On Mon, 24 Jun 2019 at 17:19, <Paul.Koning@dell.com> wrote:
I wonder if it would be helpful to discuss this more explicitly in the THEORY file.
I would support such an endeavor and I think much of your proposed language is good. However, I think that while "reflect[ing] 'reality on the ground'" does succinctly capture the essence of what we're trying to accomplish into a single phrase, it perhaps still comes off as a bit too trite, especially for the uninitiated, and therefore warrants a modest elaboration to help support its meaning. Specifically, we are interested in how *groups of people* use their notion of local/civil/jurisdictional time to coordinate and organize. For example, if the viewpoint of people in Region X, loyal to Authority A, is that they are being occupied by Authority B, but everyone in Region X regardless of allegiance references their local time in terms of the same authority, then only that authority matters for the purposes of the timekeeping portions of tz, and the "reality" of civil timekeeping there is the sole reason behind that decision. On the contrary, it should be clear that there are certainly situations in which it can, and indeed does, make sense to have tz reflect multiple overlapping and differing viewpoints in its timekeeping data — that is, *if and only if* those viewpoints differ on civil timekeeping practice in the region in question. (See Asia/Urumqi, c.f. Asia/Shanghai for a current example.) Some distillation of that (perhaps the first sentence) should be enough. I should hope we can safely avoid mentioning that we cannot and should not aim to document the *ad hoc* practices of any possible group of two or more humans, or that a plurality of a region following a given authority doesn't necessarily mean that that is the only authority that matters. The line must be somewhere in the middle, and must be evaluated (likely case-by-case) based on factual evidence of actual usage. -- Tim Parenti