Footnotes

...
The view on ECAD is greatly simplified, there is no layout generation and extraction, no back-annotation, and no schematic level shown in Figure 1.1, which entirely focuses on TCAD aspects and its interface to ECAD. For a discussion of ECAD methodology, see, e.g., [3] and [2].
...
A good review and classification of ECAD systems is given in [2]
...
Currently in use at 7 institutions in Europe, Japan, and USA
...
When in doubt, I have chosen the most popular name instead of, e.g., that of a newly created system component, to denote a TCAD system
...
derivative of SUPREM-4 file format
...
derivative of SUPREM-4 file format
...
described in detail in Chapter 4
...
described in detail in Chapter 3
...
Interesting remarks on the technology dependence of TCAD methodology can be found, e.g., in [14].
...
See also Section 8.5
...
This difference in focus often leads to severe misunderstanding.
...
this topic is picked up again in Section 8.
...
Although problem formulation does not require an implementation at all, the common terms and nomenclature implicitly (or explicitly) defined by a data level implementation can facilitate the problem definition effort significantly by providing a common conceptual basis.
...
See also Section 8.5
...
Note that the framework engineer has to cope with the sometimes conflicting needs of users, tool integrators, and tool developers.
...
This is of course not true for distinct modules within the TCAD system, which solve clearly defined problems and a top-down design can be successfully applied. (See, e.g. Section 6.2.)
...Portability.
The base system integration, as described in Section 2.5, fosters portability.
...Level.
As the TCAD data level is not subject of this text, the reader is referred to [44] for a detailed discussion data level aspects.
...
These services are not a must for an application-framework. A serious TCAD system, however, can not do without them. The generic services may also be provided in form of applications, which is merely an implementation detail. From a conceptual point of view, they may still be considered part of the TCAD framework.
...
This decomposition is not without overlap.
...
VMAKE will be part of another text.
...
Note the re-use and conceptual integrity of this choice.
...
VISTA's data level will be subject of another text, a detailed description can be found in, e.g., [44].
...are:
Standard materialtypes are technology-dependent and are managed by the VISTA material server.
...
These conventions are in addition to the rules and restrictions which are already enforced by the PAI.
...
For example, within the past five years, four entirely different window systems have been in use on Digital Equipment's Workstations (VWS, DecWindows, Motif, Windows NT)
...
See Section 8.6.1.
...
The motivation for a dedicated task level was given in Section 2.1.
...
The importance of the aforementioned requirements for tool integration and tool control (Section 4.1), as well as the need to connect the applications to the interpreter with appropriate and feasible mechanisms can easily be recognized.
...
For an explanation, see, e.g., [85]
...
February 1994
...
The term simplex is used for several other purposes; it is not intended to clarify the meaning of the term as such or its relationship to other fields.
...
An allocation size of 8 entities is assumed for every simplex block in Figure 5.2
...
Ray-casting follows the ray against its propagation direction from the observer to the point of origin.
...
For reproduction purposes, the color image had to be mapped to grey-scale. Comprehension of the figure is facilitated by interpreting it as X-ray image.
...
For the sake of comprehensibility, the definition is slightly inexact
...
All subsequent statements and considerations do not make use of the orientation of the graph. The implementation, however, uses a directed graph.
...
for the sake of simplicity, several details like the propagation of segment information or the quadtree administration must be omitted.
...
Through repreated grid merge operations the number of grid points in the wafer state grid may grow considerably, so that the removal of grid points becomes necessary
...
This implies a data redundancy of a factor of 2 which is accepted for the benefit of consistency verifications and speed advantages
...
In most practical cases the depth hardly exceeds 6.
...
To avoid confusion, one should distinguish between internal computational (simulation) grids which, depending on models and abstractions, need not necessarily be geometry-conform and global static grids which are used to define spatially varying properties within given spatial structures.
...
To be more precise, certain common assumptions [95] about the configuration of the points in have to be made, like the exclusion of collinear or cocircular point subsets. However, these assumptions are only relevant for theoretical considerations as they exclude many important practical applications.
...
The GRS will be part of another text.
...
the evaluation reflects the status VORONOI as of 1993-12-22
...
A refinement with a minimum number of boundary grid points can only be expected when a sufficient and necessary criterion for edge refinement (like, e.g. Theorem 6.5) is used.
...
Note that all process parameters that may be given are simulation parameters as the real processing parameters are not available.
...
The source code size of the etch and deposit simulation module of VISTA [141] is 210kB
...
VORONOI has a source code size of 280kB
...
The application itself is entirely coded in C
...
Due to their remarkable success (which is partly due to the rather narrow focus), the simulators are still in use today.

Martin Stiftinger
Thu Oct 13 13:51:43 MET 1994