Tom Janzen's notes on the article in Dr. Dobbs Sep. 1999

• The files in the zip file have names with initial caps due to being moved from linux through a famous OS from Seattle. You should rename them to change the initial letter to lower case. (Due to an attentive reader, as is the next one.)
• You may also have to include cstddef.h to get "size_t" on some systems. This was my mistake. Stroustrup's The C++ Programming Language, 3rd ed., page 433, clearly says that size_t is in cstddef. I havn't tried this yet.
• The code I intended to have put on the web site did not appear; the version there is one behind; perhaps this was my doing, I don't know. The last changes to the code were:
  • More commenting
  • replacement of RCS headers with identification of myself, and the bibliography from the article.
  • Include statements for Microsoft Visual C++ (TM).
  • A Readme file that explains usage of the program. README.html
• My quibbles about the term "discontinuity" are probably misplaced. The drop in the depth buffer from the edge of an object to the background is just a "silhuoette". The drop from a curved away edge (like the parametric surface in my article) to its own surface, like looking at a waving flag edge-on so that a "furl" partly occludes the flag, or an ocean wave occulding the ocean somewhat but some ocean is seen past the crest, might be what papers call a "discontinuity", and this may be appropriate because if the surface is described in an analytic way (like ocean waves described by a sine function, although inaccurate), it is piece-wise continuous and is differentiable, even in the depth or Z buffer except at the place where it curves behind and occludes part of itself. I think.