小雷诺数二维圆柱绕流问题

luo@odu.edu

Honesty is something so elemental in SCIENCE that I did not think that it should be even an issue. Now I am older (not necessary wiser) and have seen more, I start to wonder what kind of enterprise of "science" we are in. Perhaps we should remind ourselves that no one is privileged to be immuned of the common diseases like dishonesty.

The problem of dishonesty may not be necessarily more pervasive in numerical work. It is just that numerical simulation by nature is more or less a man-made subject, which is more likely to be manipulated. Moreover, "comparison with experiments" may also be used a way to "fix" the data, or to justify the wrong numerical results...

[ 本帖最后由 luo@odu.edu 于 2013-9-28 20:30 编辑 ]

通流

我想说明一下，不诚实并不等于我们国内的“造假”。搞产品设计的人也不一定能够保持诚实。虽然搞产品的人能够成功的做出好的产品，并不等于他明白了为什么。也就是说做实际东西的人，经常会误导别人，也误导自己。对于搞计算的人，有时候去调整一些参数，这几乎是习惯性的行为。而这些行为基本不会从论文中反映出来。这些行为不止一般的人有，甚至那些超级大牛，比如Jameson，都有。我说的Denton这方面就做得好一些。所以我觉得“诚实”和“严谨”之间是相关的。

luo@odu.edu

1) It is a very good and important point to distinguish "honesty" vs. "rigor".

2) There is nothing wrong to "tune parameters", however, it MUST be transparent.

Jameson is a very good example. One reason Jameson becomes a bigshot is because, not only his tricks (or tuning schemes) work, some of his tricks have mathematical/physical basis, and he has been trying to justify his tricks scientifically. Not all bigshots are equal, and we must know how to discriminate.

3) The least one can do is to maintain reproducibility of the data.

What we have been talking about are not some fine lines, but some important principles.

通流

这次看onesupeng和Luo教授的讨论让我受益匪浅。希望大家也跟我一样，从中能够学到一些东西。