Prev: jdbc connection
Next: Java/J2EE with Struts | Richmond,VA | 6 months
From: Eric Sosman on 8 Mar 2010 10:32
On 3/7/2010 11:42 PM, Lew wrote:
> Peter K wrote:
>>> But is your quarrel that if I actually went back the billions of
>>> nanoseconds from the value for today's nanasecond value, I wouldn't
>>> actually end up at 1/1/0001 - due to vagaries in the Earth's orbit, spin
>>> etc?
>
> Martin Gregorie wrote:
>> To say nothing of the transitions between the various calendars,
>> which, over the mere 2009 years in that range are probably more
>> significant than spin rate and orbit deviations.
>
> I look at such a system (100 ns "ticks" since
> 0001-01-01T00:00:00.00...Z) as a "normalized" calendar/time system.
> Arguing that you cannot precisely measure 0001-01-01T00:00:00.00...Z as
> a number of ticks ago since NOW is specious; that datetime is *defined*
> by being that many ticks ago from NOW.[...]

In short, the definition is useless, useless in the sense
that one cannot use it to say what the tick count should be at
any given NOW. If you holler NOW! and consult the clocks on
Systems A and B, and the clocks disagree by five minutes, say,
can the definition help you determine which (if either) is
correct? Since the definition is circular ("The current time
is defined as the number of ticks since a moment so-and-so many
ticks ago"), the operators of A and B can *both* claim their
clocks are correct. One might just as well define the time as
the number of ticks since the Ark hit Ararat.

In "The Devil's Dictionary," Ambrose Bierce defined a magnet
as an object exerting the force of magnetism, and magnetism as the
force exerted by a magnet (noting that the paired definitions were
the distillation of innumerable scientific treatises). Circular
definitions can still be funny ("Recursion: see Recursion"), but
it seems not everyone sees the joke.

--
Eric Sosman
esosman(a)ieee-dot-org.invalid
First  |  Prev  | 
Pages: 1 2 3 4 5 6
Prev: jdbc connection
Next: Java/J2EE with Struts | Richmond,VA | 6 months