Re: Temporal database - no end date

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Sun, 21 Jan 2007 00:30:49 GMT
Message-ID: <Zuysh.3402$1x.57048@ursa-nb00s0.nbnet.nb.ca>

Marshall wrote:

> On Jan 19, 2:04 pm, "DBMS_Plumber" <paul_geoffrey_br..._at_yahoo.com>
> wrote:
>

>>V.J. Kumar wrote:
>>
>>>Could you show with an example how the loss occurs ?  You may be right,
>>>but let's see.
>>
>>Suppose I have 5 things. 4 of them lasted for 1 'day' before they
>>break.The other broke on the same 'day' it went out.
>>
>>   What is the mean time to failure?
>>
>>     Sum of 'time quanta' = 4.
>>     Number of things      = 5.
>>
>>    Mean time to failure   = 0 time quanta.
>>
>>    Pick any intuitive unit of 'time quanta', and you can construct an
>>example where this problem arises.

>
> This example only shows one thing: aliasing. Aliasing is well
> understood
> for decades now. The way you avoid the problem above is to use
> higher resolution. Use milliseconds instead of days and this problem
> disappears.
>
> In fact this is the exact same issue as any other use of digital
> signals to model continuous functions. If this were an actual
> impossibility as claimed, then digital audio would not work, and
> jpegs would look like gray goop. But of course the actual issue
> is being ridiculously overstated. Almost the entire world models
> the reals with floating point; almost the entire world models audio
> with digital data; and an ever increasing fraction of the world
> does the same with video. IEEE 754, MP3, HDTV, anyone?

Even if one used chronons, ie. very small indivisible units of time, the issue would not exist because one doesn't necessarily coerce the result of a division into an integral type, and one calculates mean using division the last time I checked.

Children definitely arrive in discrete integral quanta, but we can still calculate total fertility of 1.6 or 2.7 for various groups of women. Received on Sat Jan 20 2007 - 18:30:49 CST