Re: Temporal database - no end date
paul c wrote:
> Well, I doubt if any persistent db system can model time thoroughly
> since it will itself be subject to time. Analogies about number
> representations are just red herrings. All a physical system can do is
> mimic some aspect of time in a way that is useful to some purpose. I
> gather that the book that uses "quanta" gave a respectable
> implementation theory for doing this. If somebody doesn't like that but
> can't say why very clearly, that's life, not Bob B's fault.
Replacing 'time' with 'what the application does' and 'quanta' with
'disk files throughout the above paragraph, and we get:
"Well, I doubt if any persistent db system can model what the
application does thoroughly since it will itself be subject to what the
application does. Analogies about number representations are just red
herrings. All a physical system can do is mimic some aspect of what
the application does in a way that is useful to some purpose. I
gather that the book that uses "disk files" gave a respectable
implementation theory for doing this. If somebody doesn't like that
but can't say why very clearly, that's life, not Bob B's fault."
This paragraph, we all agree, is a patent absurdity. Data models are
important because they constitute a well founded framework for
organizing and reasoning about data. One data model is superior to
another to the extent that it has more flexibility, and robustness.
Many of you, it is clear to me, have abandonned even the pretence of
critical reasoning: sacrificing it on the altar of a cult. Well, Date,
Darwen, Fabian Pascal et al are really smart people who's
pronouncements deserve to be treated with a robust skepticism - not
fawning adoration. In my experience useful ideas are useful precisely
because they withstand rough handling.
Many of you, on the other hand, seem to be refusing to engage the
ideas at all.
Received on Sat Jan 20 2007 - 23:05:29 CET
Original text of this message