In 1996, something amazing happened: the discussion forum of the computational geometry community was actually used! Plus, it was used in a very interesting and instructive way. The bad news is that it wasn't used for long. Maybe at some point, everything had been said, or people simply lost interest. Anyway, here is a (more or less complete) account of that historical discussion. It's about the future of theory, and I think the material is still quite interesting. It would even be more interesting to have an update to that discussion now: did it change the way we work? Are the arguments still true, or did things change with the increasing availability of programs and not only algorithms? I'm thinking about CGAL here, for example.
One thread has been started by Chazelle at al.'s CG Impact Task Force Report Application challenges to Computational Geomtery . David Avis and Komei Fukuda subsequently posted a critical review of this report to the compgeom-discuss mailing list. Check out Nina Amenta's reply and Ken Clarkson's comment to this. Supporting Avis and Fukuda's main points, Azriel Rosenfeld discusses the Computer Vision section of Chazelle's report, while Paul Heckbert's contribution deals with the computer graphics and mesh generation sections, preceded by his general view concerning the state of computational geometry. Trevor Coulson's followup to Heckbert's comments presents an industrial point of view, as well as Ernst Mücke's proposal to clean up the `CAD zoo' (which he apparently got caught in). Finally, Randolph Franklin contributes some concrete, but also philosophical remarks to the `theory vs. practice' discussion.
Another thread has been opened by Aho et al.'s strategy paper Theory of Computing: Goals and Directions . `Gravely concerned' with its contents, Goldreich and Wigderson wrote their objection Theory of Computing: a Scientific Perspective.
A discussion about geometric software and how it should look like has just started. See D.T. Lee's contribution (which explicitly addresses the European project CGAL ) and Kumar Ramaiyers's followup .
Eventually, the discussion died... I may have missed a few contributions, but the bulk of it is here. For more and other references please check out this.