[LEAPSECS] Straw men

Ian Batten igb at batten.eu.org
Mon Jan 9 16:14:11 EST 2012

On 9 Jan 2012, at 2025, Rob Seaman wrote:

> Absence of handy examples has no implications for the preceding statement.

"Just because I can't find examples doesn't mean it isn't happening" is hardly persuasive. Gathered on this list is a pretty solid cross-section of people who've worked in computing for A Long Time, and follow a variety of safety-critical, high-integrity and similar issues, and will have had all sorts of reasons (RFC822 Date: headers, for example) to worry about representation, meaning and normalisation of dates. If none of us can point to ongoing Y2K issues, it's going to be hard to claim convincingly that they're happening, just none of us know about it. Googling around for Y2K in the past few years returns lots of uses of it as a metaphor, particularly for IPv4 address exhaustion, but not a lot of evidence of Bad Stuff Happening.


>>> And since time-of-day will fundamentally remain mean solar time,


>> [Citation Needed]


> See http://www.cacr.caltech.edu/futureofutc/preprints/02_AAS_11-661_Seaman.pdf, pages 7 & 8. Also various messages on this list.

Pages 7 and 8 appear to be assertions that it's true, rather than any solid reasons why it's true.

We can all accept that most developed countries, today, need 1200 local to be apparent noon +/- two hours. Getting that value down substantially is going to be a hard row to hoe, given we can trivially point to successful industrial countries for which it's substantially greater than an hour. So long as time is unidirectional, roughly uniform and doesn't give rise to absurdities, the benefits of large geographic areas sharing a common time far outweigh details like mean solar time. It your problem is that you need |DUT1|<1s for your own applications, say so; attempting to claim that the rest of us will regret interfering with this sacred relationship in terms of our hunting licenses and our ability to get to work in the morning isn't a convincing bogie-man.


More information about the LEAPSECS mailing list