By Eduardo de Moraes Ponzoni, São Paulo

4^th Semester – System Analysis and Development

For a long time, the largest investments in technology to improve storage capacity and data processing were made in large equipment, or equipment so-called “servers”. The software and the hardware for the so-called “clients” during all this time have been one step behind, specially because major information technology players such as the IBM Corporation have long said that these “clients” (micro computers or PCs) would never be able to store large amounts of data or to perform complex or heavy processing, and even worse: be networked.

So it took a long time before PCs began to raise people’s interest, which led the manufacturers of hardware and software to start making investments in these computer models.  It has allowed, over the past few years, major advances in terms of hardware and software and specially in three points that IBM at the time said they would never be efficient at or capable of: storing large volumes of data, processing heavy and complex data with satisfactory response time, and networking, including wireless networks or wireless technology known as Wi-Fi.

Today, the new fashions of hardware are tablet computers and mobile phones, while the new fashion of software include different styles of web-based and “cloud” applications.

Okay, but why is it called the “New Old Fashion” of application development, and why is it said that this is not an absolutely brand new subject? We’ll see it pretty soon. Now, let’s look back a little bit, starting with MAINFRAME applications development

Mainframe applications development

At the very beginning of application development for MAINFRAME computers, there wasn’t any kind of concern towards the interaction between the users and the applications. That was because the main focus was to build robust applications to process large amounts of data in the shortest time possible.

The application visual was always the same, based in character mode, usually 25 lines and 80 columns with a black background and green foreground. The interaction with the applications was all done using a mainframe terminal or an emulator and a regular keyboard.

The navigation through the applications was done by function keys (known as “PF”), or menu choices, where the user would type letters or numbers according to the navigation menu option.

The programming languages for MAINFRAMES also were (are) complex. Most MAINFRAME applications were (are) written in ASSEMBLER (one of the most “down level” languages known), or COBOL (which has a very hard and verbose code).

Client-Server applications development

As the personal computers began to evolve in both hardware and software, companies like Microsoft and even IBM began releasing operating systems that had a greater number of features aimed at regular users, based on graphic applications, introducing the concept of “windows”, and “windows’ components” such as dialog boxes, text, drop-down lists and processing buttons. Moreover, with the increased processing power and storage of the small computers, there were investments towards distributed computing in computer networks and then a new application development concept was born: the model known as “Client-Server”, whose fundamental idea is to distribute the application tasks and responsibilities in two layers: “client” and “server”.

The role played by client layer is as simple as request access and use of “server components” and display data or results to the user, while the server layer provides resources and/or services or server components. Server components include database access, security concerns, file sharing, business rules and all application processing.

The server layer can also be layered in other layers distributing the server responsibilities depending on its requirements and architecture.

Despite all the advances that leverage this computing model, it has some problems, both for hardware and software.

In this model, the client-side hardware does not require large storage capacities and processing, but in any case change the application or screen, requiring you to reinstall the application. In a small network with few computers using the application that does not seem to be a big problem, but imagine a large network as a financial institution for example.

Once on the server-side, there may be a growing demand for upgrading the hardware as the application becomes larger or more complex, or as the number of clients increases, or as the volume of data to be processed and stored increases.

To be continued…