On 29 May 2020, at 9:37, Eric Rescorla wrote:
6\. To deliver a toolchain that is up-to-date and well regarded by
users.
This seems in conflict with "evidence-led". Suppose the toolchain
was well-regarded by users but empirically less efficient than
other toolchains.
I've got no problem with "efficient" being added to list of "up-to-date"
and "well-regarded", but I wouldn't want to see "well-regarded" removed
(at least with out something else that captures the right sense). We
could get a data-driven result that using a particular toolchain is
incredibly efficient, but if people are unwilling to use it for (perhaps
irrational) reasons of not liking the look-and-feel of it, or even just
because it uses a new way of thinking that people are resistant to, that
weighs against trying to deliver it. I think it's OK for those
principles to be in tension and the LLC has to use its judgment to
figure out what's best (in consultation with the community of users, of
course). But we still want to take that "regard" into account.
And like I said, a different term than "well-regarded" that captures my
concern is perfectly OK.
pr
--
Pete Resnick https://www.episteme.net/
All connections to the world are tenuous at best
