> I think it's fairly uncontroversial that randomized controlled trials are the gold standard.
Software with 99.999999% uptime [0] may be the gold standard, but that's not always practical, and we can benefit from other software too.
If everything with a flaw in the world is useless, then everything in the world is useless, including all these HN comments and whatever you and I do for our livings.
[0] I'm too lazy to actually imagine what the gold standard for software development is, sorry.
>Software with 99.999999% uptime [0] may be the gold standard
No one seriously has that many nines as their "gold standard", because that's almost always too safe, which has costs:
>We strive to make a service reliable enough, but no more reliable than it needs to be. That is, when we set an availability target of 99.99%,we want to exceed it, but not by much: that would waste opportunities to add features to the system, clean up technical debt, or reduce its operational costs. In a sense, we view the availability target as both a minimum and a maximum. The key advantage of this framing is that it unlocks explicit, thoughtful risktaking.
In this case though, other commenters have pointed out plausible reasons for why the cohort studies can't be trusted. There's also lack of evidence for the casual mechanism behind the effect, so on the whole it seems fairly reasonable to demand better evidence.
Software with 99.999999% uptime [0] may be the gold standard, but that's not always practical, and we can benefit from other software too.
If everything with a flaw in the world is useless, then everything in the world is useless, including all these HN comments and whatever you and I do for our livings.
[0] I'm too lazy to actually imagine what the gold standard for software development is, sorry.