Re: OOP/OOD Philosophy

Robert C. Martin wrote:
> On 15 Jul 2005 10:19:22 -0700, "Mark Nicholls"
> <Nicholls.Mark@xxxxxxxxx> wrote:
>
> >
> >
> >Robert C. Martin wrote:
> >> On 15 Jul 2005 01:35:15 -0700, "Mark Nicholls"
> >> <Nicholls.Mark@xxxxxxxxx> wrote:
> >>
> >> >
> >> >
> >> >Stefan Ram wrote:
> >> >> "Mark Nicholls" <Nicholls.Mark@xxxxxxxxx> writes:
> >> >> >Careful mathematics is a belief system based (usually) on the
> >> >> >axioms of ZFC.
> >> >>
> >> >> It suffices to just specify the axioms as an assertion A and
> >> >> then investigate the consequences of A. It is not neccessary
> >> >> to "belief" that A is actually true.
> >> >
> >> >OK, I agree, but if we are to believe that we can apply the model to
> >> >the universe around us, we must believe that the axioms apply to that
> >> >universe
> >>
> >> Not quite. We use the logical system of math to describe empirical
> >> observations. If the descriptions fail to match those observations,
> >> then we choose new axioms.
> >
> >if they describe the empirical observation then by definition they
> >match!
>
> Consider Euclid's fifth postulate that parallel lines always maintain
> the same distance from each other. He sought a proof that he was
> unable to find. Nowadays we believe we live in a universe in which
> parallel lines are very likely to either diverge or converge. We've
> had to change our axioms.

yes...so.....you now need to believe this new axiom applies to the
universe otherwise your just writing numbers on a bit of paper.

You've just said it "we believe we live in a universe in which parallel
lines are very likely to either diverge or converge".

You've just applied the axiom to the universe to conclude that
assertions made on this axiom, apply to the universe. You have invoked
the mapping.

>
> Michelson and Morely showed that the speed of light was constant to
> all frames of reference. We had to change our axioms.

You've invoked the mapping again.


> >
> >look, if we observer the pairs of numbers
> >
> >our specification is.....
> >
> ><1,1>
> ><2,2>
> ><3,3>
> ><4,4>
> ><5,5>
> ><6,6>
> ><7,7>
> >etc * 10,000 other pairs.
> >
> >then we may conjecture that the 2nd number = 1st
> >
> >i.e. y=x
> >
> >that satisfies the specification exactly, it is an exact model....there
> >are others, but this one works exactly.
> >
> >so if we believe that y=x is in some sense an axiom governing the
> >universe (i.e. the mapping exists between domain and model), then we
> >can pick a new number 123456778 and expect to get
> >
> ><123456778,123456778>
> >
> >seems sensible
> >
> >but if no such 1:1 correspondence exists, ever for any set of
> >observations, then there is no reason to believe it will work, and you
> >are as likely to get
> >
> ><123456778,343534535>
>
> This is exactly what happened when Michelson and Morely did their
> experiment. The long held belief that observed velocity was relative
> to the motion of the observer was shown to fail for very large
> velocities. You pair above characterizes this beautifully.

You've invoked it again.

And showed the value of it, even though we know it still to probably be
approximate.

> >
> >well done, you have just demolited the basis for scietific thought.
>
> It seems to me that discrepancies of the sort you describe are the
> force that *drives* scientific thought. We build a model for
> phenomena that we observe to work in our local context. At first we
> think this model is universal. Then we find other contexts where the
> model does not hold; so we build a new model that holds for both.
> This is how scientific understanding progresses.

Yet this is what you rejected.

You rejected models as approximations, yet understand them to be
central to scientific understanding.

> >
> >Now there is no reason for a model based on a set of experimental data,
> >to ever possibly describe anything outside the originally observed set.
>
> Are we talking past each other?

probably.

>
> >> >So I should really say science and applied maths are belief systems.
> >>
> >> They are. But they are belief systems of a fundamentally different
> >> kind than religion. Math is a belief that conclusions can be derived
> >> from initial axioms through logic. Science is a belief that the laws
> >> of the universe can be derived from observations and described by
> >> math.
> >
> >yet you reject the basis of this belief? You reject the basis of
> >scientific conjecture based on the 1:1 mapping from the problem domain
> >into the model.
>
> Why do you keep on telling me what I reject?

I am trying to clarify what you are saying by telling you what I think
your saying back, it's like TDD, I think your saying X, so I claim you
are, if you aren't you tell me.

"Software is not about modeling the real world (whatever that means).
Software is about describing
behaviors that solve problems."
We've spent a lot of time talking about the nature and validity of
models, the above was your initial reply, you went on to reject
modelling as "intentionally approximate" while SE is about "getting it
right", and TDD being the hard rigourous method for doing this.

Hopefully we've moved on a bit.

I do not believe that you do reject models, just some practioners of
modelling.
I do not believe that you reject a correspendence between problem
domain and solution?

> If I reject something
> I'll tell you.

You have.

> I disagree with your assertion that scientific belief
> is based upon a 1:1 mapping of theory to experiment. Most scientists
> think the mapping is 1:~1, and are much more interested in the
> deviations than the agreements.

New work stems from the deviations, you can't work on a theory that
already exists and is believed, seems a little pointless.

Scientific belief is based on some correspondence, we assume it is 1:1
when we apply the theory to the real world, we throw in another belief
about the universe being more likely to be simple than complex, so we
believe.

In TDD there is a 1:1 correspondence between the outputs of a 'working'
system and the outputs expected of that system, as specified by the
tests.
We hope that the tests exactly coincide with what the client wishes
(even if we realise that many requiremnets are implicit and unwritten),
so there is a 1:1 correspondence between test and client requirement.

if f is 1:1 and g is 1:1 then fg is 1:1.

that is the correspondence in practice.

> >
> >> Religion is a belief based on faith.
> >
> >belief is faith
> >
> >dictionary.com
> >
> >Faith:
> >
> >"Confident belief in the truth, value, or trustworthiness of a person,
> >idea, or thing. "
>
> I'll see your dictionary.com and raise you Hebrews 11:1
>
> "Now faith is the substance of things hoped for, the evidence of
> things not seen."

replace faith with belief.

"Now belief is the substance of things hoped for, the evidence of
things not seen."

nope, substitution still seems to largely hold to me.

>
> >you're playing with words again, and when I question your
> >interpretation, it never seems to match whats in the dictionary.
>
> Watch out for slippery generalizations, they become ad hominems.
>
> I fully realize that the definition of religion I provided was
> circular. In the end, religion is a circular belief system. In the
> end, the reason someone believes is a religion is because they believe
> in that religion.

Again I find this presumptious.

>
> >>
> >> >If you believe F=MA applies, you must believe that time is absolute for
> >> >all observers,
> >>
> >> We observe F=MA to apply in many contexts, and have observed that it
> >> does not apply in others. In those contexts where we have observed
> >> F=MA to apply, it applies for every observation and does not deviate.
> >
> >but this is simply wrong.
>
> Exactly. The model is an approximation. And yet we believe that when
> we apply F=MA to send a rocket to the moon, it will get there. When
> we apply F=MA to our local context, we believe it matches reality,
> even though we know the mapping is only 1:~1.

OK, we are probably simply arguing over sematics.

To me the model is exact, the specification an approximation.

>
> >time is not absolute in any context (it may be coincident, but even
> >that is a highly constrained model where everything moves at the same
> >velocity). F=MA always deviates, we just choose for practical reasons
> >to not care about a 0.0000001% error based on an approximate
> >specification of the universe.
>
> Agreed.
>
> >> Based upon this we conclude (believe) that within the contexts where
> >> we have observed F=MA to apply, it will continue to apply in the
> >> future.
> >
> >*why* will it apply.
>
> That question is the essence of scientific faith. Once we have
> created a model, like F=MA, we *believe* that the universe is stable
> enough to allow us to apply that model in the future.

That stability is usually expressed in the terms of a correspondence.

> We believe that
> the laws of physics are repeatable. The only evidence we have of this
> is a very long history of repeatability.

yes

>
> >because the base assumption of the 1:1 correspondence between domain
> >and model, without this you're whistling in the wind.
>
> We assume 1:1 for application within a constrained context like
> sending a ship to the moon. As Michelson and Morley discovered, once
> we leave our local context the 1:1 mapping can significantly degrade.

the problem her is what is 'our local context'

If I observe X is true for all Z < 1000

is 1001 outside the 'local context', if it is then it implies we cannot
generalise at all, yet we do this all the time.

so for what n is not inside our local context?

So I don't agree, we assume the model holds for all contexts, until
someone comes along and shows us a context it doesn't work, and then
another one, and then another one. Then someone hopefully creates a
model that exactly satisfies all these contexts and again we assume it
works for all contexts, until we are shown those it doesn't.....or
there are none.

>
> >>
> >> >if you believe relativity, you can scrub this belief,
> >>
> >> We have observed the laws of relativity to correctly describe a vast
> >> number of events and processes. And we have also observed situations
> >> that appear to deviate from it.
> >
> >so we believe that one of the assumptions (in the specifications) is
> >incorrect, the model is correct based on the specification.
>
> We *know* that one of the assumptions of GR is incorrect. GR assumes
> that the universe is smooth. QM tells us it is not. GR breaks down
> as we study distances and energies where QM rules.

So the specification is approximate, but the model is exact for that
approximate specification.

> >
> >>
> >> >but there are a whole lot more including and the existence of the empty
> >> >set, and the axiom of choice (or its negation), ZFC (or equivalent) etc
> >> >etc
> >>
> >> Those axioms seem to generate a system of logic that has the
> >> expressive power to describe most events in the world. So far we
> >> haven't found a lot of contradictions. It took quite awhile, and a
> >> lot of trial and error, to come up with a set of axioms that generated
> >> a system that was so expressive.
> >
> >Yet they are beliefs and our continued application of them is based on
> >the 1:1 correspondence, otherwise why believe any application of any
> >model to the universe.
>
> We believed in the previous axioms just as much, until we found holes
> in them.

yes, i.e. we didn't believe in local contexts until someone
demonstrated what 'local' meant.

>
> >On the issue of religion is based on some difference between 'faith'
> >and 'belief'....are you not at best splitting hairs?
>
> I don't believe so. ;-)
>
> Scientific belief is based upon observation. Religious belief is not.
> "Blessed are those who have not seen and yet believe." John 20:19-29.

Tell that to someone who is religious, they often claim to 'see' god in
the world around us.

I don't, but who are you and I to tell them they are wrong?

The problem with 'religious belief' is that we generally cannot derive
experiments to test their validity.

Is there a creational god, outside the universe that does not "play
dice", is the same thing as asking an object of class CFoo whether it
was instantiated by CBar or CGod, it does not and cannot know, infact
worse than that, it's like asking the object who wrote the software,
without even being able to see the code.

To a religious person who seeks 'meaning' in the universe then the
existence and nature of god is important, because the nature of god
implies something about the meaning and intention of the universe,
rather like trying work out the nature of a programmer by looking at
his software, assuming of course that a programmer exists!

science does not and cannot seek 'meaning' it only seeks to describe,
so the question becomes out of its scope.

Some questions cannot be answered, hard pure logic tells us that, so
there would always seem to be a gap for doubt, meaning and belief.

>
> Scientists are doubting Thomases.
>

All of them?
always, thoughout history?

I think not.

.

Relevant Pages

  • Re: OOP/OOD Philosophy
    ... >> then we choose new axioms. ... Nowadays we believe we live in a universe in which ... phenomena that we observe to work in our local context. ... >> Religion is a belief based on faith. ...
    (comp.object)
  • Re: A real world example
    ... single database instance, it cannot permanently identify things in a ... universe that can change. ... Temporal differences are just a context change. ... In the one context, you would compare the ...
    (comp.databases.theory)
  • Re: A real world example
    ... universe that can change. ... Temporal differences are just a context change. ... but their scope is limited to a single database instance. ...
    (comp.databases.theory)
  • Re: Axiomatic set theory is still contradictory
    ... of formation in order to represent valid predicates in set theory. ... Within the context of discourse a variable can only represent one ... where all the axioms are true. ...
    (sci.logic)
  • Re: Goedel - interesting problem?
    ... >> might be true both within and without the set of axioms, ... theorem will need the explanation we are discusssing, ... use, reliability, intended reader, context, context, and context. ... >> reader is not a logician, but from any of a large variety of fields. ...
    (sci.logic)