Petroleum softwares process numerous data and need Petroleum softwares process numerous data and need many color graphics in order to visualize and to compare results. To answer this needs L'Institut Francais du Petrole (IFP) has selected the window manager Motif based on the X Window System. We present a model which has been realized with Motif for a reservoir simulation program. In this paper, we discuss how to choose an interface builder above Motif to simplify the programming. programming.
The scope of this study is to define how to program interfaces for petroleum software used in the domain of modelization and simulation. What are the characteristics of this kind of software? A word can summarize them: "large"; they are often large programs, working on large amounts of data, producing lots of results, particularly graphics. The way we usually work with this particularly graphics. The way we usually work with this software is in three steps: pre-processing, processing and post-processing. This is a heavy mechanism. The processing post-processing. This is a heavy mechanism. The processing step can be done on a supercalculator, while the others are done on interactive machines with graphical displays. The evolution of computer technology in the last ten years has inferred new behaviors. RISC workstations increase the possibilities of local calculations. Network protocols allow processing to be distributed on various protocols allow processing to be distributed on various computers. Working on a workstation with a mouse and multiwindowing possibilities is bringing to the user a comfort to which he quickly adapts. So, the user of our programs, discovering all these powerful new tools naturally adapted to his needs would like powerful new tools naturally adapted to his needs would like to get, for the petroleum software he works with, the same facilities that he has on his workstation. He doesn't want to continue to type large amounts of data on the keyboard. He would like to downsize the heavy mechanism we described above, and to dispose of modular entities, each one in charge of a specific work. In addition to this, he definitively refuses to use programs that have an unfriendly interface. This means a graphical one offering multiwindowing, pop-up menus, buttons, an easy way to undo things and high level of interactivity. They also want to be able to do multitasking, for example processing a job while visualizing graphic results. What does this evolution mean for future software developments? First we have to examine three specific problems we encounter in our center, which is dedicated to problems we encounter in our center, which is dedicated to research:
The user or the developer of a program is not a software engineer but a researcher.
The programming often involves the modification of existing codes to implement new or better functionalities.
The principal language we use is Fortran.
We would like to provide our users with tools to help them easily integrate beautiful interfaces into their codes. The user-friendly look of an application with an interface well imagined, contributes to a better use of the software and surely to a better introduction in the market. In other words, a graphical interface is a main trump for the valorisation of a product. Of course, we focus on the X Window standard, which is the first layer tool we need to use.