Hacker News new | ask | show | jobs
by jjk166 1055 days ago
> and someone can make it

If the specified procedure is incorrect, then we can't make it. It doesn't need to be an elaborate con, it could just be a reasearcher misread a measurement, or recorded the wrong number, or their feedstock was contaminated. Replication ensures that the recipe includes the secret sauce that makes it actually work.
1 comments
lolinder 1055 days ago
I don't think they're arguing that no one should try to replicate the process of making it, just that it makes sense to have another lab test the sample that has already been created. If it is in fact what they claim it is, then the worst case scenario is that we have a repeat of FOGBANK[0] and have to reverse engineer it.

[0] https://en.m.wikipedia.org/wiki/Fogbank

link
jjk166 1055 days ago
I'm not arguing that another lab shouldn't test the sample already created, just stating why replication is important. A FOGBANK situation is a very bad scenario.
link