1 TRANSACTION EXPERMENT

L. Tran,Rui Zhang
2007-01-01
Abstract:As a scientific field, computer science has long been viewed as the hybrid of engineering and mathematics, whereas the science component is often less paid attention to. The lack of scientific methodology raised the questions of credibility, validity, and reliability against computer science research, especially on the issue of experiments. Normally, the experiment results are limited within the bound of performance comparison of algorithms and implementations, which are considered engineering and mathematical efforts, rather than searching for significant and interesting phenomena, and the explanation of such phenomena via theories and the verification or falsification of relevant theories and hypothesis, like other scientific fields do. In this paper, we will be discussing the efforts and approaches we tried to perform scientific experiment on DBMS and the hypothesis we raised on some phenomena we found during the experiments. Also, we will emphasize on the design of our experimental platform ExpTran, on which we conducted our experiments.
What problem does this paper attempt to address?