The Visual Computer

, Volume 8, Issue 1, pp 35–46 | Cite as

Step length in forward differencing for curved surface display

  • J. J. Steketee


Principles and basic algorithms for evaluating polynomial functions by Forward Differencing (FD) and Adaptive Forward Differencing (AFD) are presented. Both techniques are used nowadays in scanning/display algorithms for parameteric polynomial surfaces by generating and displaying a set of points. Constraints to be put on such a set of points are made explicit. One FD and two AFD surface-scanning/display algorithms presented in literature are evaluated in terms of these constraints. The FD algorithm is proved to be valid, but the AFD algorithms are shown to be incorrect. A valid AFD algorithm is presented.

Key words

Scanning/display algorithm Forward differencing Parametric polynomial surface 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Bronsvoort WF (1990) Direct display algorithms for solid modelling (PhD thesis). Delft University Press, Delft, pp 99–108Google Scholar
  2. Issacson E, Keller HB (1966) Analysis of numerical methods. Wiley and Sons, New YorkGoogle Scholar
  3. Lien S-L, Shantz M, Pratt V (1987) Adaptive forward differencing for rendering curves and surfaces. Comput Graph 21(4):111–118Google Scholar
  4. Newman WM, Sproull RF (1979) Principles of interactive computer graphics. McGraw-Hill, LondonGoogle Scholar
  5. Press WH, Flannery BP, Teukolsky SA, Vetterling WT (1988) Numerical recipes in C, the art of scientific computing. Cambridge University Press, CambridgeGoogle Scholar
  6. Rockwood AR (1987) A generalized scanning technique for display of parametrically defined surfaces. IEEE Comput Graph Appl 7(8):15–26Google Scholar
  7. Shantz M, Chang S-L (1988) Rendering trimmed NURBS with adaptive forward differencing. Comput Graph 22(4):189–198Google Scholar

Copyright information

© Springer-Verlag 1991

Authors and Affiliations

  • J. J. Steketee
    • 1
  1. 1.GEMOREGeometric Modelling ResearchZwijndrechtThe Netherlands

Personalised recommendations