Instead, observers should present results in a theory-neutral way. Observations should not converge on one model but aim to find anomalies that carry clues about the nature of dark matter, dark energy or initial conditions of the Universe. Further observations should be motivated by testing unconventional interpretations of those anomalies (such as exotic forms of dark matter or modified theories of gravity). Vast data sets may contain evidence for unusual behaviour that was unanticipated when the projects were conceived. If all results are expected and planned for, babies may be thrown out with the bathwater.
Good data are not enough (Avi Loeb, Nature, 539, 23-25) ちなみにこの標題は we do need to have the correct mindset in interpreting the data. と続く。
隣の芝は緑に見える。データから理論を縛る事が出来ないのが海洋物理。もちろん海水は universal な性質を持ち universal な方程式に従うが、境界条件がいけない。天井は空間変動も時間変動も大きくてさっぱり測れていないし、壁というか海底地形も、時間変動はこの際無視するとしても十分な精度で測れていない。ほとんどの領域で open boundary だし。あと「理論」といってもナヴィエストークスをどう近似するかというだけの話で、近似したもんだからデータと合わなくてもまあそんなものか、と。
誤差棒を使った煮え切らない議論を続けるか、big data とかアンサンブル予報とか deep learning とか、データと「答え」を直結するか、ほかになんかうまいやり方はないもんか。
ここにも「選択と集中は駄目」という意見。
The tendency to establish large projects and firm up mainstream ideas is a signature of a mature scientific discipline. In such a culture, the low-hanging fruit has already been picked by small, versatile teams that are long gone. Critics argue that when funds are limited, the focus of research should be on coordinated approaches that are likely to produce results in a predictable way. This advocacy fails to appreciate that our mainstream paradigm might be heading in the wrong direction. The opportunity for making mistakes is much greater than for real breakthroughs, so as any venture capitalist knows, investing part of the portfolio in risky endeavours is necessary to gain substantial profits.