8 ms·
> It's possible, even probable, they look at your results and then give you permission to publish. Which makes observable results inherently biased and any com
by ajuc 5y ago
> It's possible, even probable, they look at your results and then give you permission to publish.
Which makes observable results inherently biased and any comparison useless.
- 1vuio0pswjnm7 5y agoHow so. I dont understand.
- ajuc 5y agoBecause what constitutes "bad" benchmark is subjective and db providers are incentivized to stop publishing benchmarks that put them in bad light. And even if all they do is filter out actually crappy benchmarks - other db providers (like postgres or mysql) don't do that - so the comparison still isn't fair. It's like you had 2 candidates for a job and phoned their previous bosses, but candidate A forced you to only phone guys he approves of.
- 1vuio0pswjnm7 5y agoUsually there would be some language that states they cannot refuse permission to publish unless they have reasonable grounds. Your hypothetical, i.e., a "bad" benchmark, would not qualify as reasonable grounds. Blocking the publication of benchmarks because they do not suit one's business interests is not going to be an enforceable provision in a license agreement. I challenge you to name a single case that went to trial over such a provision where a court upheld the provision's enforcement.
- burnished 5y agoMight be worth considering a world in which legal pressure was enough to modify peoples behavior, even if in practice it was unenforceable. Still, wouldn't mind some examples, or a case study.