On 6/4/2019 1:42 PM, James Morris wrote: > On Tue, 4 Jun 2019, Casey Schaufler wrote: > >>> It isn't free so there should be a cost/benefit analysis. >> Some benchmarking is definitely in order, but most >> of what's you're calling out as downside is hypothetical >> or based on assumption. > When you're proposing changes such as these, which make fundamental and > far-reaching changes, the burden is on you to present the cost/benefit > analysis. Granted. There has been substantial conversation about it over the years, but I have not done well including it in this discussion. > You can't just say "Here are some changes and here are the benefits, and > any possible costs are merely hypothetical". Of course. Nonetheless, no evidence for performance impact has been provided, while it has been asserted.
