Presented here is an algorithm for scaling binary images based on piecewise polynomial interpolation. The algorithm defines a set of convolution kernels which can be used to scale the original image data by an arbitrary scaling factor and to reduce or remove the aliasing artifacts. The convolution kernels are derived from a surface geometry that is mathematically defined over the original data. This algorithm solves a quantization error problem which had prohibited practical applications of any polynomial as an interpolant for image scaling. Its microscopic behavior has been analyzed in a software simulation testbed. It can be applied for scaling binary images in the areas of facsimile imaging and font scaling. It has been fully tested and implemented in a commercial product.

Current monotonicity analysis of optimization problems has loose ends and loopholes which confuse the understanding, teaching and use of the techniques. These embarrassments are removed in this article by redefining slightly the domain, the objective and constraint functions, and even the notion of criticality itself. This leads to a simple and clear `Strong' Monotonicity Theorem giving extensions of the two monotonicity principles are straightforward corollaries. This theorem not only incorporates Hansen, Jaumard and Lu's extensions to non-monotonic functions, but also clarifies the special elimination rules applicable to variables in the constraints but not the objective. This sets the stage for analyzing optimization problems with non-monotonic functions.

The design of many industrial and engineering systems can often be accomplished using flow graphs of various types. Examples include manufacturing processes and data processing applications, Graph Transformation Expert System, is an expert system which has been developed by WUT for applying techniques of artificial intelligence to the architectural design of data and signal processing systems. Software and hardware architectures may be defined for such systems using data flow graphs, in which nodes represent data processing steps and directed areas represent the `flow' of data between the processing steps. Starting with a user- defined generic processing graphic, this expert will transform the graph by applying transformation rules in order to specialize the processing graph to satisfy specified design goals and/or hardware constraints. Although the particular application for which this expert is designed is that of data and signal processing systems, it can provide an expert system framework for other problems specified graphically; for example, manufacturing systems, information systems, and product distribution systems.

This paper reviews the two spline methods: one with interval tension; and the other with point tension. Another spline method has been described which provides point and interval tension control. This method recovers, as a special case, the above mentioned two spline methods and provides a well controlled shape designing. An efficient algorithm has been given to display the curve scheme for the case N equals 2.

Shading of planar regions is one of the basic studies in the field of raster graphics. A new algorithm dealing with the problem is proposed in this paper. By this algorithm, the complicated planar regions need not be converted to simple planar polygons, and can be directly disposed of in the same way. The algorithm runs fast, and the time and space consuming by the shading process are considerably saved.

Plane, quadric surfaces are fundamental elements of three-dimension geometric modeling systems. Intersecting between them is the most basic operation in the system. In this paper, the intersection algorithms in GEMS 4.0 modeling system are discussed, and the method of calculating valid parameter interval to obtain valid intersection curve segments is advanced. At the end, how to calculate the NURBS parameter of intersection point is given.

Rounding of polyhedra is an important operation in CAD/CAM. This paper presents a new rounding scheme based on NURBS. Algorithms for generating edge and corner roundings are developed. Rounding surfaces generated are all NURBS surfaces and rounding radius can be precisely controlled. G¹ continuity is satisfied between roundings and their base planes and along common edges between edge roundings and corner rounding.

A product modeling system GEMS 4.0 has been implemented, in which feature representation is used at the highest level abstraction of a product model, boundary representation is used at the bottom level, while CSG model is adopted at the median level. A BRep data structure capable of modeling non-manifold is adopted. NURBS representation is used for all curved surfaces. Quadric surfaces have dual representations consisting of their geometric data such as radius, centerpoint, and center axis. The geometric data are not only necessary for the mapping of form features of the corresponding CSG models, but also useful for accelerating certain time consuming calculations such as surface-surface intersections. Boundary representation of free form surfaces are easily built by sweeping and skinning method with NURBS geometry. Set operations on curved solids with boundary representation are performed by an evaluation process consisting of four steps: intersection, construction, classification and selection.

This article introduces a high-level topological representation for general multibody systems, the spatial directed graph. This simplified topological structure emphasizes the principal aspects of motion in multilink systems and forms a concise representation for the entire class of kinematically constrained multibody systems. It is shown that certain kinematic invariants with respect to this representation allow the simple formulation of kinematic constraints. In this context, it is observed that the only variables that are amenable to an operational algebra associated with this graph are the dual velocities. This formalism represents the starting point for the automatic generation of motion equations for generic multibody systems with applications in motion animation, virtual reality, robotics.

Image alignment is an essential step for 3D reconstruction of organs from serial sections. The conventional means are proved to be either unsuitable or inefficient for microscopies of pulmonary alveolus. A technical description of the alignment problem for micro-section of soft tissues is presented. An efficient automatic alignment algorithm for 3D reconstruction of pulmonary alveolus sections, based on pyramid structure, is proposed. The resultant images of 3D reconstruction demonstrate that our algorithm is promising.

A new approach for triangulating trimmed surfaces is presented in this paper. The main idea is based on constructing triangles net. The algorithm consists of: (1) reconstruction boundaries of the trimmed surfaces; (2) constructing triangles net; (3) triangulation in narrow polygon zone; (4) accuracy checking and triangle subdivision. In this paper, topological consistency of triangles net is checked in parametric domain. Property of a triangle is appraised in 3D space. Excellent triangles net without cracks can finally be received. The algorithm has been successfully used in surface modeling system SCAD 1.0.

In this paper, we exploit the generation of images of approximations to fractals and fractal-like objects by using the generalized iterated Kronecker products and give several examples to show its application. After that, we propose a fast algorithm to estimate the fractals dimension which is generated by generalized Kronecker product under some conditions, and we also give the proof of it.

The detection of dominant points is an important problem in shape recognition. In this paper, the dominant point is divided into two kinds of point: corner point and tangent point, and a two stage method to detect the dominant point is proposed. The corner points are first detected based on the curvature smoothed by Gaussian filter, and tangent points are detected based on the curvature accumulation curve with the commonly used splitting method.

Verbal protocol analysis has been widely used to capture the processes of problem solving behavior. Early research studied humans solving well-defined problems such as cryptarithmetic and Tower of Hanoi problems, but it also extended to chess. Other research has focused on continuous and discrete process control tasks. Our work explores some of the cognitive difficulties involved in CAD model construction. It addresses mental transformations in the physical world, and the nature of cognitive loads when constructing physical models. It also investigates the transport of skills from physical world model construction to a screen world construction process. This paper describes our findings and performance measures in the construction of physical models.

The convergency of iterative processes can present unexpected behaviors. In this paper an analysis of iterative processes convergency is made, and illustrated by its application to Newton model. This leads to the definition of a dynamic relaxation scheme that enables to master iterative processes convergency in limit cases.

This paper proposes a theoretical framework, based on domain subdivision for parallel radiosity. Moreover, three various implementation approaches, taking advantage of partitioning algorithms and global shared memory architecture, are presented.

A new lighting model for natural light in cloudy sky is proposed. In order to represent cloudy scenes and the distribution of the natural light in the cloudy sky, the sky in the model is expressed by a part of a sphere which is called S surface. A layer, called C layer, is set just under the S surface to deal with the distribution of clouds. The natural light is supposed to be given by finite point light sources on the S surface, and three kinds of light components--direct light, scattering light and ambient light are used in each point light source to represent the natural light in various weather conditions. The spectrum in the model varies with the sun altitude.

Many computer graphics programmers are working in the area of scientific visualization. In this paper, we apply the surface fitting algorithm to visualize the 3D terrain and objects. How to obtain the DTM dataset by digitizer is first discussed. And then the representation of objects is described. Finally, the synthetic processing method of the DTM dataset and the objects' dataset is presented, and the high-speed display system of displaying the 3D terrain and objects is also introduced.

A new shading model for linear light source is presented. It accounts for both diffuse reflection and specular reflection of the illuminated surface. By regarding a linear light source as a directional rectangular light source with infinitesimal width, a simple formula is derived for calculating the diffuse reflection component. The specular reflection component is represented by an integration taking Phong's specular model as the kernel and evaluated by a linear approximation. Finally, an efficient shadow detection algorithm for linear light source is proposed. The images rendered with the shading model are very photo realistic.

Curves that are easy to control are useful in design. In this paper we use cubic curves using basis functions proposed by Said. Curves with these basis functions are robust and easy to control. We utilized this representation in designing curves through a given intermediate point, and as a simple application is the approximation of a quarter circle. By blending two curves, we can improve the capability of the representations.

Engineering design practice usually follows a design hierarchy. Constraints more or less reflect engineering requirements. But constraints are practically useful only when they are correctly embedded into design hierarchies. The concept of geometric features is introduced to serve the following purposes: to allow a designer to work with the conventional design hierarchy; to control the effect of constraints; and to decompose a complicated constraint based geometric modeling problem into simpler ones. Geometric features are constructed hierarchically. Although constraints are applied to low level geometric entities (points, lines, etc.), the impact on a design model is usually imposed on higher level features. In terms of constraints processing, each feature is viewed as an independent problem and its constraints are solved independently. All constraints in a design model are therefore solved hierarchically bottom up.

Shoe uppers are often constructed of panels which overlap one another and which may or may not have been thinned along the periphery. One of the visual features important to aesthetic shoe CAD is the profile variation along the panel boundary caused by either stitching or compressed gluing. Based upon an understanding of geometric requirements, the following issues are addressed: offset curve generation according to a classification of local geometry of an arbitrary 3D surface patch; cross section type handling; profile representation and rendering.

This paper put an emphasis on the importance of two points which are crucial when the aim is physically based lighting simulation. The first one is the spectral approach which considers emitted, reflected, diffused and transmitted light as wavelength dependent. The second corresponds to the different steps aiming at converting into RGB components the radiance arriving at the viewpoint through the pixels of a screen.

This paper presents a new approach to construct C¹ continuous surfaces on N-sided regions of composite B-spline surfaces. For C⁰ continuous surfaces on the N-sided regions, their smooth surfaces are constructed by the integral smoothing operation with variable integral neighbors, and the final surfaces are C¹ continuous with the original surfaces on the N-sided region boundary.

A new double layer interpolation to cubic parametric curves is derived and analyzed. In the outer layer interpolation, while (Delta) 2f <EQ 2 a cubic parametric curve is split into lines; in the inner layer interpolation, every line is split into discrete points. This approach uses 32 bit integer values and arithmetic shifts.

The quality of a plastic product is determined by the material used, the mold design, and the processing conditions. The objectives of a mold design consist of a mold cooling system, a runner system, and associated gate locations with their types and sizes, and cavity design. Meanwhile, the most important processing conditions during the fill phase are the mold temperature, the melt temperature, and the filling time. In this paper, we use genetic algorithm to optimize molding conditions, which consist of the mold temperature, the melt temperature, and the filling time, based upon the results from flow simulation.

In this paper, explicit formulas are developed for representing a uniform bicubic spline surface that passes through an array of data points. The interpolated surface in the closed case is topologically equivalent to a torus. Open surface cases are reduced to closed surface cases by introducing one or two rows of `free points' such that the spline surface wraps around its boundaries. Ordinary interpolation surfaces in open cases can thus be constructed with the same formulas. It turns to be more intuitive and effective to control and modify the shape of the resultant surfaces by adjusting `free points' than by the usual derivatives and twist vectors. The interpolation surface is obtained in a two step way and the procedure is very easy to implement. Experimental results demonstrate that the proposed formulas are practically useful.

In this paper, we derive a new basis for constructing C² quintic interpolation spline curve that is called BB spline. The C² biquintic interpolation spline surface is briefly constructed by tension-product and has properties of interpolation, C²-continuity and locality. The local shape of interpolation spline curve and surface can be modified by the change of shape-parameters. The effects of the shape-parameters are also discussed.

Collision detection is very important in virtual reality. If the objects in virtual world act as they do in physical world, the simulation of virtual world will seem more believable. Due to computational complexity, completing both accuracy and efficiency in a detection algorithm makes a serious dilemma to algorithm designer. After analyzing varieties of approaches to solving collision detection problem, we prove that most successful algorithms have to use some kinds of object related information to reduce detected object pairs and surface set pairs. Then a new approach--multiscale based space descriptions model--is proposed to solve collision detection problem.

An advanced method for the generation of 2D parametric model is presented. The method is based on designer's idea from 3D conceptual model to 2D geometry model. Based on geometric reasoning, the system effectively supports parametric modification of a given design in one view or multiple views through dimension-drive. This approach has been realized on micro computers.

In this paper, the necessary and sufficient conditions and an algorithm for G⁴ continuity between adjacent Bezier patches are presented. The effects of shape parameters for surface connection are investigated in details. The algorithm can be generalized directly to the case of higher order geometric continuity for surface connecting. It has important applications in surface modeling, surface blending and surface connecting.

The ray reflection, refraction, diffusion, and transparency are important functions for visualizing 3D objects. Every ray intersection with object boundaries and its normal direction on the intersecting point are used to simulate the optical interference of rays with objects. It is theoretically impossible, however, to construct smoothly interpolated continuous boundaries from a discrete array of sampled values, such as a set of volume elements. Although many volume visualization techniques have been proposed, it is still difficult to ensure appropriate surface topology by a simple algorithm. In the present paper, two new isosurface constructors named MMC (Modified Marching Cubes) and DC (Deformed Cubes) are investigated. MMC is a modification of Marching Cubes algorithm, which is well known as a high resolution isosurface constructor. MMC algorithm produces topologically correct triangulated isosurfaces that are guaranteed to be orientable and closed. DC algorithm is much simpler than MMC, and the produced triangulated isosurfaces are also topologically adequate, and as accurate as MMC. Experimental results and comparisons of the interpolated triangulated isosurfaces in terms of the shape precision between MMC and DC are also presented.

For many purposes, problems appear easier to understand when they are formulated by graph theory. In this paper, the recurrence relation of B-spline is modeled by a weighted digraph and a one-to-one correspondence between B-spline functions and weighted digraph paths is proved. A matrix representation of the B-spline digraph is utilized, which yields to compute the B- spline functions as matrix products.

For a triangulation of a planar polygonal region, we develop a C² interpolation scheme. A model triangle T is subdivided into a split (Delta) asymmetric with the vertices of T. We prove that there exists a C² piecewise polynomial of degree 5 on (Delta) interpolating arbitrarily given values and derivatives of orders up to 2 at the vertices and on the edges of the triangle.

In this paper the wave recursive interpolation of uniform T-subdivision scheme including wave parameter is defined. The convergence of sequences of control polygons produced by wave recursive interpolation T-subdivision scheme and the smoothness of the limit curve are analyzed.

In the paper, by combining the Coons's idea and the B-spline method, we propose the method to construct the blended B-spline surface to interpolate its four boundary curves, present an algorithm to convert the blended B-spline surface to the NURBS surface and discuss the approximation means of treating the singularity of the blended B-spline surface. At the end of the paper, the conclusions are given.

This paper presents a geometric model of breaking waves. In this model, there are three amplitude coefficient functions and two phase functions. The functions are all unknown ones concerning surf similarity parameter (xi) . Genetic Algorithms is used to obtain the optimum solution of the model for a certain wave pattern, and all unknown functions simulating several wave patterns. Experiment result has proved that Genetic Algorithms are very successful for modeling breaking waves.

The proportional division property of quadratic Bezier curves is presented and proved in the paper. Based on that property, a new computer graphic algorithm to generate quadratic Bezier curves are derived. Discussed in the paper also are the approach to plot cubic Bezier curves with the derived algorithm and two inverse fitting methods for quadratic Bezier curves under certain boundary conditions.

Aiming at detail enhancement in volume rendering, a new volume illumination model, called Composed Scattering Model (CSM), is presented. In order to enhance different details in data, scattering intensity is decomposed into volume scattering intensity and surface scattering intensity, and composed with boundary detection operators. CSM can generate images containing more details than current volume rendering models. This model has been applied to the direct volume rendering of 3D data sets obtained by CT and MRI. The resultant images show not only rich details but also clear boundary surfaces. CSM is demonstrated as an accurate volume rendering model suited for detail enhancement in volume data set.

An algorithm for constructing dynamic convex hull of n points in the plane based on the new criteria for determination of vertex type and vertex position is presented. While this algorithm has O(log n) time for update and O(nlog n) time for the total processing as existing ones, the new criteria make it more compact, explicit and reasonable. In addition, it has the ability to deal with all special cases and therefore is more powerful.

This paper has inferred a corrected wave resistance formula which can be suitable to the real ship's wave resistance computation and to be used to improve the hull's surfaces, through a method which adapts linear wave resistance theory and wave analysis. This paper uses B- spline curve surface to fit the hull's surface and uses SUMT to optimize the hull's surface. A ship form with minimum wave resistance has been obtained. The results of the computation and experiments have proven that the method, presented by this paper, is an effective method for ship improvement and design.

The researches for Bezier and polynomial curves and surfaces, and their applications to curve- fitting have been reported in many papers. However the relation between control points and polynomial coefficients, because of the complexity of computation, have rarely been studied. In this paper, we propose a new method to transform the Bezier to the polynomial representation and vice-versa. An equation is given, for generating an (m + 1) polygonal Bezier control points to approximate an (n + 1) ones. This method, unlike previous works, is more transparent because it is given in form of one equation. With this method, the curve goes through the two endpoints of the polygonal, and we do not need to perform any transformation such as Chebyshev polynomial in order to obtain good approximation. A criterion of reduction is given in order to known if a polygonal Bezier is reducible without error or not. An error estimation is also given only in terms of control points. All equations are given explicitly and in matrix form, for Bezier curves. Finally, we discuss some applications of this method to curve-fitting, order increasing and decreasing, and also its extension to rational Bezier and polynomial.

This paper presents GC¹ continuity condition between adjacent NURBS surface patches along common boundary curve and deduces specific algorithms for control points and weights of NURBS surface patch. For making another NURBS surface patch and one given NURBS surface patch to attain GC¹, according to algorithms condition, we can adjust another surface patch control points and weights. It is much more convenient for engineers to apply.

A method for constructing high precision parametric planar curve is presented. The method uses the uniform parametrization to assign nodes and evaluates slope vectors at the nodes by supposing that the given data points are taken from a parametric quadratic polynomial. The constructed curve is a C¹ piecewise parametric cubic polynomial, and it reproduces exactly a parametric polynomial of degree two or less.

In this paper, we present a method for macroscopical sampling based on geometrical constraints and apply it to circular arc global recognition, by a series of tests, it indicated that this method is simple, convenient and efficient, and it has high robustness.

A new method for control of path motion is presented. To describe the motion of an object, we adopt two independent control curves, namely the path of the object in 3D space and the velocity curve of the object along the path in 2D space. Techniques for constructing the velocity curve according to various application are presented. The task of finding the exact position of the object along the path is facilitated by employment of a look-up table which matches the arclength of the path with a corresponding parameter value. An efficient arclength parametrization algorithm for path curve of Bezier form is developed.

A special kind of points sets named DPDIS is defined in this paper, followed by an elegant algorithm for constructing the convex hull of this class of points sets. Taking advantages of the characteristics of this sets category, this algorithm finishes its work within an O(N) complexity. The crux of this algorithm is a new sorting method by which all points can be sorted along x direction in linear time.

In this paper, a new method for GC¹ interpolating cubic unrestricted curve meshes is presented. By choosing the rational combination-functions which ensure the compatibility of twists at nodepoints, quintic vector-valued first order cross-boundary derivatives of surface patches are generated. The resulting surface patches are taken to be the Coons-Boolean sum patches interpolating the boundary information obtained, and are converted to rational biquintic Bezier patches. Compared with other similar interpolation methods, this method has the advantages that it uses lower degree (biquintic) Bezier patches, and that it is local.

Let a 2D random function X(x,y) to denote fBm with exponent 0 < H < 1, then its spectral density S_x(u,v) has relation: Sx(u,v) 1/(u² + v²)^H+1. Such algorithm based on fBm has shown us beautiful pictures of fractal mountains. But the mountains (fractal surfaces) were produced naturally by random process. As a result, the macroscopic shapes and global positions of fractal mounts are not controllable. This paper presents a method that generates fractal mountains with controllable macroscopic shapes and positions using spectral synthesis. First, the discrete data of Y(x,y) on finite grids are inputted, and FFT is employed to produce discrete spectral F(u,v). Second, by InvFFT, low frequency components of F(u,v) together with high frequency components of F(u,v) are transformed to produce Z(x,y)--fractal surface. The macroscopic shapes are controlled by low frequency; meanwhile, the high frequency describes texture of fractal mountains.

In this paper, a new solution to the problem for reconstructing the surface of 3D objects over a set of cross-sectional contours is proposed. An algorithm for single branch contours connection, which is based on the closest local polar angle method, is first presented. Then an approach to solving the branching problem is proposed. Next, a method for surface shading is given. Finally, these methods have been applied to the reconstruction of the external surface of a complex shaped object such as the cellar region of human brain. The results show that the presented methods are practical and satisfactory.

Solid modeling system based on accuracy model achieves a higher level accuracy than faceted system, but many problems exist, such as complexities of geometric calculation. This article presents a new method for representing and maintaining accuracy information in a solid system. A structural representation is introduced, some operations and validity checking are explained.

To address the need for producing high quality designs with better life-cycle performance and shorter product development lead-times, the methods for developing new CAD systems that support integrated concurrent engineering design are studied. A feature-based, integrated concurrent design model, and associated software structure are proposed. In this prototype system, a design is modeled using features from the three most important perspectives of a mechanical product: design, manufacturing and geometry. A hybrid database scheme has been developed for representing the qualitative and quantitative information of these models. The methods for identifying the most cost-effective design and optimal concurrent design are illustrated using examples.

This paper concentrates on the modeling of complex surfaces over arbitrary topologies. Methods and techniques newly developed are presented for high-quality curve and surface modeling in CAD system, e.g., composing of rectangular and triangular patches, modeling of closed surfaces over polyhedral topologies, blending free-form surfaces with geometric continuity. The data structure and data exchange are based on STEP, the international standard for external representation of product data. NURBS (Nonuniform Rational B-spline Surface) representation is adopted as a unified approach to advanced curve and surface design.

Openness is very critical to the CAD kernel (supporting) systems, because a closed CAD system can not provide a deliberate solution to the practical needs. In this paper, we devise an open strategy for CAD kernel systems based on the client/server architecture. With this open strategy, the CAD kernel system can be costumed (clipping or extending) to fit the various applications' needs. Both of the implementation issues and the application development methods are explored.

In this paper, a product data model for progressive die CAD/CAM system is introduced. A special form feature representation model is proposed as part of a product data model to improve design representation. The features are defined as a form feature tree associated with various types of technical constraints such as tolerance and relations for supporting the integration of solid modeling and manufacturing applications. Several examples are presented to illustrate the model and its applications.

An efficient technique is described for rendering Finite Element Modeling (FEM) volume data. The data are not a regular 3D grid. This algorithm can deal with most kinds of FEM data, such as hexahedron 8 nodes, hexahedron 20 nodes etc. Two methods to visualize the FEM data have been presented in the rendering stage. The comparison of these two methods have also been discussed later in this paper.

CAPP (computer aided process planning) is an important bridge linking CAD and CAM. It has been a major problem to convert design into suitable forms for various applications. Two common approaches (i.e. feature recognition and design by features) have been widely used to tackle this problem, but the progress has not met industrial expectations. This paper presents a hybrid method that combines both approaches to create a feature-based component data model (FBCDM) which is suitable for machining process planning and NC programming. Illustrative examples are provided to demonstrate the concept and algorithms devised. Attempts are made to reduce human intervention to a minimum at the design stage in tagging feature-related information while making full use of geometric reasoning to generate the FBCDM. Essential manufacturability assessment is carried out during the modeling process.

This paper introduces two solutions using STEP to integrate various CAD/CAPP/CAM application systems and presents our implementation methods of the EXPRESS compiler and the STEP exchange file pre/post processor as components of Gaohua Modeling System. Some essential techniques and strategies such as data dictionary structure, schema manipulations, the STEP exchange file accessing and the Application Protocol conformance are discussed in detail.

Feature conversion, also called feature transformation and feature mapping, is defined as the process of converting features from one view of an object to another view of the object. In a relatively simple implementation, for each application the design features are automatically converted into features specific for that application. All modifications have to be made via the design features. This is the approach that has attracted most attention until now. In the ideal situation, however, conversions directly from application views to the design view, and to other applications views, are also possible. In this paper, some difficulties faced in feature conversion are discussed. A new representation scheme of feature-based parts models has been proposed for the purpose of one-way feature conversion. The parts models consist of five different levels of abstraction, extending from an assembly level and its attributes, single parts and their attributes, single features and their attributes, one containing the geometric reference element and finally one for detailed geometry. One implementation of feature conversion for rotational components within GT (Group Technology) has already been undertaken using an automated coding procedure operating on a design-feature database. This database has been generated by a feature-based design system, and the GT coding scheme used in this paper is a specific scheme created for a textile machine manufacturing plant. Such feature conversion techniques presented here are only in their early stages of development and further research is underway.

Engineering database management system (EDBMS) is the kernel of CAD/CAM system integration, and object-oriented EDBMS (OOEDBMS) is the best implementation. While STEP is becoming the standard of product data exchange and representation, supporting STEP in engineering database becomes more and more important. In this paper we introduce the architecture of STEP based OOEDBMS in our CAD/CAM integrated system GHCAD. We focus on schema management and three-grade database management in OOEDBMS. Topics such as DDL compiler, transformation from EXPRESS to DDL, DDL tools are discussed. Finally further research directions of schema management in OOEDBMS are present.

A feature-centered scheme is presented to represent the feature-based variational part model. Constraint reasoning approach combined with numerical approach is triggered to satisfy the constraints between features. Coassociations between boundary primitives in feature level and part level are maintained to implement a series of feature operations and to implement precision feature modeling for parametric part geometry model. GHCAD, a feature-based parametric modeling system based on these ideas, is developed.

This paper is focused on a mechanical product model for integrated computer aids. With the investigation on product design process and feature techniques, and from viewpoint of STEP (Standard for the exchange of Product Model Data), in this paper, the information model of mechanical product is proposed. The architecture and contents are discussed in detail. Then, a prototype system GHCAD, in which the product model is implemented, is introduced.

This paper reports on the development of a concurrent engineering system which covers four application domains in production of mechanical parts, namely design, manufacturing, assembly and inspection. Two approaches are introduced: (1) semantic feature which describes information within each domain separately, and (2) feature transformation which manipulates and derives information across different application domains. Their applications in generative process planning are illustrated in a case study.

In order to improve the global effect of placement, we introduce an efficient and fast algorithm. The main idea of this algorithm is that: the layout area is hierarchically decomposed into blocks. At each level of the decomposition, the cells in the block exchange at the same time. The min-cut exchange in the block hierarchy tree prevents local optimum solutions, as well as reduce the number of exchanges. Several questions are discussed about the algorithm. The time complexion is proved to be o(n).

The data flow diagram has been widely used in software engineering. However, previous definitions of the data flow diagram have not provided a comprehensive way to represent the behavior and structure of hardware. This paper describes an improved data flow diagram which provides notations, syntactic rules and semantic interpretations for hardware description. It is a kind of visual language and supports mixed-level and hierarchical description, module configuration, attribute specification, multilevel timing specification.

This paper presents an improved algorithm with a new cost function for multiple way network partitioning. The new cost function based on the net cut model proposed incorporate a penalty function to take account of the potential effects of cell's move. Experiments show that not only the result of the new cost function outperforms that of F-M's algorithm, but also the erratic defect of F-M's algorithm has been partially alleviated. This new algorithm is flexible enough to be applicable to different objective functions. The time complexity of the new algorithm is O(bN), where b is the number of blocks and N the number of nets.

This paper discussed a method of knowledge representations to automatically reuse software information, such as, algorithms, designs, documents, reports etc. in the domain of IC-CAD. An architecture of framework ICSDE was presented which consists of four main parts: packaging unit, prototyping unit, reusable library and data base management. Within this framework, time domain knowledge of reusable-software-information (RSI) was represented and abstracted into three concept levels. Based on this representation method, the RSI can be packaged into reusable-software-component in the reusable library, and the function was provided to help IC-CAD software tools or system developers producing their products with low time and low cost. So it is possible to provide the IC-CAD products good reliability and maintainability under this framework.

Various problems occurred in encapsulation of plastic IC packages are the results of improperly handling of the many process parameters, lack of understanding of material properties, and immature mold design. The traditional way to get rid of these problems is to adjust the process parameters of the transfer molding machine. However, the optimization of these parameters is usually based on past experiences or obtained by trial and error. This approach is time consuming and less accurate. A program is developed to help users to choose an appropriate initial machine setting based on the machine, material and mold design information. The detail of its application to IC encapsulation process is described in the paper.

In this paper, we propose a timing-driven floorplanning algorithm for general building block layout. Our basic idea is based on floorplan enumeration and Zero-Slack algorithm. In our system, floorplanning is performed hierarchically followed by an incremental global routing. Whenever a node in cluster tree is floorplanned, all possible floorplan patterns are enumerated and then the floorplan pattern with best timing behavior and area, shape matches is chosen as solution. Due to our work's background, without the cost of limited floorplan patterns in BEAR, we are awarded by good performance and routability. The complexity analysis is presented also.

The major goal of this paper is to propose a new modal logic aiming at formal verification of bit-slice circuits. The new logic is called as time-space modal logic and its major feature is that it can handle two transition relations: one for time transition and the other for space transition. As for a verification algorithm, a symbolic model checking algorithm of the new logic is shown. This could be applicable to verification of bit-slice microprocessor of infinite bit width and 1D systolic array of infinite length. A simple benchmark result shows the effectiveness of the proposed approach.

This paper presents Ternary Decision Diagrams which represent sets of products. This paper also presents manipulating methods for sum-of-products forms and ringsum-of-products forms using Ternary Decision Diagrams, and gives comparison results between Ternary Decision Diagrams and Binary Decision Diagrams.

This paper explores the theory and algorithm for the measurement and reconstruction of line industrial objects based on CAD representation, and proposes a mathematical model for line photogrammetry. In the model, the geometric elements based on CAD representation describing industrial objects are considered as unknowns and match directly from image to object. Furthermore, this paper presents an application prototype which includes the system configuration and the method of system calibration. Finally, the measurement results and a discussion are put afterwards.

In respect of the mechanismic model of L²-TE type and LERT-T type cutting robot, the authors, by analyzing sources of the errors incurred, were able to formulate equations for their compensation.

In this paper, the concepts of dynamic tested object and dynamic testing procedure are first introduced and their model descriptions are then presented in class form. General design methods for real-time interactive testing systems are further studied. Some effective approaches of interacting and testing in real-time environment are discussed in detail. From the view point of application, a design instance for the telephone test system with multiple interactions is given to demonstrate the practicality and efficiency of the methods and technology.

A fault-tolerant data transmission model based on the redundant residue number system is proposed in this paper. It can transmit data correctly between two ends unless the residue errors exceed the error-correcting capability. The expression for the probability of error is presented when the channel noise is additive Gaussian noise, and every branch is M-ary orthogonal signaling modulation. The expression of the probabilities of undetected error and uncorrected error is also obtained when the RRNS system is single error-checking and single error-correcting model respectively. The fault-tolerant data transmission model has not only the properties of parallel transmission but also the properties of series transmission.

dEDBMS is a practical distributed engineering database management system in the CAD application fields. This paper systematically introduces its system architecture, data distribution, security mechanism and concurrency control strategy etc., especially discusses in more detail the concept and implementation of the two categories and three kinds of `lock' mechanism which is presented by the authors.

The ISO STEP specifications aim to provide an effective means by which product information can be shared and exchanged between applications and enterprises. EXPRESS is a modeling language within the STEP specifications. It is used to describe product data. An EXPRESS model can be implemented on a database repository, queried upon and manipulated. This paper examines the suitability of object-oriented database systems as the repository for EXPRESS models. It then discusses the design and implementation of a translator which converts the EXPRESS language into constructs suited to object-oriented database systems.

In this paper, an efficient framework for engineering database management is presented. The data model, system architecture, system components, and some related issues, such as transaction management, hierarchy of databases, checkin/checkout model are discussed in more detail. EXPRESS is used for schema definition. This paper concentrates on the overall system architecture.

One key technology required by a CAD/CAM is a database system management. However, conventional database systems are not well suited for CAD/CAM applications because of their lack of support for complex structured data and their inability to meet the demands of dynamic changes of schemas. Object-Oriented Database systems have many advantages over conventional database systems, such as the ability to support complex objects and long transactions in CAD/CAM. The main feature of an Active Database is to react autonomously (without user intervention or application request) to events that can occur in a database. An active OODB system will meet requirements of a wide range of applications, in particular, for CAD/CAM systems. In this paper, we propose a layered architecture to represent an active OODB by putting an active facility on top of a passive OODB system. In this active layer, we use a rule/event interface to specify rules/events and to store rules/events as `first-class' objects which can be created, inquired, deleted and updated like other objects. In this paper, we will focus on the specification of expressive ECA (Event-Condition-Action) rules, detection of primitive and composite events. Finally, we will discuss how to implement the architecture using a commercial OODB such as ObjectStore.

To support the requirement of concurrent engineering (CE), we need current STEP APs interpretable to integrate different applications in different phases of the entire lifetime cycle of a product. COMposition Application Protocols (COMAP) introduced and discussed in this paper is proposed to solve this problem. COMAP, serving as a root AP, is built on top of existing STEP APs for mapping information among them. A prototype of COMAP (cadpp) based on AP203 and AP224 for CAD/CAPP integration is implemented and discussed.

Hypertext is a new kind of approach to information management in which there are information nodes and links between them. This paper first presents a hypertext data model in which node can be text, table, graph, image, audio, video, executable module or item. This model also defines version node and operations on it that may be helpful to version management. Navigation graph is also introduced via an application system.

This paper introduces concept feature-objects to support operational, structural and morphological modeling of mechanical products in the early phase of design. First, the feature-centered approaches to conceptual design are shortly summarized and evaluated. Then a possible implementation of concept feature-objects and the methodology for using them is presented. The strength of concept feature-objects is in their morphology inclusive nature. They appear as parametrized 3D skeletons providing geometrical representations for the modeled engineering conceptions. A concept feature-object models the physical ports, contact surfaces related to ports, bones between ports, DOF of ports, relevant physical parameters, scientific and empirical descriptions of intentional transformations and environmental effects. Modeling entities for a given application are constructed by generic modeling. The method of modeling and simulation with feature-objects is described.

This paper introduces a new method to represent entities and deal with constraints in parametric CAD system. The method is based on equivalent entity, named as method of equivalent element. Using the method, many kinds of complex geometrical elements and constraints could be treated as simple equivalent line(s), so the whole drawing would be translated into its equivalent drawing that is composed of only lines. It can make the parametric work become very regular and efficient.

In this paper, we try to apply Case-Based Reasoning (CBR) to macro cell layout. It presents challenges to current macro cell layout methodologies. It broadens CBR's application domains. In this paper, we introduce new approaches in case representation, case retrieval and selection, adaptation.

This paper introduces the main techniques adopted by design flow management subsystem in the OOTIF Framework. OOTIF is an object oriented tool integrate CAD framework. In this paper, we present static model of design process based on flowchart; we implement tool control through tool template which makes use of object oriented concept; we develop design flowchart builder through which designers can express design intent when submitting tasks; knowledge base is built, which makes intelligent tool selection and flow management possible. Some ideas of future developments of OOTIF are also presented. With the help of design flow management system, designers are able to concentrate exclusively on those issues concerned with the creative and exploratory phases of design.

Parametric design is an effective and productive tool for the definition and modification of geometric models. This paper presents an intelligent method for the creation of a parametric model. In the proposed method, an adaptive neural network model is built to map the set of dimensional parameters to a set of coordinates. Defining a parametric model is equivalent to teaching the neural network. The user needs only specify a set of dimensional parameters that defines the parametric model and teach the neural network how to react to the changes of the dimensional parameters. Once the neural net work is taught, any dimensional changes will result in corresponding coordinate changes. This novel method eliminate the need of programming or graphic interaction that are normally required by contemporary parametric design systems.

A new solving method for geometric constraint reasoning is presented in this paper which has been named as maximal-reduction algorithm (MRA). It is the application about constraint network theory, analysis of degree of freedom, sparse matrix, and graph theory. With the MRA, geometric constraint system can be transformed into a series of subsystem which are organized with a forest data structure finally. The strategy enables the time complexity to solve the system separated with the scale of the system. It improves the solving stability and time expenses greatly. An interesting character of the MRA is its generalization which gives itself wide applications including parametric drawing, parametric modeling, kinematics of multibody system and product assembly.

Design can be viewed as a constraint specification and satisfaction problem. Constraint-based design systems are proven to be more powerful and flexible than the conventional CAD systems. This paper discusses current constraint-based design approaches, together with their strengths and weaknesses. Our research work in constraint-based design is also introduced. Finally, some future development issues in this area are suggested.

In this paper, the technology of Artificial Intelligence is introduced into a modular design and manufacturing for machine tools. The authors present a methodology to realize the modular conceptual design combined with traditional CAD, and develop an intelligent machine tools modular conceptual system. The problem-solving strategies are described in detail. The design model and system architecture are set up. Techniques and their incorporation of expert system, case-based reasoning and artificial neural networks are clarified.

A knowledge-based computer integrated process planning system for aircraft structural parts including frames, ribs and beams in CIM environment, FA-CAPP, is presented. The main function modules of FA-CAPP are outlined. With interactive feature definition method, the integration of CAD/CAPP/CAM is achieved on the basis of hierarchical feature-based part model and supported by RMI. A technological decision-making expert system was developed to make decision for process planning. Knowledge base management system makes it easy to retrieve and update the technological knowledge base.

In this paper, the redesign technique which is one of the key technologies in Intelligent Design System is analyzed and studied in detail. Considering previous achievements in redesign field, the main difficulties and key techniques are discussed. Combining the `8.5' National Key Project, that is `Heavy Duty Machine Tools (HDMT) Modular Design and Manufacturing'. Considering the specialties of HDMT modular design and the underconstrained nature of machine tools structures, a novel hybrid redesign strategy suited for HDMT modular design problems is put forward, and the basic idea, realization method and solving techniques of the hybrid redesign strategy are described in detail. The strategy is applied to various phases in HDMT modular design procedure successfully and flexibly.

In the paper, comments on the knowledge-based design model and the case based design model are given at first, then a concept of case prototype is proposed, which integrates with the generalized knowledge and design cases. In hence, case prototype based design model is described. As an embodiment of this model, a case prototype based design support tool-- CaseDEST is presented. It presents the system architecture and main techniques in CaseDEST including of the strategy of object-oriented representation of the case prototype, neural network based and reconstructed discrimination network based case indexing models and the strategy of variable sensitivity analysis based object scheme modification.

In this paper, a formal model for design reasoning is proposed based on the logic of design. The model embodies the holistic nature of design problem by considering the design synthesis and evaluation in a uniform framework. The kernel parts of the model are exploration and search with the former determining the state space for design solution and the latter looking for the final design proposals in the state space. The whole design problem is solved through recursively resolving the original design requirements and repeatedly using exploration and search.

Existing anti-aliasing methods are mostly applied during the image generation process. The limitations of these methods are that they require a lot of processing time and/or large memory space. Bloomenthal suggested a post-filtering method based on line fitting. In this paper, we describe a similar post-filtering method but is based on curve fitting. The advantage of the new method over Bloomenthal's method are that the shapes of the filtered objects do not change according to which pixels being selected as a first pixels for fitting and that curves fitting is considered more appropriate for object contours than line fitting. To improve efficiency, we identify all possible curves that may appear in a fitting operation in advance and store the information in a lookup table. Fitting a curve to a sequence of pixels becomes a simple table lookup operation.

Real world engineering design projects require the cooperation of multidisciplinary design teams using sophisticated and powerful engineering tools. The individuals or the individual groups of the multidisciplinary design teams work parallelly and independently with the different engineering tools which are located in the different sites for often a long time. This paper presents a distributed architecture for integrating such engineering tools in an open design environment organized as a population of asynchronous cognitive agents. A prototype of such an environment with seven independent agents located in the different workstations and microcomputers is presented and a small mechanical design example is used for demonstrating such an environment.

Determination of injection molding machine setting is usually carried out by an experienced machine operator. By considering the mold geometry, polymer type and machine model, the operator decides an initial setting based on his own knowledge. This knowledge could include material and machine data, empirical rules and simple formulae. A computer system based on this knowledge is developed for computing an initial setting for an injection molding machine. The system first determines the possible processing conditions by considering only the material type and machine model. Then an initial processing condition is determined from these possible conditions by reasoning using the empirical rules. This reasoning is formalized by fuzzy logic. Finally, the machine setting corresponds to the determined initial processing condition is calculated by simple formulae. The performance of the system has been verified by experiments. Results show that the computed setting is acceptable for simple geometry.

To support and control different design processes using Concurrent Engineering concept, all the information required in different design process must be complete, unique and most importantly, sharable. By close look at the assembly process and design process, this paper proposes an integrated information model of a product based upon its assembly relationship-- the assembly model. With all assembly information recorded and inheritated, an assembly model enables us to develop a product (rather than parts) oriented CAD system, to support top-down design process and, to control multiple design tasks and processes concurrently and ensure the consistency of the design model.

This paper presents an assembly modeling method by inputting the assembling constraints between components interactively. It analyzes several common assembling constraints in detail, and derives the corresponding constraint equations. To accelerate the assembling process, this paper also provides the effective measures to reduce the unknowns and equations for solving the transformation matrices of the assembly.

In this paper, the position of color design with computer in surface modeling for electronic products is briefly discussed, and the necessity and feasibility of developing color modeling system are shown. A practical computer-aided color design system suitable to common microcomputers, which is rapid in modeling speed, convenient in interactive modification mode and rich in kinds of colors, is introduced.

In an attempt to create a generic designing approach for mechanisms, an integrating backbone structure has been proposed and is being created. This employs constraint modeling techniques to interface between a specifier, schemer and optimizing facility. Many of the individual units have already been created and tested in previous research programs. Work is now underway upon the integration structure. The completed system will be described and demonstrated. This is designed to allow the user to enter a combination of constraint rules, preferred arrangements, performance profiles and function diagrams. The constraint modeler then attempts to resolve any conflict in the requirements and to undertake various optimization procedures to create a design solution.

This paper presents a computer software system for an often encountered 2D layout problem in the shipbuilding and manufacture industries. For an ordered set of irregular as well as rectangular pieces, the system produces rectangles called modules for the pieces at first, and then places these modules into given rectangular sheets of fixed dimensions such that the number of sheets used is minimized. The system also provides the user with an interaction procedure in order to make itself as flexible as possible. This system has been implemented on IBM PC/AT 486 and compatible computers and is supported by AutoCAD system.

Today, CAD/CAM/CAE tools are highly graphical and interactive. But how an interface can be more helpful is not ascertained yet and still interesting to many software workers. From the object-oriented point of view, based on knowledge deducing and event stack structure, we propose a new interface model SRT. Using this model, we have developed an entirely new interaction technique. In this paper, after a review on the development of users interface models and interaction techniques, we will give a detailed definition of SRT, followed by a discussion of what a role it plays in users' design work. Based on SRT model, we developed SDDT, a Software Document Design supporting Tool of ICSDE system. SDDT consists of a concise expert system and a Software Graphic Document edit Too, along with a Software Textual Document edit Tool. SDDT now has been realized on HP workstations, and the new interface of it brought by SRT model has proven to be a very successful interface. In this paper, we will also describe the interface of SDDT.

The performance of hydraulic system employing counterbalance valves frequently encounter problem due to oscillation which are primarily due to the load inertia, volume of pressure system, response of the counterbalance valve and the flow at which the valve is required to regulate the load. This paper describes an application of CAD via a computer simulation analysis in the design of a hydraulic system employing counterbalance valve. The computer simulation results are compared with those obtained from steady state and transient experimental tests. The analysis shows the advantages of CAD in the design of the hydraulic system, especially to predict the performance of the system for all range of operational conditions.

Virtual Environment is a typical application of time-critical computing. To satisfy time-critical computing in distributed graphics environment, we may employ methods including hierarchical model representation, rendering algorithms with different computational complexity, and parallel processing. In this paper, we present a new time-critical computing model which uses the above three techniques.

The declarative language Ruby provides a coherent framework for representing and developing designs. Sketching diagrams for Ruby programs by hand is, however, time- consuming and error-prone. This paper describes a design sketcher which automates the production of a diagram from a Ruby description.

This paper provides a methodology for manufacturing modeling and simulation which combines techniques descended from object approach and domain analysis technologies.

This paper presents a formal description of multimedia presentations. The description uses a synchronization model synthesized from a number of more well known synchronization techniques, and data abstractions used to represent the multimedia objects based on a hierarchy of classes. Starting with a brief overview of synchronization techniques, we propose a synchronization model, a hierarchy of multimedia objects, a formal description and its implementation.

Streamline construction is an important means for visualizing 3D flow field. Many existing methods are based on regular grid field. Constructing streamlines in structured grid field is more complicated. The main difficulties are the location of an arbitrary point and the interpolation in irregular cells. This paper presents a method for streamline construction in structured 3D flow field. We use structured-point to represent the position of a point and the relationship with its surrounding faces. The point location is made directly in physical space. This method does not need extra memory expense and preprocessing. The implementation and results are also discussed.

Intersecting and visibility ordering are key problems that affect the efficiency of volume rendering, especially when the grid of data field is irregular. In this paper, a partition approach for direct volume rendering the irregular 3D data field is given. In this approach, the image plane is partitioned and the current boundary cell set is divided correspondingly. By exploiting the characteristic of adjacency, the efficiency of intersecting is increased, therefore the rendering speeds up. The implementation of the approach, together with some results is given in this paper.

Color quantization is a process of choosing a set of K representative colors to approximate the N colors of an image, K<<N, such that the resulting K-color image looks as much like the original N-color image as possible. This paper presents a new color quantization algorithm by efficient statistical computations and nearest-neighbor searching. Different from recursively bipartition, this algorithm directly splits the color space into many voxels, and chooses a color palette which permits the most pixels with the colors corresponding to the voxels that they fall. The algorithm has a good effectiveness.

For the CAD applications, a C¹ monotonically shape preserving scheme has been developed. This scheme uses piecewise rational cubic functions. An algorithm has been developed, for the economical curve implementation and pleasant demonstration, which ultimately replaces the rational cubic locally in each of the interval into a constant, linear, rational quadratic or rational cubic functions according to the nature of the slopes and the parameters in that interval. Furthermore, O(h⁴) convergence result is also obtained with the appropriate derivative data approximations.

This paper describes parametric frame space interpolation, a new technique for motion control which allows an animator to define a range of movements using a set of reference movements (a frame space). Once this frame space has been constructed, an animator can specify an envisioned movement in terms of these reference movements or, more precisely, as an interpolation between these reference movements. The result is a high-level motion control mechanism which, while powerful, retains the artistic expressiveness and visual emphasis characteristic of parametric keyframing. Then we will develop simple human locomotion model that illustrates the use of this parametric frame space interpolation technique. This locomotion model blends a range of walking and running movements to form a unified motion framework.

The paper presents the application of digital simulation to survivability assessments of warship power system. The digital simulation models of the damaging probability of electric equipment and the probability of the electric load getting power supply are developed. Based on these models, the digital simulation model of survivability of warship power system has been deduced, which is a generally applicable means of computer simulation and can be used to comprehensively assess the survivability of warship power systems.

This paper proposes a general accelerating rendering approach based on scene operations according to the frame-to-frame coherence of animation sequences. This approach is fit for cases in which the viewpoint is fixed and the scene only consists of polygon planes. We describe the accelerating algorithm of space scanline tracing as one implementation of the approach. This approach has good accelerating efficiency.

Algorithm animation is a powerful approach for exploring a program's behavior. It can clarify how and why an algorithm works. This paper reviews briefly the development history of algorithm animation and its application, and then focuses on the presentation of our research on implementation techniques of algorithm animation. Finally, we describe summarily our further plan and work in algorithm animation area.

A VHDL simulator with debug function will play a very important role in the area of hardware description and design. The debug function makes the simulation model and the algorithm much more complicated, and sets a still higher demand on the simulator. In this paper an effective hierarchical model and a simulation algorithm suited to debug function requirement are proposed. The hierarchical model and the algorithm support a whole set of VHDL, including behavioral structural and data flow description manner, all kinds of data types of signals and variables, hierarchical component configuration and subprogram call, and to support many kinds of interrupt requirement, on time, condition, component, process, subprogram and statement line.

This paper introduces an NC machining process simulator. It enables NC programmers or machining operators to check the effects on the part of the cutter motions prescribed in an NC program before machining is performed. The device model (both geometry and logic) of a machining center, NC program controller and the stock-fixture model on the pallet can be defined in this system so as to provide a visual simulation circumstance of machining projects. The machining process is driven directly by NC program by means of an NC program interpreter. Not only can the material removal be visualized, but also the possible danger of collision and interference among cutting tool, stocks and fixtures can be detected during the cutting progress. The simulator is suitable for 3 through 5 Axes NC machining, especially milling, drilling and boring of cast box parts.

According to hierarchical design methodology, complex digital systems are composed of a number of modules interconnected by nets. This paper presents a high level test generation which is accomplished in two steps. The first step is the test generation of individual modules in isolation. A graph theoretic approach for generating optimal test sequence for sequential modules is developed. The second step is test generation of these modules in the complex digital system one by one. When one module is under consideration for test generation, the others constitute the signal propagation paths. This task is to transform the module I/O sequence into primary I/O sequence for testing and is performed by the forward and backward evaluation procedures developed in this paper. Besides, to guarantee the completeness of the test, the concise and precise functional models for sequential circuit testing are used.

This paper presents the principle, functions, features and implementation of a general graphical user interface management system based on OSF/Motif-JB-UIDS, which is a part of the integrated software engineering environment CASE (Computer Aided Software Engineering), named JB (Jade Bird). The visual and interactive UIMS can help the interface designer to generate user interface automatically and then refine it interactively. It adopts a new method of describing internal application interface based on Object-Oriented ideas to support the separation of user interface component from computational component. JB-UIDS has been implemented on SCO-ODT and has good portability and flexibility.