OOA is good but OOP is bad for enterprise applications

Object Oriented Analysis is good but Object oriented programming is bad for enterprise applications such as ERP. ERP are object based and not object oriented. The objects created in ERP don’t live long time. They live mostly for the current request of the user.

What is next to SaaS, for programmars?

If we trace the software developed to ease programmars in the chronological order

Operating System(OS) – These took care of memory allocation, file system and external device interaction

RPC,CORBA,RMI- These took care of  connecting processes across network. The programmar need not think about networking code.

RDBMS- An easy way to store and retrieve data.

Java( garbage collection)- Memory leaks were greatly reduced, thus the servers were able to run for longer time.

Application Servers – Multithreading, handling multiple request, clustering was handled automatically the application servers.

SaaS(Software as a Service)- Now scalability and availability are out of programmars mind.

So what is next?

SaaS solutions are too proprietary. Each vendor have their own programming model. It is high time to develop programming independant of  technology and architecture. Thus code once and Run anyway. The business logic has to be mentioned in XMLs. These XMLs can be converted to target architecture/platform.

This is exactly the goal of  sednaspace. Using sednaspace developers can design,develop and execute their business logic using XML. For more information please visit www.sednaspace.com

View Driven Development

I think model driven development is not best
approach for software development. First pages
have to be designed and build prototype before model
can be arrived.

Develop ERP using XML

Languages like C#, Java need not be used for creating ERP or web enabled database applications. We at sednaspace have designed an xml based language to define business logic. This way you do not lose your business logic.

IT needs ERP platform not ERP

SaaS is projected as the big saver of IT costs. The cost of intergration and customization is higly downplayed. Also it is projected that IT spends too much of money in maintaining their existing systems. The solution to this problem is an ERP platform and not an ERP. The business users understand their domain much better than an outsourced partner or ERP  vendor. These business users want their existing processes to be automated. Currently it is being outsourced or in-house development. But maintance cost is high due to non-standardization of software development.

IT department needs an ERP platform to develop their applications. This makes all their applications easy to maintain and develop.

Future of programming – for browser and database based applications

We have witnessed software development for the past thirty years. Languages such as Procedural languages(COBOL,C,SQL),Object Oriented languages(C++, Java, C#),Scripting languages(PHP, Ruby on Rails) have been used for software development. On top of these languages frameworks were built for ease of development. These languages are very similar and they have not greatly improved the development of software applications. The steps involved in dealing with the databases is same for the past twenty years. The steps involved are

• connect to the database,
• prepare the query ,
• bind parameters
• execute query
• close the connection

Visual Basic,SQL were huge sucess among developers. Both are interpreted languages and both did not expect programmers to deal with memory.

(LINQ reminds me of Pro*C days).All the important business logic are buried in these languages. Switching from one language to another is a very painful process. Also interoperability between languages were almost impossible. If we still continue to write our business logic in these languages, we are going to have a lot of legacy code in the future to maintain.

The future is about developing applications independant of technology and architecture. This is possible if we collect metadata about the application to be developed. The metadata is not about tables and columns in the database, but about everything including pages, business logic. If we can maintain metadata for all the parts of the application, then we can create applications independant of technology and architecture.

This the primary goal of our sofware www.sednaspace.com We collect metadata for all the layers(presentation, business and database layer). Using this metadata code can be downloaded in any language such C#,VB.Net, Java ,PHP.

Web Application generators

There are several tools available to generate code in 3GLs. Often these tools do not expose the metadata used for generating code. Most of the code generators are based on database tables. These tools generate code for CRUD operations. The applications developed especially in the web world demand more than this. The end applications needs to take care of security, internationalization, workflow and  BPM.

The code generators need to expand their scope beyond the CRUD generation. The code generator must generate complete web applications. This was the goal behind the tool SednaSpace. This tool generates code end to end for web applications. This tool also has an interpreter, to execute the application immediately.

Intranet SaaS

SaaS on the cloud has clear benefits(browser based development and deployment). SaaS on the cloud also has limitations in terms of customization and integration. Integration is done thru webservices only. Customization issues are similar to current ERP environment. Now the same environment needs to be replicated within the enterprise. But when it comes to within enterprise, the SaaS has to work on the existing technologies of the enterprise. For this Intranet SaaS has to be built on metadata. Metadata should not be collected only for tables and columns , but also for business logic. This is very important, because IT department has to migrate their applications from one technology to another , when a technology(or sometimes version).  becomes obselete