public interface IUndoCommand extends ICommand {
     function undo():void;
  }
}

//Interface for redo
package com.thepixelcode.sampleCPExample.commands {

  public interface IRedoCommand extends ICommand {
     function redo():void;
  }
}

I didn’t want to write a long post over the weekend, also neither i wanted to loose on my writing spirit. Shall continue on other elements and a demo app in next post. Till then have a great weekend.

Appcelerator TUG India

1. Firstly the Titanium platform is amazing, my first impression about it was, it can truly build Rich Unified Applications (RUA), though the hardware coverage is not that exhaustive but with insights from Appcelerator, seems like the next mobile platform they intend to cover can be BlackBerry. Also what is impressive is that within last 6 months, Appcelerator has extended Titanium support for 6 platforms (Windows, Mac, Linux, Web, iPhone and Android).

2. Secondly, as the web developer community in India is huge, and if they are provided with right support from Appcelerator, I am sure the Titanium platform can be put to use and build amazing social, enterprise and collaborative applications.

With that being said, feel free to join the Appcelerator Titanium User Group (TUG) India here. Learn and share your experiences with Titanium Platform, and if you require any further assistance, you can try to ping me (twitter or gmail) or reach out the Appcelerator team itself (over their IRC channel). We look forward to your participation and hope we build together an active community.

So lets connect and Code Strong.

Kitchen Sink App on Iphone using Titanium Mobile

Kitchen Sink App on Android using Titanium Mobile

For those of you who still don’t know what in the world Titanium is all about, then (no I am not gonna write about it here) visit their site here and if you are familiar with other RIA development tools like Adobe AIR/Flex, Curl and WPF, you will understand what Titanium can do with simple JavaScript. In short its a rich web, desktop and mobile development tool using JavaScript. What’s unique about it is that using same JavaScript code you can package the app for deploying it on Web, Desktop (Mac, Win, Linux) and Mobile (iPhone & Android). Isn’t that cool, and if you are a SI, I am sure it comes as blessing. No more writing code that is build differently for each platform. I am just wondering how would Enterprise computing would change with it (got to work on it) ?.

Also for ActionScript developers, it would be easy to adapt to Titanium platform as both ActionScript and JavaScript are based on the ECMAScript standards. In coming weeks I shall try to evaluate how Titanium can be fit for Enterprise solutions. Well for now, I think you should try out the platform yourself and think all about the possibilities of creating Rich Unified Applications. Go be a Titan !!