Main Development Concepts
Openbravo ERP is a software application which is being developed with the following design principles in mind:
- Open Source
- ERP application framework
- Model Driven Development
- Rich Internet Application (RIA)
- Java - Lightweight J2EE
- Support for multiple databases
The aim of Openbravo is, while following these principles, to deliver an application which is state-of-the-art from both a technological as well as a functional point of view.
Openbravo is an open source project built on open source technologies. We aim to leverage on the excellent infrastructure components developed by the open source community to ensure our platform benefits from the advantages and stability of components supported by a large community. Whenever necessary we contribute our developments back to the community.
ERP application framework
Openbravo ERP is an application developed through an integrated development framework included in Openbravo ERP distribution. This integrated development framework takes care of a wide range of concerns in all the areas involved during the development process. Most relevant from low level to high level:
- Integration with Eclipse IDE
- Integration with SCM (Mercurial)
- Automated build process
- Automated update process
- Automated deploy process
- Built-in infrastructure for several common development needs:
- MVC framework (xmlEngine, httpBaseServlet, sqlc)
- Data_Access_Layer (based on Hibernate)
- Web server and servlet-container (integration with Apache-Tomcat and support to other J2EE implementations)
- Reporting (integration with JasperReports engine)
- XML and JSON REST Webservices
- Emailing (integration with Sun mail)
- Process scheduling (integration with Quartz)
- MDD development framework (Openbravo Application Dictionary)
- Multi-language user interface support
- Built-in security model
- Built-in enterprise model
- Multi-currency support
- Multi-general ledger support
Model Driven Development
Openbravo follows a Model Driven Development (MDD) approach. This means that Openbravo uses a technology agnostic model to define application components, such as windows and processes. Based on this application model, code and other software artifacts are generated.
Openbravo data model information -so called metadata- is stored in the Openbravo Application Dictionary.
Model Driven Development aims to increase productivity and re-use through separation of concerns and abstraction. The model is an abstract definition of system components which contains enough information to drive the generation of one (or more) implementations of the system in a concrete technology.
This separation of concerns -abstract functional description in the model and implementation of model components in a concrete technology- hides technology complexities to ERP domain experts in their process to define and implement new ERP functionality and simplifies the evolution of the implementation technology
In some cases it is needed to code a solution externally from the model. This is fully supported by Openbravo. Developers can freely develop their own solutions on top of the Openbravo technology stack.
Modularity is a capability introduced in the Openbravo ERP 2.50 release which allows to define and package additional functionality and configurations as extension modules, independently from the core product.
Modularity changes the way in which Openbravo can be adapted to user needs. Instead of customizing the code to match user requirements, it is possible to externally -from an independent module- extend functionality and to configure it.
This new approach has several advantages. Most important:
- Enables pure distributed development: new functionality can be developed through modules in a pure distributed manner. The team developing the module can work isolated from other teams -they only need a stable API from the other modules they use- and the life cycle of this module -including releases- is independent from other modules.
- Highly improves maintenance of code: developing through modules means packaging independently. With a proper definition of module dependencies and keeping API's stable the process of updating an instance is straightforward and can be performed in one user click
- Encourages sharing and re-use of new functionality: developing through modules make it quite simple to share this new functionality with other people. If developers want to share their modules all they need is to package and publish them in Openbravo Forge (Central Repository). After that these modules will be publicly available and other users can search for them and install them through a simple process.
Pure web application - Rich client
Openbravo is, by its very nature, a pure web application. Ubiquity of web browsers provides an universal point of access. Openbravo understands the network as a platform, delivering and allowing users to use applications entirely through a browser.
The requirements are minimum: a web browser is available on virtually all computer systems. Moreover, being web-based means the product can be delivered over the Internet, allowing to update the application without distributing and installing software on potentially hundreds of client computers.
Traditionally web applications had big limitations in regards of user interface. This has changed with the introduction of new web technologies such as AJAX. With AJAX and similar frameworks it is possible to develop a rich, interactive and user-friendly interface
Java - Lightweight J2EE
Openbravo uses Java as its backend programming language. There are many reasons for choosing Java as the server-side language:
- Open Source nature
- Wide support for enterprise-level development
- Mature architecture for web applications
Openbravo follows Java 2 Enterprise Edition architecture (J2EE) without making use of the EJB container. Instead of that Openbravo uses lightweight infrastructure to implement access to data and business logic. In Openbravo 2.50 Openbravo has delivered a new Data Access Layer (DAL) based on Hibernate that provides a powerful but still lightweight persistence mechanism.
Support for multiple databases
Openbravo is committed to avoid vendor locking in any technology it uses including database. Openbravo 3 runs on PostgreSQL (8.3.5+) and Oracle SE (10g-11g).
In future releases Openbravo aims to be database independent. The Data Access Layer (DAL) based on Hibernate is a first step in that transition.
Openbravo main concepts
Openbravo runs on top of a group of well-known third party applications:
- Apache-Tomcat. We use Apache Tomcat as the servlet container but others can be used instead
- Apache-Ant is used to automate a number of tasks such as build the system from source code
- PostgreSQL (8.3) or Oracle SE (10g-11g) database
All of these applications can be installed both on Linux or Windows.
Openbravo developers have three different ways to develop their code. Following the MDD approach, most common is to edit Openbravo Application dictionary through a web browser connected to Openbravo ERP. Based on the new model definition the software artifacts can be generated automatically. A developer can also connect directly to the Openbravo database through a sql client (eg. pgAdmin, sqlDeveloper) to manage database schema objects (tables, procedures, etc.). Finally developers can develop their own code through an integrated development environment such as Eclipse.
All openbravo software artifacts are stored in text files in the development project. This includes the database definition and content. The large advantage of using text files for storing all software artifacts is that it is much easier to share and compare changes made by developers in a distributed environment.
Openbravo uses a tool called DBSourceManager to manage database source code. DBSourceManager is able to read from the database schema objects and application dictionary data and export them to xml files. It can also create or update an Openbravo database from those xml files
The process to build the system from openbravo source code includes a number of steps to generate the code at different levels (DAL, WAD and others) and put together that code with other code directly written by developers. Openbravo has automated this process through an ant task.
The new architecture is explained in the Openbravo_3_Architecture article