Most authors of SGP or SIGGRAPH publications have to develop software as proof of concept and for qualitative and quantitative comparison with prior art. This software is often just a prototype that works on some data sets.
Sometimes software is mature enough to get distributed, so that others can reproduce results, check wether the algorithm works on their own real world data, and compare with another algorithm for the same problem, or another implementation of the same algorithm.
Only a small fraction of software gets developed with the ambition to produce “high quality” software.
GeometryFactory sponsors the SGP Software Award in order to encourage scientists to go the extra mile it takes to turn a research result into “high quality” software.
It takes an extra mile, as it means working on robustness so that the algorithm also works for degenerate or noisy input data, as it means cross-platform support, that is different compilers and third party software), as it means working on API design, as it means to write a documentation and test cases, as well as examples and tutorials, as it means dealing with users and bug reports, and reviewers in case the software is part of a library. It finally means a long term commitment as software needs maintenance.
All this is an effort we should not underestimate. We heavily profit from it as it allows to leverage on existing solutions when we work on new problems, inside as well as outside of the geometry processing community.