On 5/31/23 18:39, Jeff Xu wrote: > I think this solution should work. By "work" I think you mean that if laser-focused on this one use case, without a full implementation, it looks like it can work. I'll give you a "maybe" on that. But that leaves out the bigger picture. How many other things will we regress doing this? What's the opportunity cost? What other things will get neglected because we did _this_ one? Are there more users out there? Looking at the big picture, I'm not convinced those tradeoffs are good ones (and you're not going to find anyone that's a bigger fan of pkeys than me).
