The development of open methodologies for application design, the three different development platforms, the importance of open standards to build web services, the current hype around agile development and extreme programming.

Also some ideas about comparison elements between open and proprietary platforms.

3. Open Internet Development

Most of the well designed Internet applications are developed under object-oriented methodologies , to better cope with the modular structures of HTML and XML, and to more effectively reuse or even share programming modules with other companies. The applications using object-oriented languages are often designed using a formal methodology and tightly related to modelling languages.

In this chapter, we will analyse the development of open methodologies for application design, which can be extremely helpful to ensure scalability, security and robust execution of Web applications under stressful conditions. After, we will see three different platforms used to develop the web applications, and discuss the importance of open standards to build web services.

We cannot finish without analysing the current hype around agile development and extreme programming. To conclude, we will review some ideas about comparison elements between open and proprietary platforms.

3.1. Design

Modelling is the designing of software applications before coding, being an essential part of large software projects, and helpful to medium and small projects. Using a model, those responsible for a software development project's success can assure themselves that business functionality is complete and correct, end-user needs are met, and program design supports requirements for scalability, robustness, security, extendibility, and other characteristics, before implementation in code renders changes difficult and expensive to make.

3.1.1. MDA

The OMG Model Drive Architecture (MDA) provides an open, vendor-neutral approach to the challenge of business and technology change. Based firmly upon other open standards, MDA aims to separate business or application logic from underlying platform technology. The platform-independent models enable intellectual property to move away from technology-specific code, helping to foster application interoperability.

· Specification

In September 2001, OMG members completed the series of votes that established the MDA as the base architecture for the organization's standards. Every MDA standard or application is based, normatively, on a Platform-Independent Model (PIM), which represents its business functionality and behaviour very precisely but does not include technical aspects. From the PIM, MDA-enabled development tools follow OMG-standardized mappings to produce one or more Platform-Specific Models (PSM): one for each target platform that the developer chooses.

The PSM contains the same information as an implementation, but in the form of a UML model instead of running code. In the next step, the tool generates the running code from the PSM, along with other necessary files. After giving the developer an opportunity to hand-tune the generated code, the tool produces a deployable final application.

MDA applications are composable: If PIMs are imported for modules, services, or other MDA applications into the development tool, it can generate calls using whatever interfaces and protocols are required, even if these run cross-platform. MDA applications are "future-proof": When new infrastructure technologies come on the market, OMG members will generate and standardize a mapping to it, and the vendors will upgrade his MDA-enabled tool to include it. Taking advantage of these developments, cross-platform invocations can be generated to the new platform, and even port the existing MDA applications to it, automatically using the existing PIMs. Although MDA can target every platform and will map to all with significant market buy-in, CORBA plays a key role as a target platform because of its programming language-, operating system-, and vendor-independence. The mapping from a PIM to CORBA has already been adopted as an OMG standard.

· The core

Figure 16 - Model Driven Architecture

§ The Unified Modeling Language (UML)
As discussed, each MDA specification will have, as its normative base, two levels of models: a Platform-Independent Model (PIM), and one or more Platform-Specific Models (PSM). These will be defined in UML, making OMG's standard modeling language the foundation of the MDA. UML is discussed separated in the next chapter.

§ The Meta-Object Facility (MOF)
By defining the common meta-model for all of OMG's modeling specifications, the MOF allows derived specifications to work together in a natural way. The MOF also defines a standard repository for meta-models and, therefore, models (since a meta-model is just a special case of a model).

§ Common Warehouse MetaModel (CWM)
The CWM standardizes a complete, comprehensive metamodel that enables data mining across database boundaries at an enterprise and goes well beyond. Like a UML profile but in data space instead of application space, it forms the MDA mapping to database schemas. The product of a cooperative effort between OMG and the Meta-Data Coalition (MDC), the CWM does for data modelling what UML does for application modelling.

3.1.2. CORBA
CORBA is the acronym for Common Object Request Broker Architecture, OMG's open, vendor-independent architecture and infrastructure that computer applications use to work together over networks. Using the standard protocol IIOP, a CORBA-based program from any vendor, on almost any computer, operating system, programming language, and network, can interoperate with a CORBA-based program from the same or another vendor, on almost any other computer, operating system, programming language, and network. CORBA is useful in many situations because of the easy way that CORBA integrates desktop and servers from so many vendors.

· Technical overview

In CORBA, client and object may be written in different programming languages. CORBA applications are composed of objects . For each object type, an interface is defined in OMG IDL (Interface Definition Language) . The interface is the syntax part of the contract that the server object offers to the clients that invoke it. Any client that wants to invoke an operation on the object must use this IDL interface to specify the operation it wants to perform, and to order the arguments that it sends. When the invocation reaches the target object, the same interface definition is used there to parse the arguments and perform the requested operation.

This separation of interface from implementation is the essence of CORBA. The interface to each object is defined very strictly. In contrast, the implementation of an object - its running code, and its data - is hidden from the rest of the system (that is, encapsulated). Clients access objects only through their advertised interface, invoking only those operations that that the object exposes through its IDL interface, with only those parameters (input and output) that are included in the invocation.

Figure 17 - CORBA request from client to object

The above figure shows how everything fits together, at least within a single process: The IDL is compiled into client stubs and object skeletons . Passing through the stub on the client side, the invocation continues through the ORB (Object Request Broker) , and the skeleton on the implementation side, to get to the object where it is executed. Because IDL defines interfaces so strictly, the stub on the client side matches perfectly with the skeleton on the server side, even if the two are compiled into different programming languages, or even running on different ORBs from different vendors.

Figure 18 - CORBA remote invocation flow using ORB-to-ORB communication

The figure above diagrams a remote invocation. In order to invoke the remote object instance, the client first obtains its object reference. To make the remote invocation, the client uses the same code that it used in the local invocation we just described, substituting the object reference for the remote instance. When the ORB examines the object reference and discovers that the target object is remote, it routes the invocation out over the network to the remote object's ORB. Although the ORB can tell from the object reference that the target object is remote, the client cannot. This ensures location transparency - the CORBA principle that simplifies the design of distributed object computing applications.

3.1.3. UML

The Unified Modelling Language (UML) is a open method for specifying, visualizing, constructing, and documenting the artefacts of software systems, as well as for business modelling and other non-software systems. It represents a collection of the best engineering practices that have proven successful in the modelling of large and complex systems, addressing the needs of user and scientific communities.

· Medieval ages

Identifiable object-oriented modelling languages began to appear between mid-1970 and the late 1980s as various methodologists experimented with different approaches to object-oriented analysis and design. Several other techniques influenced these languages, including Entity-Relationship modelling and the Specification & Description Language. These early methods could not satisfy most of the design requirements, and started to incorporate each other's techniques. A few clearly prominent methods emerged, including the OOSE, OMT-2, and Booch'93 methods. Each of these was a complete method, and was recognized as having certain strengths. In simple terms, OOSE was a use-case oriented approach that provided excellent support business engineering and requirements analysis. OMT-2 was especially expressive for analysis and data-intensive information systems. Booch'93 was particularly expressive during design and construction phases of projects and popular for engineering-intensive applications.

· All for one

The UML started out as collaboration among three outstanding methodologists: Grady Booch (Boock'93), Ivar Jacobson (OOSE), and James Rumbaugh (OMT-2). At first Booch and Rumbaugh sought to unify their methods with the Unified Method v. 0.8 in 1995; a year later Jacobson joined them to collaborate on the slightly less ambitious task of unifying their modelling languages with UML 0.9.

The user community quickly recognized the advantages of a common modelling language that could be used to visualize, specify, construct and document the artefacts of a software system. They enthusiastically applied early drafts of the language to diverse domains ranging from finance and health to telecommunications and aerospace. Driven by strong user demand, the modelling tool vendors soon included UML support in their products.

· One for all

At the same time that UML was becoming a de facto industry standard, an international team of modelling experts assumed the responsibility to make the language a formal standard as well. The "UML Partners", representing a diverse mix of vendors and system integrators, began working with the three methodologists in 1996 to propose UML as the standard modelling language for the OMG. The Partners organized themselves into a software development team that followed a disciplined process. Since the process was based on an iterative and incremental development life cycle, the team produced frequent "builds" and draft releases of the specification.

The Partners tendered their initial UML proposal to the OMG (UML 1.0) in January 1997. After nine months of intensive improvements to the specification, they submitted their final proposal (UML 1.1) in September 1997, which the OMG officially adopted as its object-modelling standard in November 1997. UML 1.5 is the current specification adopted by the OMG, and in mid-2001, OMG members started work on a major upgrade to UML 2.0. Four separate RFPs - for UML Infrastructure, UML Superstructure, Object Constraint Language, and UML Diagram Interchange - keep the effort organized.

The OMG defines object management as software development that models the real world through representation of "objects." These objects are the encapsulation of the attributes, relationships and methods of software identifiable program components. A key benefit of an object-oriented system is its ability to expand in functionality by extending existing components and adding new objects to the system. Object management results in faster application development, easier maintenance, enormous scalability and reusable software.

OMG members are preparing to standardize a Human-Usable Textual Notation (HUTN) for UML models, or at least those that fit into the Enterprise Distributed Object Computing (EDOC) UML Profile.

· (Almost) Ten Years Later

UML defines twelve types of diagrams, divided into three categories:
§ Structural Diagrams represent static application structure and include the Class Diagram, Object Diagram, Component Diagram, and Deployment Diagram.

§ Behavior Diagrams represent different aspects of the application's dynamic behavior and include the Use Case Diagram (used by some methodologies during requirements gathering); Sequence Diagram, Activity Diagram, Collaboration Diagram, and Statechart Diagram.

§ Model Management Diagrams represent ways the applications modules can be organized and managed and include Packages, Subsystems, and Models.

The Advanced UML Features add to the expressiveness of UML:

§ Object Constraint Language (OCL) has been part of UML since the beginning and express conditions on an invocation in a formally defined way: invariants, preconditions, post conditions, whether an object reference is allowed to be null, and some other restrictions using OCL. The MDA relies on OCL to add a necessary level of detail to PIMs and PSMs.

§ Action Semantics UML Extensions are a recent addition and express actions as UML objects. An Action object may take a set of inputs and transform it into a set of outputs, or may change the state of the system, or both. Actions may be chained, with one Action's outputs being another Action's inputs. Actions are assumed to occur independently - that is, there is infinite concurrency in the system, unless you chain them or specify this in another way. This concurrency model is a natural fit to the distributed execution environment of Internet applications.

UML Profiles tailor the language to particular areas of computing or particular platforms. In the MDA, both PIMs and PSMs will be defined using UML profiles; eventually OMG will define a suite of profiles that span the entire scope of MDA. Examples of three supporting UML Profiles and one specialized profile are :

§ The UML Profile for CORBA defines the mapping from a PIM to a CORBA-specific PSM.

§ The UML Profile for EDOC is used to build PIMs of enterprise applications. It defines representations for entities, events, process, relationships, patterns, and Enterprise Collaboration Architecture.

§ The UML Profile for EAI defines a profile for loosely coupled systems . These modes are typically used in Enterprise Application Integration, but are used elsewhere as well.

§ A UML Profile for Schedulability, performance, and time supports precise modelling of predictable systems, precisely enough to enable quantitative analysis.

· Opening the iron mask

UML is clearly an open methodology, and the three main characteristics are the independency of the technical infrastructure, the independency of the methodology and the openness of its definition, managed by a recognised and independent organisation (OMG).

At first, UML can be used to model any type of application, running on different combinations of hardware and software platforms, programming languages and network protocols. Built upon the MOF metamodel which defines class and operation as fundamental concepts, UML is a natural fit for object-oriented languages and environments like C++, C#, Java and Python. Some UML tools analyse existing source code and reverse-engineer it into a series of UML diagrams. Other tools execute UML models in interpretative way (to validate the design) and other may even generate program language code from UML.

The process of gathering and analysing an application's requirements, and incorporating them into a program design, is a complex one and the industry currently supports many methodologies that define formal procedures specifying how to go about it. The second characteristic of UML is that it is methodology-independent. Regardless of the methodology used to perform the analysis and design, UML can express the results. Using XMI (XML Metadata Interchange, another OMG standard), the UML model can be transferred from one tool into a repository, or into another tool for refinement or the next step in the chosen development process.

At last, the UML definitions are openly discussed by an independent organisation - OMG, freely published in the Web and implemented by any software vendor, including several open source projects.

The OMG is structured into three major bodies, the Platform Technology Committee (PTC), the Domain Technology Committee (DTC) and the Architecture Board. The consistency and technical integrity of work produced in the PTC and DTC is managed by an overarching Architectural Board. Within the Technology Committees and Architectural Board rest all of the Task Forces, SIGs, and Working Groups that drive the technology adoption process of the OMG.

There are three major methods of influencing the OMG process, in addition to the impact of general review, commentary and open discussion. The first is the ability to vote on work items or adoptions in the Task Forces that are ultimately reviewed and voted on at the Technology Committee level. The second is the ability to vote on work items or adoptions at one or both of the Technology Committee levels. The third is the ability to actually submit technology for adoption at one or both of the Technology Committee levels. Membership fees are based on these levels of influence.

3.2. Web Platforms

3.2.1. Java

The Java platform is based on the power of networks. Since its initial commercial release in 1995, Java technology has grown in popularity and usage because of its true portability. Java is a general-purpose concurrent object-oriented programming language. Its syntax is similar to C and C++, but it omits many of the features that make C and C++ complex, confusing, and unsafe. Java was initially developed to address the problems of building software for networked consumer devices. It was designed to support multiple host architectures and to allow secure delivery of software components. To meet these requirements, compiled Java code had to survive transport across networks, operate on any client, and assure the client that it was safe to run.

· Java platform

A platform is the hardware or software environment in which a program runs. We've already mentioned some of the most popular platforms in the chapter 2. Most platforms can be described as a combination of the operating system and hardware. However, Java is a software-only platform that runs on top of other hardware-based platforms.

The Java platform was designed to run programs securely on networks and allows the same Java application to run on many different kinds of computers. For example, PersonalJava applications power home appliances, Java Card applications run on smart cards, smart rings, and other devices with limited memory, and Java TV applications run in television settop boxes. This interoperability is guaranteed by a component of the platform called the Java virtual machine (or "JVM") - a kind of interpreter that turns general Java platform instructions into tailored commands that make the devices do their work.

Figure 19 - Java platform components

The other component - Java API - is a large collection of ready-made software components that provide many useful capabilities, such as graphical user interface (GUI) widgets. The Java API is grouped into libraries of related classes and interfaces; these libraries are known as packages.

The power of compiled languages is the execution speed. The power of interpreted languages is the flexibility to run a same program in different platforms. The Java programming language is unusually powerful in that a program is both compiled and interpreted. With the compiler, first you translate a program into an intermediate language called Java bytecodes -the platform-independent codes interpreted by the interpreter on the Java platform. The interpreter parses and runs each Java bytecode instruction on the computer. Compilation happens just once; interpretation occurs each time the program is executed.

Figure 20 - Java compiler and interpreter

· Java editions

Three editions group the different technologies according to the hardware platform: J2ME (tiny commodities, like smartcards), J2SE (development of applets and applications) and J2EE (enterprise server-side applications).

Figure 21 - The Java platform and editions

Java is owned by Sun Microsystems, which has in place the appropriate mechanisms to meet the evolving needs of the industry, and to fulfil the needs of Internet application developers .

· Java buzzwords

In the scope of this document, let us analyze the two main Java characteristics in detail, and briefly describe the others:

§ Architecture neutral

The independence of the architecture is fundamental to allow the same program to run virtually anywhere, without special adaptation tasks. For Java, this is guaranteed by the Java platform components seen above. This is seen as a menace by companies like Microsoft, which tried to block the compatibility of the JVM with recent Windows versions. This was a strategy to force the Windows customers to use the Microsoft platform - .NET - and luckily has been avoided by the justice.

§ Portable

Portability is closely related to the architecture independence. It describes the characteristics of the language that must remain unchanged independently of the platform, like data types (e.g. character, byte, integer). Even if this may seem simple at first sight, this is not true for most of the languages, in which one of the activities when porting a program from one platform to another is the adaptation of the data types. Java did not make this mistake. The data types have specific, defined lengths regardless of the system.

The Java virtual machine is based primarily on the POSIX interface specification - an industry-standard definition of a portable system interface. Implementing the Java virtual machine on new architectures is a relatively straightforward task.

§ Simple - Reduces software development cost and time to delivery

§ Object oriented - To function within increasingly complex, network-based environments, programming systems must adopt object-oriented concepts

§ High performance - Obtained from the compiled bytecode

§ Interpreted - The bytecode is transformed into executable code during the execution of the program

§ Multithreaded - Multiple tasks can be executed in parallel, improving overall system performance

§ Robust - Provides a solid, reliable environment in which to develop software

§ Dynamic - Classes are linked only as needed. New code modules can be linked in on demand from a variety of sources, even from sources across a network.

· Applets

Sun's HotJava browser showcases Java's interesting properties by making it possible to embed Java programs inside HTML pages. These programs, known as applets, are transparently downloaded into the HotJava browser along with the HTML pages in which they appear. Before being accepted by the browser, applets are carefully checked to make sure they are safe. Like HTML pages, compiled Java programs are network and platform-independent. Applets behave the same way regardless of where they come from, or what kind of machine they are being loaded into and run on.

With Java as the extension language, a Web browser is no longer limited to a fixed set of capabilities. Programmers can write an applet once and it will run on any machine, anywhere. Visitors to Java-powered Web pages can use content found there with confidence that it will not damage their machine.

· Open Belgian Cathedrals

According to Tony Mary , openness and flexibility are the main Java advantages. This openness could be widened further, with a creation of a unique media technology by a group of companies. This could be compared - accorded to him - to the cathedrals. In opposition to the medieval castles - strongly protected and closed, and almost destroyed today - the Cathedrals always remained open and survived.

According to Edy Van Asch , Open Source is the main Java trend. Java has always been closely related to Open Source, but since 2001 this relationship became stronger. Examples are Eclipse (Open Source Java Development environment), Junit (Open Source Test Structures), Jboss (Open Source Application Server) and Jini.

Jini network technology is an open architecture that enables developers to create network-centric services - whether implemented in hardware or software - that are highly adaptive to change. Jini technology can be used to build adaptive networks that are scalable, evolvable and flexible as typically required in dynamic computing environments. Jini Offers an open development environment for creative collaboration through the Jini Community, and is available free of charge with an evergreen license. It extends the Java programming model to the network (by moving data and executables via a Java object over a network) and enables network self-healing and self-configuration.

3.2.2. .NET

.NET (read as "dot-net") is both a business strategy from Microsoft and its collection of proprietary programming support for the Web services . Its goal is to provide individual and business users with a seamlessly interoperable and Web-enabled interface for applications and computing devices and to make computing activities increasingly Web browser-oriented. The .NET platform includes servers (running Microsoft Windows), building-block services (such as Web-based data storage), device software and Passport (Microsoft's identity verification service). Functionally the .Net architecture can be compared to the Java platform. In the scope of our study, they are completely different: Java aims to give the companies and users the liberty of choosing the hardware and software platforms. .NET is tightly related to Microsoft Windows operating systems (the only software platform supported) and favours the usage of complementary Microsoft products such as Internet explorer and the Office suite. Omnipresence is the goal, and the weapon is the relative complexity of Java platform.

The full release of .NET is expected to take several years to complete, with intermittent releases of products such as a personal security service and new versions of Windows and Office that implement the .NET strategy coming on the market separately.

· Dot objectives

According to Bill Gates, Microsoft expects that .NET will have as significant an effect on the computing world as the introduction of Windows. One concern being voiced is that although .NET's services will be accessible through any browser, they are likely to function more fully on products designed to work with .NET code.

§ The ability to make the entire range of computing devices work together and to have user information automatically updated and synchronized on all of them
§ Increased interactive capability for Web sites, enabled by greater use of XML rather than HTML
§ A premium online subscription service, that will feature customized access and delivery of products and services to the user from a central starting point for the management of various applications, such as e-mail, for example, or software, such as Office .NET
§ Centralized data storage, which will increase efficiency and ease of access to information, as well as synchronization of information among users and devices
§ The ability to integrate various communications media, such as e-mail, faxes, and telephones
§ For developers, the ability to create reusable modules, which should increase productivity and reduce the number of programming errors

· Dot components


Figure 22 - The Components of Microsoft .NET-Connected Software

§ Smart clients: "Smart" is the term used by Microsoft for client computers using one of its proprietary operating systems: Microsoft Windows XP, Windows XP Embedded, and Windows CE .NET. All other clients platforms are excluded from the .NET architecture.

§ XML Web services let applications share data, and invoke capabilities from other applications without regard to how those applications were built, what operating system or platform they run on, and what devices are used to access them. While XML Web services remain independent of each other, they can loosely link themselves into a collaborating group that performs a particular task.

§ Developer Tools - Microsoft Visual Studio .NET (with high-level programming languages like Visual Basic, Visual C++ and C#) and the Microsoft .NET Framework (The application execution environment, which include components like the common language runtime, a set of class libraries for building XML Web services, and Microsoft ASP.NET) aim to supply a complete solution for developers to build, deploy, and run XML Web services.

§ Servers - In opposite to the portability of Java applications, only servers running Microsoft operating systems (Windows 2000 Server, Windows Server 2003, and the .NET Enterprise Servers) are certified to fully obtain good results from the .NET platform.

· Dot integration

Microsoft considers XML "deceptively simple" , and would certainly prefer a complex and proprietary solution to integrate the .NET platform - composed by Microsoft operating systems and basic software - with other vendors environments. XML is revolutionizing how applications talk to other applications - or more broadly, how computers talk to other computers - by providing a universal data format that lets data be easily adapted or transformed. Therefore, Microsoft saw in XML an opportunity to connect its completely proprietary platform with other proprietary or open environments. As an example, XML is extensively used by Microsoft Host Integration Server 2000 to provide application, data, and network integration between .NET platforms and host systems (Mainframe and AS/400 environments, considered "legacy" by Microsoft, and estimated to contain 70 percent of all corporate data).

Figure 23 - Microsoft Host Integration Server

3.2.3. Java.NET

The comparison between Java and .net is not only technical. Economical and cultural aspects are far more important. In this chapter, we will analyse some of the differences and exploit possible alternatives.

· Co-existence

Considering the easy integration of both .NET and Java platforms, by the usage of XML and Web services, many analysts suggest a co-existence of both platforms in most companies. Scott Dietzen foresee this to happen in the next five years. One of his arguments is the current .NET power in the client-side and J2EE in the server-side. But he also prophesises that the integration of Java to the companies culture, and its implementation should become easier, before Microsoft .NET technology is mature enough to attract large customers in projects where reliability is too important. Nico Duerinck agrees with the co-existence to be possible, mainly if Web services are implemented, even if a reduced number of companies are currently using both technologies.

· Portability

For new projects, Java may be clearly preferred to ensure the portability of the applications. It guarantees that a server may start small, running Windows operating systems, and move to other platforms like Linux or UNIX to satisfy more advanced requirements like parallelism, high-availability and reliability. An application developed under .NET will always require a windows-based server to be able to exploit all its capabilities.

Additionally, as described by Dietzen, the Java community is composed by more than 100 vendors, working to improve the Java features and developing complementary and concurrent solutions. The innovation resulted is far beyond what a single company can afford.

· Symbiosis

According to David Chappell , Windows DNA (seed of .net) and Java appeared together in 1996. Since then, J2EE was built using windows DNA interesting concepts, and .net framework incorporated Java-like features. This "mutual and cross-pollenization" helps innovation in both senses and minimize the problem of "single source of creativity", suggested by Dietzen.

· Infrastructure Costs

According to Duerinck, most of the companies will try to keep the existing infrastructures, and will tend to implement .net platforms in windows systems and Java for all the other architectures. It is often a simple financial decision. If a Microsoft license exists, it will probably be kept and exploited. The cost of .net framework in existing windows systems is very low. Some Java application servers can be rather expensive, however cheap and open source alternatives exist.

The training costs and simplification of the hardware and operating systems diversity is also an important point to be considered.

· Complementary products

Products are often developed by independent companies to be compatible with both platforms. This is the case of Compuware development products and Real Software development architecture . This can also minimize the costs to transport an application from one platform to another, by reducing the effects of .net proprietary philosophy.

3.2.4. The outsiders: LAMP

 

The term LAMP (Linux, Apache, MySQL, Perl / Python / PHP) refers to a set of Open Source software tools that allow for the development and deployment of web applications .

All of the omponents of LAMP can be downloaded without any costs. The support is done by a network of developers aggregated in online communities like "OnLamp " to help each other get the most out of LAMP. Training is available from a wide array of providers, and many consulting firms offer advanced capabilities for those businesses that require sophisticated LAMP development.

I have used LAMP to develop my web site . All dynamic pages have been developed in PHP, and the databases in MySQL. The website is hosted by ALL2ALL on Linux and Apache servers. The usage of Open Source allows the web hosting to be performed with low costs. The learning cost was a book about PHP and MySQL complemented by a large amount of free online information, from FAQS to complete systematic training. The learning time was reduced by using the large amount of online examples. Most of the sites using LAMP provide the sources used for their development.

§ Linux

See chapter 2.2.3 for a large discussion on Linux.

§ Apache

The most widely used web server software in the world , Apache can be run on a variety of operating systems, including AIX, Digital UNIX, FreeBSD, Irix, Mac OS X, Netware, OpenBSD, Solaris, SunOS, Windows, and of course, Linux. Apache's security record is far better than that of Microsoft IIS.

§ MySQL

This database was built with the philosophy that a web database should be lean and fast. It doesn't incorporate the diverse array of application server features that Oracle and Microsoft database tools do, instead focusing on core performance and leaving enhanced functions to scripting languages. MySQL runs on a variety of operating systems, including AIX, FreeBSD, Mac OS X, Solaris, Windows, and Linux. It is well known for its reliability; while the recent "Slammer" worm that crippled much of the Internet was spread through Microsoft SQL Server, it did not affect MySQL databases.

§ Perl

This scripting language was invented to help overworked system administrators. It still performs that function admirably well, but is also often used in web development for things like dynamic forms, server monitoring, and database integration. Perl runs on a wide variety of operating systems.

§ PHP

Developed from the ground up as a Web developer's language, PHP is easy to use and executes very quickly. Skilled developers can use PHP to build everything from online forms to complex database-driven web applications. PHP is extremely popular, runs on a variety of operating systems, and is most often used in conjunction with MySQL.

One of the most powerful PHP features is the extensive library of functions, open to be expanded by volunteer work. It contains functions to address any possible need, to communicate with a large range of databases, and to read and write files in many different formats, like PDF. The documentation is written by volunteer work and consists in examples supplied by the community. For any developer trained in other languages the learning by example process is extremely efficient.

§ Python

A general-use high-level language, Python is often used to tie different backend pieces together. It lends itself particularly well to Java/XML integration and development of dynamic content frameworks. Python runs on several operating systems.

§ Trends

To give a large impulse to this initiative, Sun unveiled a dual-processor server - LX50 - that runs the new Sun distribution of the Linux OS. The server will run Intel x86 chips and will be available not only with Linux but also with an Intel-enabled version of Sun's own Solaris operating system. In a briefing with analysts and press, Sun's new executive vice president of software, Jonathan Schwartz, further detailed the vendor's Linux strategy. While Sun will keep pushing Solaris in the data centre and J2EE for middle-tier business logic, it is making a major bet on LAMP as the application environment of choice at the edge of today's corporate infrastructures .

A company called BD-X used LAMP as the basis for its web services strategy. BD-X's chief technology officer Kevin Jarnot started prototyping a wide-area content-creation environment that would allow BD-X to provide XML-based financial content on demand for a wide variety of subscription customers. "At first, our decision to use Open Source was to build a prototype without a lot of cost," according to Jarnot. But after his prototype was built, Jarnot found, "it would actually be fine for [the] project's required performance and scalability."

BD-X founders reasoned that if they could transcribe these calls, mark them up into XML, and then offer them to business customers, they might just have a viable business. In the end, Jarnot says that 60 percent of the reason BD-X ended up with an Open Source solution was because of cost. Thirty percent, he says, was because "we were doing some pretty bleeding edge stuff... if we had been using proprietary software, it would have been a lot harder to do the integration." Oh, and the last ten percent of why BD-X went with Open Source? "Coolness factor," says Jarnot.

Jarnot said he expected BD-X would eventually need to port the system to a more production ready commercial environment like Solaris running Macromedia's ColdFusion. But, Jarnot said, it turned out that his Open Source prototyped version was good enough to take into live production. Using the standard Open Source LAMP stack as a backend platform, they also decided to use the Open Source ELVIN message-oriented middleware to filter the transcriptions and deliver them to customers. On the XML side, the developers used Jedit (Open Source Java text editor) for XML markup and editing. Jarnot mentions quite a few options he is considering for using J2EE and Microsoft .NET technologies in conjunction with his Linux, PHP and Apache systems. What proves again the clear advantages of the symbiosis theory, between commercial, proprietary and open platforms.

3.3. Web Services

According to Brad Murphy , "by now, virtually every company is fully aware and engaged in various stages of integrating the Internet into their business model to help serve customers better. A key part of every integration strategy is the use of Web services. The phenomenal growth of Web services is made possible by open standards that are built into the products that comprise the Web services infrastructure".

With a standards based infrastructure, companies have the ability to use the Web rather than your own computer for various services, linking applications and conducting B2B or B2C commerce regardless of the computing platforms and programming languages involved. In a way, Web services are the natural evolution of the Internet - again all based on open standards. At this point, Web services may seem straightforward, but a critical issue companies face is choosing the best enterprise/strategic development platform for creating and deploying these new interoperable applications.

· Open standards

Several options are available, but the key thing to remember is that any Web services development platform - by definition - must follow open standards to be interoperable, be innovative with rich and robust functionality and work properly in responding to and addressing the complex needs of a typical enterprise organization. The W3C has working groups focused on refining the SOAP 1.1 and WSDL 1.1 specifications, which should improve things considerably. The XML Protocol Working Group is working on SOAP 1.2 while the Web Services Description Working Group is creating the WSDL 1.2 specification. Meanwhile, the IETF and OASIS are also working on standardizing Web service specifications, including DIME and WS-Security.

While work at the W3C focuses on new versions of the core Web services specifications, a separate organization is focusing more attention on interoperability. The Web Services Interoperability Organization (WS-I) is focused on defining best practices for ensuring Web service interoperability. The WS-I Basic Profile Working Group is currently developing a set of recommendations for how to use the core Web service protocols like SOAP 1.1 and WSDL 1.1 to maximize interoperability.

The clear leaders in the race to deliver Web services functionality are IBM and the entire Java / J2EE vendor community. They recognized nearly five years ago the need to support standards that provide customers the flexibility of choice and as a result, devoted serious resources to creating its solution. As discussed on chapter 3.2.2, the Microsoft approach is very different when it comes to implementation. IBM and the Java community are committed to a Web infrastructure based on open standards - especially Java and Linux - to make disparate systems work together. Microsoft presents .Net as an open platform that supports Web standards, but it is still a highly proprietary technology that runs solely on the Windows platform. Microsoft supports Web standards interoperability - such as XML and SOAP - but only within the proprietary Windows framework that is not easily portable.

· Proprietary platforms

Locking-in Web services to the Windows platform may be a perfectly acceptable solution for a small or mid-tier business. However, for larger organizations, this approach is not only impractical but also caries with it risks and hidden costs when addressing the integration and collaboration challenges of significant internal applications as well as a large, dispersed organization. It's important to consider that almost every large company has a rich, complex heterogeneous computing environment.

Another critical factor to consider is that a commitment to only one operating environment such as Windows means companies lose the ability to choose and negotiate among Web service application vendors for the best function and price. Will companies really want to marry their Web services application development -including the possibility of a substantial portion of their future Internet-based revenue and vital customer relation activities - so closely to Microsoft?

Microsoft's commitment to standards and interoperability in the Web services world is both welcome and critical to large-scale adoption in the marketplace. The fact that Microsoft has embraced standards and interoperability with the Java world is evidence that Microsoft recognizes customers will no longer tolerate proprietary solutions. It's also true that the .NET offerings will likely capture a large segment of the existing pool of Microsoft developers and customers. Unfortunately for larger customers however, this won't produce increased platform choice or vendor flexibility.

If future Web usage, cost savings and revenue generation is only half of what is projected, it is likely that most companies will continue their investment and support of Java as their strategic platform. Will .NET gain acceptance and find an important role to play in a Java/Standards-based world? Absolutely, but only if Microsoft delivers on their promise to truly support standards. Which raises an interesting question - does history give anyone reason to believe that Microsoft is truly interested in customer choice and flexibility based on vendor neutral standards? Ignorance is bliss.

· Open source

Not only big names (IBM, Microsoft, BEA) are behind web services software. Open Source software and tools exist such as the Apache Tomcat servlet engine , the JBoss J2EE-based server and Apache AXIS (a Java toolkit for building and deploying Web service clients and servers) .

According to Thomas Murphy, "Consider the fact that Web services themselves are built on top of Open Source technologies like HTTP and TCP/IP, people don't realize the amount of Open Source technology that underlies everything that they do. Many companies use Open Source technology areas other than Web services without giving it a second thought, and so should open themselves to using the technology for Web services as well. Look at the Internet - the most predominant server is Apache, which is Open Source, and so people don't really have problems running Open Source software. And look at the underlying protocols; they're all Open Source as well."

Those involved in the Open Source community, not surprisingly, tout the benefits of Open Source tools for Web services development. Marc Fleury, founder and president of JBoss says that many commercial Web service tools are "pricy implementations rushed to the market with poor quality." He believes that because of this, most commercial vendors will disappear over time, but that Open Source technology will survive because of its superior quality. Additionally, he adds, Web services technology is a "moving target. Many implementations are fighting for standard status. Going with a free software implementation guarantees you the maximum probability of going with a standard." It's not only the Open Source community and analysts who believe that Open Source technologies are the best solution to Web services development - those involved in Web services creation and deployment are backers as well. For example, FiveSight Technologies provides comprehensive Web services workflow integration and software tools, and they've built those tools using Open Source software.

Paul Brown, president of FiveSight, says that "Without using Open Source, we wouldn't have been able to launch our company. If FiveSight on its own, or any other company, had to implement XML schema or WSDL or any other number of Web services technologies in combination, it would be an impossible task. The Open Source community is a catalyst for innovation in software, and so I know things like where we can get a good Open Source implementation of a transaction manager. It's an opportunity to solve a hard problem by building on work from the community at large. We've used Open Source in our development work, from the first piece of software we deployed. We didn't have the money to pay for developers and staff," and so instead turned to Open Source software, which already had the software available, to do the job.

Murphy also notes the drawbacks. When companies devote themselves to using Open Source technology for Web services, they're taking on responsibility for product support and management, since there is no commercial vendor that takes care of that for them. That means no technical support, and no clear upgrade paths. There may also be legal issues involved with Open Source licenses, and so businesses need to have their legal staffs examine the implications of using Open Source before committing.

Another issue is that the Open Source community has yet to fully embrace Web services technologies with open arms. JBoss's Fleury, for example, says that "Web services isn't real so far. We see zero dollars in Web services." There are signs, however, that that is changing and an increasing number of Open Source tools and developers have turned their attention to Web services.

3.4. Agile Development

Agile Development is a collection of methods and practices aimed to reduce the time and effort required to develop an application, through communication, simplicity, feedback, courage and humility. Some of its processes are Adaptive Software Development , Crystal , Scrum , Xbreed , Dynamic Systems Development Method (DSDM) and Usage-Centered Design (UCD) . However, the best known agile method is XP (Extreme Programming).

3.5. Extreme Programming

In the early 1990s, a man named Kent Beck was thinking about better ways to develop software. Kent worked together with Ward Cunningham to define an approach to software development that made every thing seem simple and more efficient. Kent contemplated on what made software simple to create and what made it difficult. In March of 1996, Kent started a project using new concepts in software development: the result was the Extreme Programming (XP) methodology. Kent defined Communication (programmers communicate with their customers and fellow programmers), Simplicity (the design is simple and clean), Feedback (quick feedback by testing the software starting on day one), and Courage (the programmers are able to courageously respond to changing requirements and technology) as the four values sought out by XP programmers.

Extreme Programming (XP) is actually a deliberate and disciplined approach to software development, and it's successful because it stresses customer satisfaction. The methodology is designed to deliver the software required by the customer in the target timeframe, and empowers the developers to confidently respond to changing customer requirements, even late in the life cycle. This methodology also emphasizes teamwork. Managers, customers, and developers are all part of a team dedicated to delivering quality software. XP implements a simple, yet effective way to enable groupware style development.

XP is an important new methodology for two reasons. It is a re-examination of the software development practices used as standard operating procedures, and it is one of several new lightweight software methodologies created to reduce the cost of software. XP goes one step further and defines a process that is simple and enjoyable.

According to Kent, software development is a difficult task and the efforts should be concentrated in the four main activities:

§ Listening - Consists in obtaining information from clients, users, managers, and business people. The problem must be identified and the data must be collected for testing purposes.

§ Designing - Some contradictory opinions exist around this topic. Some think that XP consists in low design, short-sighted. As this is far from a preferred option, let us consider the other opinion stream: the design must be performed by the developer and be validated by the user, but the time spent shall not be extremely high.

§ Coding - The heart of development.

§ Testing - Validate the application, with the help from the customer in the final phases.

3.5.1. Extreme Programming and Open Source

· XP Open Values

The four XP values fit with the hacker's ethic:

§ Most open source projects rely on communications. As normally the teams are geographically dispersed, the exchange of written messages (via e-mail, newsgroups and forums) may have a positive effect as it allows a record of the discussions and the comment from the community.

§ Simplicity is often present - and sometimes even exaggerated - in open source projects, which are normally started to satisfy basic needs of one person or one part of the community.

§ The feedback is one of the most important features of open source development, as the responses from the community can be based on the functionality (new features required, conceptual and technical problems) but are often recommendations on the programming techniques (via the sharing of the source code) or suggested new routines, functions or updates.

§ It normally takes a lot of courage to start a new open source project, as the source code may be analysed and criticised by a huge number of peer programmers, but also to join an ongoing project and commit to using the spare time to perform the coding.

· XP Open Practices

§ Planning - XP assumes you have a customer (or a proxy) as part of the project. Open Source projects generally don't have a well defined "customer". Nor is there a single voice for their users. Instead, Open source projects tend to be guided by a combination of the vision of their key developers, and the "votes" of their other developers and users.

§ Small releases - Release early, release often. This is a current practice in XP and Open Source projects

§ Testing - Automated tests are a key part of XP, as they give the courage for re-factoring and collective ownership. They are what allow the small releases (release early, release often XP and Open Source principle) to happen. Although many Open Source projects include test suites, this is possibly the best XP practice to be adopted. An Open Source project that incorporated XP style tests would have enormous advantages. It would require less oversight and review and it would encourage more people to contribute because they would have immediate feedback on whether their changes worked or not. Moreover, it would ensure that as the software changed, grew and ported, it wouldn't break.

§ Re-factoring - "Re-factoring is the process of changing a software system in such a way that it does not alter the external behaviour of the code yet improves its internal structure." Many Open Source projects are reviewed and re-factored, but often reluctantly. In XP re-factoring is recommended all the time, so that the design stays as clean as possible. The fear, of course, is that something will be broken, but this is counteracted by having automated tests, and pair (or peer) programming.

§ Pair programming - One of XP recommendations is to always have two persons working in the same programs. This is rather difficult to achieve in the real world of commercial development, but compensated by the Open Source peer programming, with an exponential number of programmers revising important projects.

§ Collective ownership - In XP collective ownership means that anyone can change any part of the project - programmers are not restricted to a certain area of expertise. Open Source projects normally work differently, having restrictions on the persons allowed to make changes directly, and the process required for submitting changes. This is necessary because of the large numbers of contributors. However, in the Open Source software all the parts of the software are visible to every programmer, who can always request access to change "restricted" parts of the code in special cases.

§ Continuous integration - Some Open Source projects integrate contributions continuously and others group them into releases. Some projects allow anyone to commit changes whereas others require changes to be submitted for review. Nevertheless, on an individual level, most hackers do practice continuous integration.

§ 40-hour week - It's hard to apply this to hackers that are working in their spare time. The main objective of this XP principle, however, is satisfied by the Open Source definition: motivation.

§ On-site customer - Most Open Source projects don't have a physical "site" or a specific "customer". However, the essence of this practice is to have good communications between programmers and users, and this has already been explained.

§ Coding standards - Most Open Source projects adhere to coding standards, for the same reasons that XP requires them: To allow different programmers being responsible for the same code at different times. On smaller projects, the standards are often informally enforced by the key developer, who adapts the contributions into his own style. Large projects usually have an explicit set of standards.

· Open Source XP software

§ JUnit is a regression testing framework written by Erich Gamma and Kent Beck. It is used by the developer who implements unit tests in Java. JUnit is Open Source Software, released under the IBM's Common Public License Version 1.0 and hosted on SourceForge

§ CurlUnit is a an instance of the xUnit architecture commonly used as a testing framework within the Extreme Programming methodology (XP) methodology. Like Junit, CurlUnit is Open Source Software, released under the IBM Public License and hosted on SourceForge.

Other resources for Open Source development under XP rules may be found in http://www.xprogramming.com/software.htm.

· Final word

Open source software can be used in Extreme Programming development, and XP concepts can be adopted by Open Source projects. The design part should be given more importance, tough, with the integration of XP, Open Source and MDA. The goal s being to reduce the time and increase the quality of open source development, and to reduce costs and improve efficiency and portability of applications developed under XP principles.

3.6. Open decision

A good starting point in deciding whether to use Open Source software is to take a look at what phase a company is, regarding the development of Web services, says Thomas Murphy, senior program director for Meta Group consulting firm: "Open Source is ideal for when you need to keep up with where technology is headed, and for that portion of your technical staff working in future technologies. It's well suited for these early stages. But when you're looking to use something for production and development, that requires a stable release path. So I think that often, people look for commercial tools at that point."

Another deciding factor is money. Open Source may be the perfect solution for companies that haven't decided yet to go full-bore into Web services, or those that are operating on a shoestring or looking to hold down costs,.

An exceedingly important issue, but one that is easy to overlook is the "cultural" factor, says Eric Promislow, senior developer specializing in Web services with ActiveState, which provides Open Source-based applications, tools and support. "Go with Open Source for Web services development if your staff already uses Open Source for other purposes, and want to stay with the tools they know and love," he says. "They know how it works, and they know how to get support for it from the Open Source community." They'll be far more productive - and happier in their jobs - than if they had to use commercial tools.

Promislow adds that Open Source software is an ideal way for companies to dip their toes in the water when it comes to Web services development, and so is suited for companies still deciding whether to seriously pursue Web services. "They can make no upfront investment except in time," he says, and so can inexpensively develop pilot projects.

Murphy warns that there are a series of not-obvious issues that companies should be aware of when they ultimately decide to go the Open Source route. "Businesses need to understand that in using Open Source, they are also taking on support and product management responsibilities," he warns. "You need someone to track where the bug fixes are, and know what the newest features are likely to be. It's not like a commercial product where someone makes money off support and so provides it for you - you have to do it yourself." Of course the existence of service companies around Open Source is increasing and reducing the effect of this problem.

Another costly activity may be installing the software. Not all Open Source software have easy installation routines, and often a commercial implementation of Open Source software is required to ensure the community support, the low cost, and technical support.

Open Source software is built by community, and is a community effort, and so businesses have to decide up-front how involved their developers can be in that community. Should companies allow developers not only to work with Open Source software, but also participate in the community, and in the community development of the software? Will managers allow developers to spend company time working on Open Source community projects? The rules shall be put in place for this ahead of time, to avoid problems in later project phases.