[Discuss] Abolish DST (was This year's Beowulf Bash is not for the lily-livered)

[grg-webvisible+blu at ai.mit.edu (grg)]

On Mon, Nov 22, 2021 at 02:55:41 PM -0500, Kent Borg wrote:
> Nope. Youngsters know time is simple and still write code that assumes 
> things are today as they were when I was born: a second is a fixed 
> fraction of a day.
...
> Time is complicated.

I couldn't agree more that time is complicated, but I don't agree that it's
a matter of youthful or, shall we say, substantially experienced
programmers.  at any age, imho the only way to really appreciate the inane
and esoteric vagaries of time is to either implement a time library or
build an application where time critically matters, such as synchronizing
worldwide communications or scheduling global airline flights.

do either of those and you realize that everything about time, without
exception, is just a construct of people and the agreement between them
(spelled "politics", but in the broad sense).  we sciency/engineeringish
types think time is an absolute, scientific reality -- a second is exactly
a second in any reference frame you're in, right?  even if that were true
(it's not), *everything* we ever say or compute about time is just an
arbitrary human fabrication.

across the globe, people don't agree on how long a year is, how long a day
is, how long a second is (forget about "month!"), when a day/year/etc
starts and ends, whether we use the sun, the moon, stars, atoms, or some
combination of those as a reference (all of which we tweak as it pleases
us).  not to mention the most obvious, its variation by location - where
the subject of the current argument (EST/EDT/AST/ADT/EPT...) is based not
only on capricious applications of geographic and governmental boundaries,
but even the fact that they're 1 hour apart is the product of someone's
whimsy (and more whimsy means they're not even all 1 hour apart).  tzone
varies by dictum, years vary by decree, seconds vary by convention, and it
all changes whenever a random neuron fires in someone with the ability to
get other people to follow that neuron.  (which means that rationalizing
the past is yet an additional layer of annoyance...)

and we're supposed to teach computers to deal with this garbage??  what a
colossal waste of time (measured in which standard? ;) - even if necessary
if we want computers to interact with these inconsistent and mercurial
humans.  (do I sound bitter about this? ;)

everything about time is arbitrary and it changes all the time.  forgive me
if I can't get worked up about a 1-hr shift every half revolution or so.


> (fuzzy) Google definition of the second. Mostly it is dang precise and 
> stable, but every year or so, it starts to slew wildly away from its 
> usual precise duration and then slew wildly back
...
> I say fuzzy because I am pretty sure how and when the slewing happens is 
> not well defined, is probably not consistent from one leap second to the 
> next. And this odd time standard is distributed via NTP, which was not 
> intended to distribute a non-stable reference, so the result is going to 
> be a mess from any time-standardization perspective.

fwiw, I think google's proposal is well defined & documented, and they make
a reasonable argument that the change is well within ntp and commodity
system clock tolerances ("11.6?ppm...within the manufacturing and thermal
errors of most machines' quartz oscillators, and well under NTP's 500 ppm
maximum slew rate"):
	https://developers.google.com/time/smear

and don't blame google only, there were plenty of other made-up versions of
this before google said "hey, let's all hack this smear thing the same way"
(to which everyone else said "forget you, I thought of a different hack so
whatever I made up is obviously superior").

ok, so I guess I do get worked up about this stuff, just not about the dst
part of it specifically. ;)

--grg