Design and Implementation of the Statistical Graphic System TIGS

Literature Overview

This paper, published in Computer Engineering (Vol. 18, Issue 6, 1992, pp. 10-15) by Cai Shijie, Yin Jianwen, Ge Ruding, Li Binyu, and Dong Yi from Nanjing University, describes the design philosophy, user interface architecture, and multi-language development methodology of the Statistical Graphic System (TIGS). While this paper is not directly related to welding or pipe fabrication, it represents an early Chinese contribution to statistical visualization software design and is included in the provided literature set.

System Design Philosophy

The TIGS system was conceived as a statistical graphic system designed to facilitate the visualization of statistical data through interactive graphical interfaces. The design philosophy emphasized user-friendliness and flexibility, recognizing that statistical analysis is most effective when results can be presented in visual form for rapid interpretation and communication.

User Interface Architecture

The system featured a menu-driven user interface with two distinctive design features:

1. Operation-oriented menus: The menu structure was organized around statistical operations and graphic types rather than around data structures, allowing users to select the desired analysis or visualization method directly.
2. Rollback mechanism: A rollback (undo) mechanism was implemented to allow users to revert to previous states during the analysis process. This feature was important for exploratory data analysis where users might need to try different parameter combinations or analysis approaches and return to earlier results.

Multi-Language Development Approach

A notable aspect of the TIGS system was its development using multiple programming languages. This multi-language approach allowed the system to leverage the strengths of different languages for different components:

• Low-level language (such as assembly) for graphics rendering and performance-critical routines
• High-level language (such as C or Pascal) for statistical algorithms and data processing
• Scripting or interpretive language for user interface and menu management

This layered architecture was a practical approach in the early 1990s when development tools and libraries were less mature than they are today.

Technical Significance

Although TIGS is not a welding or materials engineering system, its design principles—operation-oriented interfaces, rollback mechanisms, and multi-language development—represent general software engineering concepts that were ahead of their time in the Chinese computing community. The system demonstrated that statistical visualization could be made accessible to non-programmers through intuitive menu-driven interfaces, which was a significant advance for scientific computing in China at that time.

Relevance to Engineering Practice

While TIGS itself is not directly applicable to pipe welding or materials engineering, the concept of statistical graphic systems is relevant to quality control and process monitoring in welding operations. Modern welding quality assurance involves the collection and analysis of large volumes of data from non-destructive testing, process parameter monitoring, and mechanical property testing. Statistical visualization tools can help engineers identify trends, detect anomalies, and make data-driven decisions about process optimization. The principles of operation-oriented interfaces and rollback mechanisms that TIGS pioneered are now standard features in modern data analysis and process monitoring software.

Study Insights

This paper represents an important historical document in the development of Chinese software engineering. The multi-language development approach and the emphasis on user-friendly interfaces demonstrate a pragmatic engineering philosophy that balanced technical capability with user accessibility. For the broader engineering community, the paper serves as a reminder that effective software tools are essential for data-driven decision-making in technical fields, and that the design of user interfaces should prioritize the operational workflow of the end user rather than the internal structure of the data.