Date: Wed, 12 Nov 1997 09:09:50 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: nate@mt.sri.com (Nate Williams) Cc: tlambert@primenet.com, nate@mt.sri.com, freebsd-chat@freebsd.org Subject: Re: Newest Pentium bug (fatal) Message-ID: <199711120909.CAA07713@usr01.primenet.com> In-Reply-To: <199711120750.AAA02612@rocky.mt.sri.com> from "Nate Williams" at Nov 12, 97 00:50:34 am
next in thread | previous in thread | raw e-mail | index | archive | help
> > > Simpler != correct.
> >
> > We've been here. Simpler == provisionally correct in the absence of
> > empirical evidence to the contrarary.
>
> We've been here, but I don't agree to your 'waving of the hands' that
> claims it's provisionally correct.
It's a definition for a rule set called the "scientific method".
Do you want to argue the existance of light bulbs? 8-).
> But, the environment is similar enough that in many cases their siblings
> are more alike in certain behaviors to one twin than the two are alike.
I don't accept this statement without empirical evidence to back it;
can you point to the studies that back this up?
> Hence, you cannot model human behavior since it essentially a
> chaotic system. But, even completely chaotic systems exhibit some
> 'patterns', which makes is down-right frustrating when you start
> to rely on those patterns, or make the assumptions that those
> patterns are adequate to fully model the behavior and fall on
> your face. :) :)
When is the last time you saw an insurance company go bankrupt?
The insurance industry relies on actuarial tables. These tables
predict with a high degree of accuracy what will occur within a
large population. They model a chaotic system that exhibits
patterns, and they do not "fall on their face" as a result.
For a less human linked chaotic system, we can look at gasses;
the behaviour of a single atom in a gas can not be accurately
predicted because it will interact with other atoms of gas.
Enough gas, and you can not model the individual components of the
entire system in real time.
But you can model the behaviour of the gas, as a whole, using
statistically derived "laws" (like "PV = nRT"; Pressure times
Volume equalt number of Moles of gas times the gas law constant 'R'
time the Temperature above absolute zero).
Another chatoic system would be one in which you could draw a circle
of an indeterminate radius. I don't need to know the radius to tell
you the relationship between the radius and the area (assuming you
draw perfect circles, and assuming you draw them in Euclidian spaces).
> > > Yes, but you've only described a subset, not the entire thing.
> >
> > How do you think "proof by induction" works?
>
> Proof by induction assumes that the behavior of the system is the same
> across all of the space, and it fails since the behavior and/or model we
> know is incomplete. It works well with numbers since we've arbitrarily
> limited the model to something simple for communication purposes.
Yet we cannot observe that which we cannot observe, and therefore we
must leave it out of our models if we want them to work at predicting
that which we can observe.
You *could* choose to add whatever terms you want to the model, so
long as the terms you ass resolve to the identity matrix for that
which we can observe.
And then the rest of us can pretend those terms don't exist, until
you can come up with empirical evidence to prove their existance.
And until then, we *will* pretend the don't exist, because that
makes the calculations simpler (Occam's Razor).
For example, if we have a system we are modelling, and there are no
relativistic effects involved, we may use Newtonian mechanics to
model the system, and we will not obtain results which are different
from those of using special relativity or general relativity based
models. Such a model is "sufficiently predicitive", despite the
fact that we know that should we speed it up above some point, the
values at the far limits of our precision will start to show error.
> > > > > With the same token energy maybe infinite as well....
> > > >
> > > > Again, the best evidence is that we live in a closed universe.
[ ... ]
> Should I bring in Brian Handy, who *almost* has a Ph.D in Solar
> Physics. And, don't think I'm not willing to use/abuse his knowledge
> and talents. :) :)
Feel free. A Solar Physics person (him) is probably going to be at
least as well versed on Cosomology as a Quantum Physics person (me),
and we can both argue ourselves blue about the missing mass, and both
being outside our specialties, will probably resolve nothing. Then
we'll have to appeal to authority. Like Hawking. 8-).
But if we fail at that, then we can take the standard cosmological
question by the roots ("where did the universe come from") to its
reductio ad absurdum conclusion that it's simpler to say the Universe
has always existed (steady state or not) than that God has always
existed.
> Again, you're using circular logic that I don't agree with. I assert
> that the human being is *more* than just atoms spinning around, but that
> requires Faith. :) :)
To accept?
I agree: it requires faith to accept anything for which one does not
have empirical evidence. That's a definition.
But to cause me to not apply Occam's Razor and accept the simpler
explanation that humans are finite state automatons as my working
hypothesis? For that, you will have to provide empirical data on
the mass/energy equivalent of "human - just_atoms_spinning_around".
Or you will have to prove that a model based on that assumption is
more predictive that one based on the simpler hypothesis. 8-).
Terry Lambert
terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199711120909.CAA07713>